|Home | About | Journals | Submit | Contact Us | Français|
Standard treatment consisting of surgery, irradiation plus concomitant and following T is established in GBM with median survival of 15.6 months, indicating the need for further effective treatment. In several phase II studies B plus I showed impressive objective response rates (>50%) in pre-treated GBM pts with low toxicity rate, duration of response, however, was moderate. Therefore, B plus I might be an effective induction regimen for T resistant GBM patients as basis for further treatments. In colo-rectal cancer a lower dose of B plus I was effective, therefore B and I dose was reduced.
From December 2006 to October 2007, 44 pts with progressive GBM resistant to T were treated with B 4 mg/kg body weight intravenously (iv) followed by I 80 mg/m2 iv repeated every 2 weeks. ECOG performance status (PF) was 0–2 in 43 pts, 3 in 1 patient. MRI scans were required at baseline (no hemorrhage was allowed) after 4 weeks and afterward every 6 weeks. Treatment was given until progressive disease (PD) or intolerable toxicity occurred.
All 44 pts were eligible for toxicity and efficacy; median follow-up was 7 months; 33 pts were male, 11 female; median age was 45 years (21–79); 32 pts had primary GBM; and 12 pts secondary GBM. All pts had prior irradiation (56–60 Gy); 4 pts had 4 prior chemotherapy regimens, 11 pts 3, 22 pts 2, and 7 pts T only. The only patient with PF 3 died of clostridium sepsis during grade IV leucopenia after the first treatment; 1 patient developed grade III leucopenia; 2 pts grade III thrombopenia; 1 patient grade III pneumonia; 2 pts asymptomatic intracerebral bleeds requiring treatment delay; and 1 patient grade III fatigue. In 22 pts a partial response (PR) was achieved; in 15 pts disease stabilization for at least 2 months; and 7 pts showed primary PD. Median duration of PR was 3 months (2–8); best MRI response was acchieved after 4 to 8 weeks of treatment. Data will be updated.
B plus I seems to be the most effective regimen for induction of objective response in multiple pretreated GBM pts with excellent toxicity profile. Efficacy of the low dose regimen was comparable to other published regimen. In targeted therapy the principle of maximum tolerated dose (MTD) is not necessarily identical with the optimal treatment dose (OTD) which might be much lower. Confirmation is required. A following maintenance treatment should be considered.
Primary as well as secondary glioblastoma typically exhibits striking neo-angiogenesis with elevated expression of vascular endothelial growth factors (VEGF) and their receptors (VEGFr). These signals promote aberrant blood vessel formation by recruitment of endothelial cell precursors and facilitate sustained tumor growth. Associated edema within the tumor and surrounding normal brain is a major cause of disease-related symptoms and morbidity in glioblastoma patients. Sunitinib is an oral small molecule that inhibits the activity of the VEGFRs, FLT1, FLK1/KDR, PDGFR-a and -b, c-Kit, and the FLT3 and RET kinases.
We evaluated whether sunitinib at a continuous daily dose of 37.5 mg is capable of attenuating intra-tumoral VEGFr-mediated signaling, thereby causing antiangiogenic and antiproliferative effects in patients with recurrent glioblastoma. We assessed the antiangiogenic response after 4 weeks of sunitinib therapy by calculating the cerebral blood volume (CBV) and cerebral blood flow (CBF) from dynamic susceptibility (DSC) based perfusion MRI. We compared the CBV and CBF values in the whole tumor, in the contra-lateral white matter, as well as the lesion-to-normal-white matter CBV (CBVLTN) and CBF (CBFLTN) ratios.
As part of an ongoing phase II study, 21 patients with recurrent GBM were recruited since July 1, 2007 (median age 43 [range 34–71], M/F 15/6). All patients had progression following surgery, radiotherapy and temozolomide chemotherapy. Nineteen patients have initiated study treatment and results are available on 16 patients. Sunitinib resulted in early anti-angiogenic effects in 6 patients that were manifested by a clinical improvement with associated early decrease of CBVLTN and CBFLTN (p<0.05, r = −0.683). In 2 responding patients, the treatment was interrupted because of grade III or IV hematological toxicity, with subsequent tumor progression. One of these 2 patients was able to resume sunitinib, with again a rapid clinical improvement. In this particular patient with a durable response on sunitinib, reduced amino acid metabolism was documented by 2-18F-Fluoromethyl-L-Phenylalanine PET-scan differentiating true cytoreductive activity from a steroid-like effect. Despite continued therapy, all other responding patients had PD within 4 mths. Four additional patients have remained clinically stable at the latest follow-up (<4 mths) and 6 patients had immediate progression. Artefacts prevented an adequate interpretation of perfusion imaging in 3 patients.
DSC based perfusion MRI seems valuable in measuring the early antiangiogenic activity of sunitinib (37.5 mg on a continuous daily dose) in patients with recurrent glioblastoma. Sunitinib causes early reduction in CBVLTN that correlates with an improved clinical status.
We report on activity and feasibility of a rechallenge of recurrent gliomas with temozolomide (TMZ) in a dose-dense 21/28-day regimen and development of a novel 5/7-day schedule.
40 recurrent glioma patients received TMZ 100 mg/m2 with individual dose adaptation: 21 pts. received 148 cycles days 1–21/28 (18 WHO IV [GBM], 3 WHO III). GBM pts. had a mean age of 54.4 y, median KPS was 60%. Patients with critical blood counts (~50%) were switched to a 1–5/7-day regimen. Better tolerability led us to treat 19 subsequent patients initially days 1–5/7: 13 GBM, 2 WHO III, 4 WHO II; 55.5 y; KPS 70%; total 61 cycles. All malignant gliomas were pretreated with temozolomide.
Nausea and fatigue were less frequent than with the 5/28-day schedule. Blood counts usually decreased continuously. Dosage was reduced in approx. 50% of patients and increased to up to 130 mg/m2 in 3 patients. Hematotoxicity grade 3/4 (asymptomatic): 2/21 pts. (21/28-day) and 3/19 pts. (5/7-day); nonhematological 3/4: 4/21 pts. (21/28) and 0/19 pts. (5/7). Efficacy of 21/28-day application in GBM was: 3 CR (17%), 1 PR (5%), 7 SD (39%), and 7 PD (39%) for at least 3 months. Progression-free survival at 6 months was 38.8%, at 12 months 22%, and at 24 months 6% (1 pt.). Survival after relapse was 34.8 weeks and overall 17.9 months. 3/6 pts. with unmethylated MGMT were progression-free >12 months.
Rechallenge of recurrent gliomas with dose-dense temozolomide initially at 100 mg/m2 and individual dose adaptation is feasible and active. 22% of objective long-term responses demonstrate that this strategy is effective even with unfavorable prognostic criteria. Dose adaptation is most appropriate with a near-continuous application days 1–5/7, which is expected to cause permanent depletion of the alkylation-resistance factor MGMT. In fact, patients with unmethylated MGMT also had long-term responses, indicating that chemotherapy resistance may be overcome with this strategy. We therefore consider TEGWONDO (TEmozolomide Glioma WOrkiNgday DOse-dense regimen) to be active and feasible also in TMZ-pretreated patients with critical blood counts.
The combination of radiation with temozolomide (TMZ) 75 mg/m2 × 6 weeks + adjuvant TMZ 150–200 mg/m2 given 5 days out of 28 (5/28) is the standard of care for glioblastoma multiforme (GBM). However, a large proportion of patients will progress after this primary intervention and second-line regimens are modestly active at best. Continuous dosing and dose intensification have been shown to decrease levels of O-6-methylguanine-DNA methyltransferase (MGMT), which has been associated with TMZ resistance. Thus, altering the schedule of TMZ administration and dose intensification would appear to be a reasonable strategy to reinduce response.
After signing informed IRB-approved consent, patients with high-grade glioma who failed the standard TMZ 5/28 adjuvant regimen received continuous dose-intense TMZ 50 mg/m2 for 28 days out of 28 (28/28) for up to 1 year. The primary endpoint was 6-month progression-free survival (PFS). The trial used a two-stage design in which at least 1 of the first 15 patients enrolled needed to achieve 6-month PFS before the next stage of full accrual was initiated.
120 pts were enrolled at 11 centers; 90 of these patients had GBM at first replapse following standard chemoradiation. Patients were divided into four cohorts: GBM patients failing during the first 3–6 months of adjuvant therapy (B1); GBM patients failing after more than 6 months of therapy (B2); GBM patients who recurred after stopping treatment (B3); and anaplastic glioma patients (A). Median age for the entire sample was 52 years (range 25–73 years). The majority (66%) were male. ECOG performance status was 0 in 47% and 1 in 53%. All pts are expected to reach the primary endpoint by June 2008 and will be included in the EANO presentation. An interim analysis was performed on the first 90 patients in 02/2008. The 6-month PFS rates were 28.6% (B1), 9.5% (B2), 30.4% (B3), and 42.1% (A). Nausea and vomiting was observed in less than 5% of patients. Other toxicities were rare. Progressive lymphopenia was observed in some patients, but no clinically significant event has been observed to date. Prophylaxis for Pneumocystis carinii pneumonia was not required.
Continuous dose-intense TMZ 50 mg/m2 administered on a 28/28 day schedule is active and well tolerated after failure of the conventional 5/28 day regimen. Efficacy compares favorably to other commonly used second-line agents. Final 6-mo PFS rates are pending, as is correlation with MGMT status. Continuous dose-intense TMZ may represent an ideal regimen for use in combination with other agents such as the new targeted therapies.
Due to the primary treatment of patients with glioblastoma multiforme (GBM) with alkylating drugs (AD) the role of a rechallenge with AD upon recurrence is subject to discussion. In addition, the impact of cytoreductive surgery for GBM, as demonstrated for the primary treatment, has not yet been evaluated for recurrent GBM. We therefore investigated the role of re-resection and intensified alkylating chemotherapy by both local and systemic regimens in a single institution series of patients with radiographic recurrence of GBM, stratified for MGMT status.
Since 12/2006, 25 subsequent patients with recurrent GBM under Stupp regimen were included after informed consent was obtained. Patients were subject to re-resection and implantation of carmustin wafers and were treated with temozolomide one week on/one week off (150 mg/m2/BS), 4 weeks after surgical treatment. Patients were evaluated pre- and postoperatively by Karnofsky and NIH scores and serial MR-images (all parameters were collected presurgical, <72 h after surgery and every 3 months). Toxicity was closely monitored every week. Overall survival (OAS) and progression-free survival (PFS; McDonald criteria) were determined after 3 months (PFS-3) and 6 months (PFS-6). The MGMT promoter status was determined by methylation specific PCR.
All patients were resected with less than 5 ml residual tumor volume. Severe toxicity after WHO CTC was observed in 16 patients [Grade 3 (n = 10) and 4 n = 6)]. All patients demonstrated a decreased Karnofsky and NIH score immediately postoperatively, which stabilized until disease progression at a median of 70% and 7.0, respectively. As yet, the median PFS is 17 weeks, PFS-3 39%, PFS-6 0% and median OAS 18 months (first diagnosis to death). 72 % of the patients were MGMT negative.
As yet, there are no data on the prognosis of patients with a recurrent GBM after Stupp. Therefore, the interpretation of our outcome data must be approached with caution. The high percentage of MGMT patients probably reflects a negative selection bias, due to the consecutive inclusion of recurrent patients in a single center study. However, the OAS demonstrated in our population compares favorably with the OAS of the MGMT negative subpopulation published by Hegi (20/12.7 months). Though the toxicity of this regimen is acceptable, the clinical deterioration of the patients is not only attributable to the natural course of this disease.
No drug has yet been proven effective in patients with GBM at time of failure after standard radiotherapy plus concomitant and adjuvant temozolomide. Although nitrosoureas may be considered as the gold standard, no data are available in an homogeneously pretreated population.
Pts with recurrent/progressive GBM after ≥3 mo from the end of radiochemotherapy received 3 weekly doses of FTM (75 mg/m2 i.v.) followed, after a 5-week rest, by FTM (100 mg/m2 i.v.) every 3 weeks for < 1 year; treatment was suspended if disease progression or unacceptable toxicity was observed. The main endpoint was to ascertain progression-free survival at 6 months (PFS-6). According to Fleming’s design, a sample size of 40 patients was planned assuming P0 = 0.10, P1 = 0.25, α = 0.1, β = 0.1
Between April 2005 and May 2006, 6 Italian GICNO network centers enrolled 43 pts (29 M, 14 F; median age, 52 [range 34–68] years; median KPS 90, range 70–100). PFS-6 was 21% (SE 6.2%); 3 patients (7%) had partial response (PR), and 15 (35%) disease stabilization (SD). Two out of 3 responsive pts are alive after 21 and 22.7 months, and 1 died after 15.5 months. Median survival time was 6 months (95% CI: 5–7). MGMT methylation status was determined in 34 pts (73.5%): 8 (23.5%) were methylated and 26 (76.5%) unmethylated. In methylated pts, disease control was significantly better than in unmethylated pts (75% vs. 34.6%, p = 0.05); however no significant difference in terms of PFS-6 and survival was observed. Grade 3–4 thrombocytopenia and neutropenia were observed in 25.6% and 30.2% of pts, respectively.
The findings of this present trial, the first to analyze second-line nitrosourea treatment in a homogeneous population following standard treatment, may represent a new benchmark of nitrosourea activity. However, as disease control was not long-standing, alternative doses, schedules, and drug-combinations should be evaluated in future studies.
This trial has been supported by an unrestricted grant of Italfarmaco Pharmaceuticals, Milan, Italy.
To analyze the value of preoperative multimodal MR-imaging (1H MR- spectroscopy [1H MRS], fMRI, CBF) in patients with dominant hemisphere gliomas in close relation to language areas with respect to surgical treatment strategy.
Data derived from a prospective series of 19 consecutive patients with newly diagnosed gliomas in the dominant hemisphere in close relation to anatomical language areas treated between November 2005 and November 2007. Mean patients’ age was 45 ± 12 years, and 12 patients were male (63%). All patients underwent awake craniotomy for intraoperative language testing (counting, naming, action naming) and at least partial tumor resection according to intraoperative stimulation results. Data of multimodal preoperative imaging were implemented in surgical navigation (Brain Lab, Germany). Signs of focal tumor dedifferentiation were considered in cases of partial contrast enhancement in standard MRI, increased choline in 1H MRS, and increased regional cerebral blood flow (rCBF) in perfusion weighted MRI. In patients with extended tumors with limited surgical treatment option the intended surgical goal was to remove the suspected dedifferentiated tumor part defined by preoperative multimodal MR imaging.
According to standard preoperative MRI, 8 patients had signs suspicious of focal malignant tumor dedifferentiation. In addition, 9 patients’ 1H MRS and rCBF measurements indicated focal malignant transformation. Although complete surgical resection of the tumors was deemed impossible due to intraoperative language test results, the dedifferentiated tumor part could be completely removed in all patients as proven by postoperative MRI. A temporary decrease in language function was seen in 16 patients (84%) but resolved within 1 week in 14 of them. In 13 of the patients with targeted resection, histology proved malignant transformation and therefore these patients underwent adjuvant treatment.
Multimodal preoperative MR imaging is helpful to detect focal malignant transformation missed in standard diagnostic MRI in patients with presumed low-grade gliomas. In patients in which complete resection is precluded due to infiltration of functional language areas a targeted resection of the dedifferentiated tumor part(s) is a prerequisite for adequate histological diagnosis. However, long-term follow-up assessment is necessary to show, whether or not targeted resection of malignant tumor parts and adjuvant treatment prolongs survival in these patients.
to determine the uselfulness of the combined use of motor and language tract DTI FT and intraoperative subcortical mapping for surgical removal of gliomas.
110 patients with gliomas (90 low and 20 high grade) were submitted to surgery with the aid of motor and language mapping (ISM). DTI-FT data for CST, IFO, SLF, and UNC were available at the time surgery. DTI-FT was acquired by a 3-T MR scanner with a single-shot EPI sequence (TR/TE 8986/80 msec, b = 1000 s/mm) with gradients applied along 32 non-collinear directions. Data were transferred to the neuronavigational system. Functional subcortical sites identified during subcortical mapping were correlated with fiber tracts depicted by DTI-FT. The impact of the combined use of DTI FT and subcortical mapping on duration and modalities of surgical procedures and on functional outcome of the patients was also evaluated.
In high-grade gliomas DTI-FT depicted tracts mostly at the tumor periphery; in low-grade gliomas fibers were frequently located inside the tumor mass. There was a high correlation between DTI-FT and ISM (sensitivity for CST = 95%, language tracts = 97%). For a proper reconstruction of the tracts, it was necessary to use a low FA threshold of fiber tracking algorithm and to position additional regions of interest (ROIs). The combination of DTI-FT and ISM decreased the duration of surgery, patient fatigue, and intraoperative seizures.
Our data indicate a good concordance between DTI-FT data and those obtained during subcortical mapping. When used in combination with subcortical mapping, DTI-FT offers the opportunity to quickly find the fibers associated with motor or language functions during surgery. The clinical relevance of this combined approach comes from the fact that it further enhances surgical safety maintaining a high rate of functional preservation.
For a long time, insular gliomas have been considered inoperable. Thus, few experiences of insular surgery have been reported, especially concerning WHO grade II glioma (GIIG). Here, the author details a 10 years personal consecutive experience of 52 patients operated on for a GIIG involving the insular lobe. On the basis of the functional and oncological results, both advances and limitations of this challenging surgery are discussed.
Between 1997 and 2007, 52 patients (31 men and 21 women, median age 36 years) with an insular GIIG underwent surgery. The tumor was revealed by seizures in 51 cases; 18 patients had chronic epilepsy (34%). The preoperative examination was normal in 46 patients (88%). Five patients presented with mild language disorders, and one patient with a left hemiparesis. All surgeries were conducted under cortico-subcortical stimulation and in 16 patients while awake. In all cases, the resection was stopped according to functional boundaries.
There was neither operative nor postoperative mortality. Despite a transient worsening in 30 cases (58%), all patients but two (96%) returned to baseline or better. Furthermore, 14 of the 18 patients (78%) who presented with preoperative chronic epilepsy had a relief of seizures. On control MRI, 77% of resections were total or subtotal. Pathological examination revealed a WHO grade II glioma in all cases. Ten patients underwent a second or third surgery, with no additional deficit. The tumor was still a GIIG in 7 patients, whereas anaplastic transformation was diagnosed in 3 cases. A complementary treatment was given in 20 patients (chemotherapy alone in 9 cases; chemotherapy and radiotherapy in 11 cases). Forty-three patients (82%) are still alive with a median follow-up of 48 months (from 3 to 121 months).
This is the largest ever reported experience with insular GIIG surgery. The better knowledge of the insular pathophysiology and the use of intraoperative functional mapping enable minimizing the risk of permanent deficit (and even improving the quality of life) while increasing the extent of resection and thus the impact on the natural history. Therefore, surgical removal has to systematically be considered for insular GIIG. However, this surgery remains challenging, especially within the anterior perforating substance and the posterior part of the (dominant) insula. Repeated operations can be suggested when the first resection was not complete.
The benefit of cytoreductive surgery for glioblastoma multiforme (GBM) has not been demonstrated conclusively and selection bias in past series has been observed. The ALA study investigated the influence of fluorescence-guided resections on outcome, generating an extensive database on GBM patients with a high frequency of complete resections. With the present analysis we evaluated whether the RTOG recursive partitioning analysis (RPA) would predict survival of ALA study patients and whether there was any detectable benefit from extensive resections depending on RPA class.
243 per protocol patients with newly diagnosed GBM were operated either with or without ALA and treated by radiotherapy. Early postoperative MRI was obtained in all patients. Patients were allocated into RTOG-RPA classes III, IV, and VI based on age, Karnofsky Performance Status, neurological condition, and mental status (as derived from the National Institute of Health stroke score).
Overall survival was different among RPA classes III, IV, and V, with median survival times of 17.8, 14.7, and 10.7 months, respectively, and 2-year survival rates of 26%, 12%, and 7%, respectively (p = 0.0007). When stratified for complete vs. incomplete resections of contrast-enhancing tumor, survival for patients with complete resections was longer in RPA classes IV and V (17.7 vs. 12.9, p = 0.0015, and 13.7 vs. 10.4, p = 0.0398; 2-year rates: 21.0 vs. 4.4% and 11.1 vs. 2.6%, respectively). In the small subgroup of RPA class III patients, differences were 19.3 vs 16.3 months (p = 0.14).
Survival of patients from the ALA study is correctly predicted by the RTOG-RPA classes. Differences in survival depending on resection status, especially in RPA classes IV and V, strongly support a causal influence of resection on survival.
For the patients with glioblastoma multiforme (GBM) involving pyramidal tracts, maximal resections are difficult to accomplish preserving their motor function. We use tractography-integrated functional neuronavigation and 5-aminolevulinic acid (5-ALA) fluorescence-guided resection following the removal of GBM involving pyramidal tracts. In this study we analyze the postoperative motor function and extent of resection in a series of patients who underwent surgery in our department.
Eight patients with GBM involving pyramidal tracts received radical surgery. To preserve pyramidal tracts, we have used the tractography-integrated functional neuronavigation since 2006. We also developed a functional neuronavigation-guided fence-post procedure in 2007 to resolve the problem of brain shift, a disadvantage of the existing neuronavigation system. And we have attained precise resection on the tumor using 5-ALA fluorescence navigation. Intraoperatively, tumor fluoresence was visualized using a modified operating microscope. All fluorescing tumor tissue was resected.
Motor function was preserved after appropriated tumor resection in all cases. Postoperatively, improvement of motor weakness was observed in six patients, whereas transient mild motor weakness occurred in two patients. Gross total removal was accomplished for four patients, subtotal removal was accomplished for two patients, and partial removal was accomplished for two patients.
Combined use of tractography-integrated functional neuronavigation and 5-ALA fluorescence-guided resection contributes to maximal safe resection of GBM with pyramidal tract involvement.
Gross total resection is increasingly recognized as an important first step and prognostic factor in the treatment of glioblastoma. In spite of this, most published papers lack an objective measurement of residual tumor after surgery. The removal of the entire MRI enhancing lesion is accepted as the gold standard of gross total resection. When this measurement of residual tumor by post-op MRI is done, the frequency of verified total resection is quite low, between 18%–49%. Fluorescence guided resection of glioblastoma using 5 aminolevulinic acid (5-ala) is a new technique developed by Dr. Stummer. The European Medical Agency approved the marketing of 5-ala for this indication under the name Gliolan in January 2008. In an unusual agreement, the agency approved it to be used only by neurosurgeons who have done a specific training course.
We report our initial experience with 10 glioblastoma cases. The cases were consecutive patients in which surgery was indicated. There were three recurrent cases. We did volumetric MRI measurement pre- and postoperatively. For each case we took 6–8 biopsies of the borders of the lesion and the normal tissue, as shown by the fluorescence.
In all our cases, we could see under the fluorescence three different zones: pure tumor tissue in bright red; infiltrating tumor cells in the border of the lesion with the color gradually less intense, becoming shades of pink; and normal tissue in blue. The correlation with the histology was excellent, with 100% specificity in non-recurrent cases, 80% in recurrences, and 94% sensibility. Interestingly, the pink areas always corresponded with tumor cells infiltrating brain tissue, never with solid tumor. We achieved 100% resection as measured by MRI in 7 cases, while the mean volume resected over the 10 cases was 98.9%. Mean preoperative volume was 52 cc. There was no new neurological morbidity in the series.
The fluorescence produced by 5-ala is a very efficient and safe method of achieving maximal safe resection of glioblastoma. This method allows the surgeon to see the solid tumor as well as the infiltrating border, which has a high tumor cell density. It could achieve greater resection than MRI guided surgery. With this technique it’s possible to obtain separate samples of the infiltrating border of the tumor during the surgery, and we think this could be useful for specific research on the diffuse part of the tumor.
Gliomas are aggressive brain tumors that actively suppress anti-tumor immune responses. A key event for glioma immune escape is the intratumoral accumulation of CD4+FoxP3+ regulatory T cells, which actively suppress local CD4 and CD8 effector T cell responses. In mice, we attempted to restore local T cell responses by intratumoral injection of various toll-like receptor (TLR) ligands.
Mice bearing intracerebral GL261 gliomas received a single intratumoral injection of various TLR ligands five days after tumor challenge and were monitored for survival. T cells were isolated at different time points from the tumor and cervical lymph nodes or spleen for functional characterization. In addition, direct effects of various TLR ligands on GL261 glioma cells were determined in vitro.
A single intratumoral injection of CpG-oligonucleotides (CpG-ODN, TLR9) most effectively inhibited glioma growth in vivo and cured 80% of glioma-bearing mice. Intratumoral injection of Pam3Cys-SK4 (TLR1/2) or R848 (TLR7) also produced a significant survival benefit, whereas Poly I:C (TLR3) or purified LPS (TLR4) stimulation alone was not effective. Applying CpG-ODN, potent CD4 and CD8 T-cell responses were induced that protected cured mice from a subsequent tumor challenge without further addition of CpG-ODN. In addition, the entry of CD4+Foxp3+ regulatory T cells into the tumor was strongly counteracted by IFN-γ secreting CD4+ and CD8+ effector T cells upon CPG-ODN treatment. In vitro, CpG-ODN appeared to inhibit GL261 glioma cell proliferation in a cell-type specific, but CpG-motif independent manner. Studies using TLR9+/+ wildtype and TLR9−/− knockout mice revealed that the efficacy of local CpG-ODN treatment in vivo is mainly immune-mediated.
In summary, this study underlines the potency of local TLR treatment in anti-glioma therapy and demonstrates that local CpG-ODN treatment most effectively restores anti-tumor immunity in a therapeutic murine glioma model.
A major shortcoming of traditional mouse models based on xenografted human glioblastoma cell lines is that tumor cells do not invade. Another deficit is that genetic alterations, such as amplification of the epidermal growth factor receptor (EGFR), are typically not maintained in glioma cell lines and xenografts derived thereof. These models are therefore of limited value for preclinical therapeutic studies. We established a highly invasive orthotopic nude mouse model to evaluate the effects of monoclonal antibodies against EGFR (cetuximab) and vascular endothelial growth factor receptor-2 (VEGFR-2, antibody DC101).
Freshly resected glioblastoma tissue was minced and briefly cultured as spheroids in vitro. Spheroids were stereotactically injected into the striatum of nude mice. Animals were treated for 4 weeks with either interstitial infusion of cetuximab or intraperitoneal injections of DC101. Tumor extension was measured using image analysis on H&E-stained serial sections, defining 36 landmark points at 6 different coronal levels with 6 different areas each (e.g., striatum, corpus callosum, thalamus).
Highly invasive xenografts were obtained from 9 different glioblastomas. Of 7 different xenograft-cases treated with cetuximab, 3 responded to treatment with significant tumor growth inhibition, whereas 4 did not. All responsive tumors were derived from glioblastomas exhibiting EGFR gene amplification as well as expression of the truncated EGFRvIII variant. EGFR amplification was maintained in mouse xenografts as determined by FISH analysis. The proportion of apoptotic cells was increased in responding tumors, whereas the fraction of proliferating cells was decreased. All non-responsive tumors lacked EGFR amplification and EGFRvIII expression. None of 4 xenograft cases treated with DC101 responded to treatment, and quantification of intratumoral blood vessels showed that the diffusely invading tumors grew largely independent of angiogenesis.
This is the first study showing that inhibition of invasive glioblastoma growth can be achieved in vivo using interstitial delivery of an anti-EGFR antibody. Importantly, tumor responsiveness depended on the presence of amplified and/or mutated EGFR. In contrast, anti-angiogenic treatment was not effective against the diffusely invading tumors.
Tumor stem cells (TSCs) are responsible for tumor initiation and therapy resistance in a variety of cancers, including leukemia, carcinomas of the breast and colon as well as several types of brain tumors. Differentiation of TSCs has been discussed as a possible approach to eradicate the tumor-driving cell population disrooting the actively proliferating tumor bulk. Since all-trans retinoic acid (ATRA) is known as a modulator of differentiation and proliferation, we sought to elucidate whether ATRA induces differentiation of glioblastoma-derived TSCs, so-called brain tumor stem cells (BTSCs), and if tumor-relevant properties of these cells are affected by differentiation.
BTSC lines (n = 3) with high CD133 content (68%–93%) were treated with ATRA-containing medium. Change in proliferation and CD133 content was monitored by BrDU incorporation assay and FACS analysis, respectively. Impact of differentiation on the angiogenic capacity of BTSCs was measured by quantification of angiogenic cytokines and assessed in a HUVEC-based tube formation assay. Potential effects on BTSC invasiveness were studied in a 3D-collagen invasion model. Finally, we studied whether in vitro effects could be confirmed in vivo using a NOD/SCID-mouse xenograft model.
We present evidence that BTSCs exposed to ATRA increase their proliferative activity and simultaneously lower the expression of stem cell–related antigen CD133 by up to 75% in favor of incremented linage markers GFAP, MBP, and beta3-tubulin. Furthermore, we report on significantly reduced VEGF and bFGF secretion by 70%–82% and 95%–99%, respectively, as well as significantly lowered tube formation by up to 55% following differentiation. Additionally, we show that differentiation elicits strong anti-invasive effects reducing collagen invasion by up to 46% and that these effects are associated with a marked downregulation of invasion-related MMP2 protein (up to sevenfold). Finally, we report that xenografted tumors of differentiated BTSCs are significantly smaller (only 15% of BTSC tumor volume) and less invasive than undifferentiated BTSC tumor xenografts. Correspondingly, animals bearing differentiated cells show both significantly better PFS and OS than mice with BTSC xenografts (p<0.016).
Altogether, these results highlight the potential of differentiation treatment to target the tumor-driving compartment in glioblastoma and point out a potential therapeutic value in the eradication of TSCs.
Angiogenesis and invasion are strictly related phenomena involved in glioma genesis and progression. In vivo models based on the intracerebral injection of glioma cancer stem cells are well suited for studying the relationship between these two phenomena, because they both possess in these tumors features similar to those observed in the human setting. In the first part of this work we analyzed the changes in tumor vasculature and cell infiltration during the various phases of tumor progression in a glioma nude mice model generated by the injection of glioma cancer stem cells isolated by a case of human anaplastic astrocytoma. Animals were sacrificed at various time points after glioma cancer stem cell injection. Sections were evaluated for vascular parameters (vessel number, diameter, length, basal membrane, pericytic coverage, tumor hypoxia) and pattern of tumor cell infiltration. At early stage a small tumor mass, consisting of an agglomerate of dispersed elongated tumor cells, diffusely infiltrating the surrounding brain parenchyma, was evident. Most of the cells were co-opting preexisting tumor vasculature; in the tumor mass actively growing capillaries were present. At intermediate stage, larger tumors composed of a tumor core surrounded by a large area of infiltration were documented, in the context of which several foci of anaplasia were evident. Angiogenesis was active, both in the tumor mass and particularly in the foci. At late stage, the tumors were large, densily and extensively infiltrating the brain parenchyma. Necrosis was also present. Angiogenesis was active and heterogeneous, with glomeruloid structures, mature and actively growing capillaries, co-opting vessels. This model strictly resembles human glioma, in terms of tumor cell and vasculature heterogeneity and tumor cell infiltration. It represents a good model of progression, which occurs at both the tumor cell and vasculature levels. Angiogenesis is striclty associated with invasion. In the second part of the work, we investigated if angiogenic treatment was able to inhibit the progression of the tumor toward more aggressive phenotype. Angiogenic inhibitors were administered alone or in combination to animals with early or intermediate stage tumors. Mice were sacrificed at various time points, and the infiltrative and angiogenic pattern of the tumors evaluated as previously described. In both settings, antiangiogenic treament reduced angiogenesis and tumor cell infiltration. The effect was more prominent at early stage, where tumor progression was also inhibited.
Bone-marrow derived mesenchymal stem cells (MSCs) have emerged as a promising cellular vehicle for delivery of anti-tumoral substances in glioma therapy. However, the capacity of MSCs to specifically migrate within an invasive GBM-like model has not been quantatively analyzed. Here we quantified the extent to which MSCs migrate to invasive tumor extensions and to distantly located tumor micro satellites. For this purpose, we isolated MSCs from the bone-marrow of adult Fischer344 rats, retrovirally labeled MSCs to express GFP, and grafted GFP+ MSCs into the highly invasive N29 rat brain tumor in adult Fischer344 rats. MSC grafting was done intratumorally into established tumors in order to resemble a clinical scenario where MSCs are grafted following partial tumor resection. We counted the numbers of tumor extensions and distantly located tumor micro satellites that were infiltrated by GFP+ MSCs at three time-points following grafting. The percentage of tumor extensions infiltrated by MSCs, 5, 10, and 15 days after MSC grafting, was 59±14 %, 76±3 % and 70±11 %, respectively, and the percentage of distantly located tumor micro satellites infiltrated by MSCs were 20±7 %, 26±4%, 34±6%, respectively. Importantly, MSC migration was largely restricted to tumor tissue and minimal numbers of MSCs were found in the normal brain parenchyma. These findings were confirmed by FISH analysis of male MSCs grafted into female hosts. Thus, we demonstrate proof-of-principle that one single MSC injection results in infiltration of the majority of invasive tumor extensions and of a significant fraction of tumor microsatellites. Multiple intratumoral injections at different coordinates will likely increase the numbers of tumor satellites infiltrated by grafted MSCs.
Varizella Zoster Virus (VZV) is an oncolytic virus solely replicating by cell-cell fusion processes. Recently, it has been shown that VZV is suitable for tumor therapy in malignant melanomas. Focusing on a future oncolytic therapy for malignant gliomas, the aim of the following study was to asses permissive VZV replication in glioma cell lines and primary glioblastoma cell cultures in vitro.
Three gliomas cell lines (U87, U252, U373) and six primary cell cultures from human glioblastoma specimens have been tested for permissive cellular VZV replication in vitro. An established melanoma in vitro model has been used as positive control, and three normal brain cell cultures (NBC) have been used as negative controls. Since bone marrow derived human mesenchymal stem cells (hMSC) are regarded as potential carriers of oncolytic therapy, three hMSC cultures were included in this study. Infection rates have been qualitatively and quantitatively evaluated by cellular expression of early (e.g., ORF63) and late (e.g., ORF68) viral proteins using immunocytochemistry. Oncolytic potency has been evaluated by cell counting and comparative morphometric culture analysis in simultaneously VZV infected glioma cells, hMSCs and normal brain derived cells in a time dependent protocol.
Except for U87 cell line, permissive VZV replication has been verified in all glioma cell cultures tested including all primary cultures (8/9). Efficient total oncolytic cell death was detected within 6±3 days after VZV infection. Interestingly, plaque formation and syncytial fusion, which are characteristic for melanoma VZV replication, rarely occurred within glioma cell lines. Cell cultures harvested from normal brain specimens also exhibited ORF63/ORF68 expression but showed delayed single cell death (16±4 days after VZV infection). Interestingly, besides established VZV infected melanoma cells, permissive VZV replicating hMSCs have also been highly suitable for cellular vehicles in glioma cell infection.
Wild-type VZV efficiently replicates within glioma cell cultures causing broad oncolytic cell death in vitro. The underlying mechanisms causing infections devoid of characteristic plaque formation are still unknown. Since hMSCs are known to have an intrinsic tumor cell tropism, autologous VZV-replicating hMSCs could be applicable as cellular vehicles for local oncolytic glioma cell infection. Further investigations have to address sufficient glioma cell selectivity in VZV replication with respect to the normal brain parenchyma.
An increased copy number and mutation of the epidermal growth factor receptor (EGFR) gene are frequently found in high-grade gliomas (HGG). We investigated the activity of the EGFR-blocking monoclonal antibody cetuximab (Erbitux) for the treatment of patients (pts) with recurrent HGG following surgery, RT, and chemotherapy.
Eligible adult pts were treated with cetuximab (400 mg/m2 2 h i.v. on d1 and weekly 250 mg/m2 1 h i.v. thereafter). Pts were stratified in 2 treatment arms according to the amplification status of the EGFR gene of their high-grade glioma (determined by fluorescence in situ hybridization on archival tumor material).
Between May 2005 and December 2007 a total of 55 pts with performance status 0–2 initiated treatment with cetuximab (28 pts with and 27 pts without an increased EGFR copy number, 17F/38M, median age 53 years [range 32–73]). Cetuximab was generally well tolerated. During a total of 765 treatment weeks the most frequent treatment-related adverse events were: skin toxicity (grade 2, n = 12; grade 3, n = 5), thrombocytopenia (grade 2, n = 1; grade 3, n = 2), confusion/diminished consciousness (grade 3, n = 4 pt), lymphopenia (grade 4, n = 1), infusion related allergic reaction (grade 2, n = 1), intratumoral hemorrhage (grade 2, n = 1), diarrhea (grade 2, n = 1), and fatigue (grade 2; n = 3). A dose reduction of cetuximab (to 200 mg/m2 weekly) was necessary in 3 pts. After a median follow-up of 17 months, 8 pts are alive (4 are still receiving cetuximab) and 47 pts have died. The disease control rate (DCR) was 36% (3 pts [5.5%] had a PR and 16 pts [29.1%] had SD); 36 pts (65.4%) had PD. The median PFS was 1.9 months (95% CI 1.6–2.2); median OS was 5.0 months (95% CI 4.4–5.7). The 6-month PFS and OS rates were 8.7 % and 35%, respectively. No significant correlation was found between EGFR amplification and DCR or survival. Whereas PFS was <5 months in 49/54 of pts (91%) with a follow-up of >5 months, a subgroup of 5 pts (9%) have a PFS of >9 months (range 9.5 to >18.5). This subgroup consisted of 4 pts with de novo glioblastoma (3 without and 1 with EGFR amplification) and 1 pt with an anaplastic astrocytoma with EGFR amplification.
Cetuximab as a single agent was safe and well tolerated in this population of pretreated patients with recurrent HGG. Durable disease control was observed in a small subgroup of patients but was not predicted by EGFR amplification at initial diagnosis.
Anaplastic astrocytomas (AA) (WHO grade III) constitute less than 10% of all gliomas. The prognosis for patients with these tumors is poor; even with the most aggressive therapies, including surgery and adjuvant radiation therapy and chemotherapy, the 5-year survival for newly diagnosed patients is currently less than 30%. For recurrent AA the overall survival is less than 2 years. The purpose of this study is to evaluate the outcome of adults with recurrent AA treated with ANP in an FDA-monitored phase II clinical trial. ANP affects multiple targets, and its components have different mechanisms of action. A10 interferes with signaling in the AKT2 and MYCC pathways, blocks expression of TGFB1, activates the PTEN and MAD tumor suppressor genes, and normalizes nuclear transport by decreasing the expression of RANBP1, which may restore the activity of the mutated INI protein. AS2-1 interferes with signal transmission in the RAS and BCL2 pathways and activates expression of the tumor suppressors TP53 and p21. Twenty assessable adults, all diagnosed with AA, whose ages ranged from 20 to 51 years (median age 41 years), were involved in this study. The tumor recurred in all patients after radiation therapy, and 55% of patients received additional chemotherapy before the recurrence. ANP was administered intravenously daily through a subclavian venous catheter via a double-channel infusion pump. The median duration of treatment was 6.5 months and the median of average dosages of A10 was 5.8 g/kg/day and of AS2-1 was 0.24 g/kg/day. The treatment was well tolerated with only one case of a serious toxicity of hypernatremia. Complete response (CR) was achieved in 15%, partial response (PR) in 10%, stable disease (SD) in 45%, and progressive disease in 30% of patients. Overall survival (OS) at 2 years was 35%. The median of progression-free survival based on K-M was 10.3 months. These results compare favorably with the outcome of standard treatment. In conclusion, ANP is well tolerated and provides encouraging results in the treatment of recurrent AA and merits further investigation in a randomized phase III trial in comparison to standard treatment.
Nitrosourea-based chemotherapy (NBC) is a widely used treatment for patients (pts) with recurrent high-grade glioma (HGG) who have not received prior chemotherapy. Although temozolomide (TMZ) is increasingly used, it has never been directly compared with NBC. We report results of a Cancer Research UK–funded randomized trial of PCV vs. TMZ and comparison of two dosing schedules of TMZ at first recurrence of HGG.
Chemo-naïve pts with radiologically confirmed, recurrent HGG, considered fit for chemotherapy were randomized through the MRC Clinical Trials Unit in the ratio 2:1:1 to PCV, TMZ-5, and TMZ-21. The TMZ schedules were 200mg/m2 for 5 days (TMZ-5) or 100 mg/m2 for 21 days (TMZ-21), both schedules repeated every 28 days for up to 9 cycles or until progression. PCV comprised procarbazine 100 mg/m2 p.o. days 1–10, CCNU 100 mg/m2 p.o. day 1, and vincristine 1.5 mg/m2 (max 2 mg) i.v. day 1; cycles repeated every 6 weeks for up to 6 cycles or until progression. The primary (1°) comparison is PCV vs. TMZ (either schedule); the 1° outcome is survival, and the trial is powered to detect a 2–3 month increase in median survival with TMZ corresponding to hazard ratios (HR) of 0.75 and 0.67, respectively. We require 380 deaths to detect HRs of 0.67 with 95% power and 0.75 with 80% power, α = 0.05 (2-sided). To observe 380 deaths, a maximum of 500 pts randomized over 4 years was anticipated, with a planned subgroup analysis in 300 glioblastoma multiforme (GBM) pts. Progression-free survival (PFS) at 3 months is the 1° outcome for the comparison of the two TMZ schedules. 220–250 pts randomized with respect to dose would provide 80% power to detect absolute differences of 15%–20% in PFS rates at 12 weeks.
Accrual began in July 2003 and closed on January 31, 2008. 447 pts were randomized. Approximately 75% of pts have GBM tumors. We anticipate 380 deaths by July 2008, when the majority of patients will have completed treatment, and we will conduct the primary analysis at this time. The data to be presented will therefore include treatment compliance and toxicity data, the intent-to-treat logrank analysis of survival for the TMZ and PCV arms in the whole trial population and the GBM subgroup and the PFS logrank analysis for the PCV vs. TMZ comparison, and for the two TMZ schedules individually.
ENZ, an oral serine/threonine kinase inhibitor, targets PKC and AKT pathways to induce tumor cell apoptosis and suppress proliferation and angiogenesis. This phase III, multicenter, open-label study compared efficacy and safety of ENZ vs. CCNU in patients (pts) with recurrent, intracranial GBM (WHO grade IV).
Pts were stratified by age, KPS, and disease recurrence, and randomized 2:1 to receive 6-week cycles of 500 mg ENZ daily (1125-mg loading dose day 1) or CCNU (100–130 mg/m2 on day 1). Treatment continued until disease progression or unacceptable toxicity occurred. Assuming a 45% improvement in progression-free survival (PFS) of ENZ over CCNU, 397 pts were to be enrolled to provide 80% power to achieve statistical significance at a one-sided level of 0.025. Secondary endpoints included overall survival (OS), tumor response, safety, and patient-reported outcomes (PROs). The primary PRO analysis was time to deterioration (TtD) for the FACT-Br Trial Outcome Index (TOI; physical and functional well-being plus brain tumor–specific concerns).
Enrollment was terminated at 266 pts (ENZ = 174; CCNU = 92) after a planned interim analysis for futility. Pt characteristics were balanced between arms. Median PFS, OS, and 6-month PFS rates were not different between arms. For ENZ and CCNU, respectively: median PFS = 1.5 vs. 1.6 months, HR = 1.3 (95% CI: 1.0, 1.7), P = 0.08; median OS = 6.6 vs. 7.1 months, HR = 1.2 (95% CI: 0.9, 1.7), P = 0.25; and PFS rate: 11.1% vs. 19.0%, P = 0.13. Pts with KPS 90–100 had significantly better PFS and OS on CCNU compared with ENZ. Five (2.9%) and 4 (4.3%) pts had objective response and 67 (38.5%) and 33 (35.9%) had stable disease on ENZ and CCNU, respectively. A total of 251 pts received therapy (ENZ = 167; CCNU = 84) and were included in the safety analysis. Median duration of therapy was 42 days on both arms. Three pts remain on ENZ therapy for ≥1 year. The incidence of drug-related adverse events (AEs) was 44% on the ENZ arm and 62% on the CCNU arm. Four pts discontinued ENZ due to drug-related serious AEs (erysipelas, aortic thrombosis, cerebral hemorrhage, and convulsion). Eleven (7%) pts on ENZ died (4 due to AEs, 1 of which was drug-related). On the CCNU arm, all 4 (5%) deaths were disease related. Grade 3/4 hematological toxicities were significantly higher on CCNU (P ≤0.001); no anemia, neutropenia, or leukopenia occurred on ENZ, and only 1 pt had drug-related thrombocytopenia vs. 21 on CCNU. There were no significant differences in grade 3/4 nonhematological toxicities between arms. Baseline scores for the TOI were not statistically different between arms (P = 0.638). Median TtD was 2.3 months for both arms (HR = 1.12 [95% CI: 0.77, 1.63], P = 0.54).
ENZ had a better toxicity profile but did not have superior efficacy or PROs compared with CCNU in pts with recurrent GBM.
High-grade gliomas are very aggressive tumors, characterized by overexpression of transforming growth factor-beta 2 (TGF-beta 2). TGF-beta 2 is one of the most potent immunosuppressors that induces escape from immunosurveillance. TGF-beta 2-specific phosphorothioate antisense oligodeoxynucleotide AP 12009 is targeted to suppress this protein and is used as monotherapy in this indication in clinical trials.
The international Phase IIb study with AP 12009 had an open-label, randomized, and active-control dose-finding design. The main aim of the trial was to compare two doses of AP 12009 (10 μM and 80 μM) with standard chemotherapy (temozolomide or PCV [procarbazine/CCNU (lomustine) vincristine]) concerning overall response rate, survival, and safety. 145 patients with either recurrent or refractory anaplastic astrocytoma (AA, WHO grade III) or recurrent or refractory glioblastoma (GBM, WHO grade IV) were randomized into three groups. AP 12009 was administered as monotherapy via intratumoral, convection-enhanced delivery (CED) using an external portable pump for 6 months with up to 11 treatment cycles (1 cycle: 7 days AP 12009, 7 days isotonic saline).
134 patients (95 GBM and 39 AA) on the study received the drug: 10 μM AP 12009 (N = 40), 80 μM AP 12009 (N = 49), or standard chemotherapy (N = 45). In the entire patient population the survival results for the AP 10 μM group showed a trend for superiority to standard chemotherapy (p = 0.082). Even more encouraging results were obtained for AA patients. The overall response rate (CR+PR) and progression rate of the 10 μM AP 12009 group at 14 months were significantly superior to standard chemotherapy. The current overall survival time (in months) is 36.4 (10 μM AP 12009) and 21.7 (standard chemotherapy) (status: Feb. 2008). In both AP 12009 groups, the observed long-lasting tumor regression clearly exceeded the active treatment period. The follow-up period for the Phase IIb study has been extended.
For the 10 μM AP 12009 vs. the standard chemotherapy group, patients with recurrent or refractory AA showed a clear current survival benefit of 14.7 months. The overall response rate and progression rate at 14 months were significantly better for the 10 μM AP 12009 group than for the standard chemotherapy group (p<0.05). A multicenter, international, active-control Phase III study with 10μ M AP 12009 monotherapy in patients with recurrent or refractory AA will start in 2008.
Despite improvements obtained with frontline treatments prognosis of recurrent HGG still remains dismal. HD chemotherapy suggested a dose-effect relationship in extra CNS lymphoma and germ cell tumors. HDTMZ could be a promising way to overcome resistance of HGG to standard schedule of CT.
This phase I had as principal objective to determine the maximum tolerated dose (MTD) of HD of TMZ with PBSCS rescue in patients with recurrent HGG under 60 years. The MTD was defined as dose level at which 50% of patients treated experienced a DLT (dose-limiting toxicity) defined as grade IV hematological toxicity (lasting more than 2 weeks for neutrophils or 3 weeks for platelets) or a grade III non hematological toxicity. The dose escalation was planned for eight dose levels from 300 to 650 mg/m2/day over 5 days with CSP reinfusion at day 7 according to the Modified Continual Reassessment Method 3 (MCRM). Treatment was administered for one cycle.
30 patients were eligible, but in 10 the phase I could not be performed (8 because sufficient number of PBSC could not be collected, and 2 for other reasons). Thus only 20 received HDTMZ: 16 GBM, 2 AA, and 2 OA. All but one had received previous RT, and 11 had received chemo, mainly nitrosourea. Dose level (mg/m2/day × 5 days) was as follows: 300 (1), 350 (1), 400 (6), 450 (3), 500 (3), 550 (3), and 600 (3). 18/20 had a non-DLT (84 declarations: 22% for prolonged neutropenia and 13% for prolonged thrombopenia) and/or a DLT (3 patients: one hepatic cytolysis grade III at level 400, and 2 others at level 600: see infra). The MTD was reached at 600 mg/m2/day × 5 days: 1 gonalgia grade III and 1 prolonged but partially reversible comatose state. There were 3 partial responses (1 at 550 and 2 at 600 mg/m2), and 8 stable diseases (none at 350 or 600).
This final analysis demonstrated that HD of TMZ with CSP reinfusion is feasible but MTD is reached at 600 mg/m2/day × 5 days. The recommended dose is thus 550 mg/m2/day × 5 days. The association with other drugs that have no cross-resistance may then be explored further.
Pilocytic astrocytomas comprise the most frequent brain tumors in childhood. However, the molecular mechanisms of pathogenesis and tumor recurrence are still poorly understood. In the present study, 66 pediatric astrocytomas of World Health Organization (WHO) malignancy grades I and II were investigated for DNA copy-number aberrations by using array-based comparative genomic hybridization (array-CGH). The most frequent genomic aberration in this tumor series was a circumscribed duplication of the BRAF locus at 7q34 present in 30/66 (45%) of cases with a predominance in pilocytic astrocytomas. Screening the same tumor sample for activating mutations of the BRAF gene, such mutations were identified in 4/66 (6%) tumors, exclusively affecting cases without BRAF duplication, indicating different mechanisms for BRAF activation. Tumors with BRAF duplication showed significantly increased mRNA levels of BRAF and CCND1 as compared to tumors without duplication. Proliferation of cultured tumor cells derived from low-grade gliomas was effectively blocked by pharmacological inhibition of MEK1/2, the immediate downstream target of BRAF as well as by stable knockdown of the BRAF gene using lentivirus-mediated transduction of BRAF-specific shRNAs. Cell cycle analysis revealed a G2/M arrest in cells treated either with MEK1/2 inhibitors or shRNAs targeting the BRAF gene. Taken together, our findings implicate aberrant activation of the Mitogen-Activated Protein Kinase (MAPK) pathway due to gene duplication or activating mutation of BRAF as a common molecular pathomechanism in low-grade astrocytomas and provide a promising novel target for future treatment strategies.
O6-methylguanine-DNA methyltransferase (MGMT) is a DNA repair protein that specifically removes mutagenic, carcinogenic, and cytotoxic O6-alkylguanine DNA adducts induced by alkylating agents like nitrosureas. Recent data suggest that loss of MGMT expression due to promoter hypermethylation may occur in the pathway leading to secondary glioblastomas.
Using methylation-specific PCR (MSP) we investigated the inactivation of the DNA-repair gene MGMT by promoter hypermethylation in 54 low-grade diffuse astrocytomas (grade II WHO) obtained from patients who underwent surgery in our institution.
Methylation of the MGMT promoter was detected in 22 of 54 tumors (meth+ 40.7%). For all patients, median PFS was 51 months and median OS 141 months. During follow-up, tumor recurred in 34 patients (18 meth+ 82% and 16 meth– 50%), while death occurred in 14 cases (9 meth+ 41% and 5 meth– 15%). Both the risk of recurrence and the risk of death was statistically significantly higher in tumors with methylation of the MGMT promoter (p = 0.02 and p = 0.03, respectively, Fisher’s exact test). Furthermore, at univariate analysis, median progression-free survival(PFS) and overall survival (OS) were shorter in tumors with methylation of the MGMT promoter (31 vs. 71 months; log-rank test, p = 0.01) and (68 vs. not reached months; log-rank test, p = 0.03). No association between age, histological subtype, treatments after surgery and OS or PFS were found. Of the 34 patients with progressive disease, 22 had a second surgery: in most of them (17 out of 22) the second diagnosis was high-grade glioma (7 glioblastomas, 5 anaplastic astrocytomas, and 5 anaplastic oligoastrocytomas). Three patients also had a third surgery: the last diagnosis was anapalstic astrocytomas in 1 and glioblastoma in 2. Malignant PFS was shorter in tumors with methylation of the MGMT promoter (40 vs. n.r. months; log-rank test, p = 0.052). Furthermore, during the follow-up in 9 patients the tumor changed from low grade to glioblastomas: the percentage of tumors that became glioblastomas (36% vs. 14%) was higher in the meth+ group and also the time occurred was shorter (81 vs. n.r.). Both camparisons did not reach statistical significance due to the small number of cases.
The findings indicate that in low-grade astrocytomas MGMT methylation is associated with tumor recurrence and is predictive of progression to a more malignant phenotype.
Protracted administration of low doses of TMZ offers potential advantages over the standard TMZ schedule, leading to increased MGMT depletion, which may enhance the antitumor activity of alkylating agents. Few data are available on the efficacy and tolerability of protracted low-dose TMZ in patients with low-grade gliomas.
Since September 2004 we enrolled in an ongoing multicenter phase II study 31 patients (25 evaluable for response), with a median age of 41 yrs and a median KPS of 90. Mild enhancement on MRI was present in 36% of patients. Pathological diagnosis according to WHO (2007) was oligodendroglioma grade II in 15 and oligoastrocytoma grade II in 10. TMZ was administered orally at 150 mg/m2/day, days 1–7 and 15–21 every 4 weeks, up to 18 cycles or tumor progression or unacceptable toxicity. Patients received prophylaxis against Pneumocystis carinii when lymphocytes <500 mm3. Patients underwent clinical and MRI assessment every 1 and 3 months, respectively, and monitoring by PET with methionine was performed in selected centers. The primary endpoint was the response rate on MRI, and secondary endpoints were the progression free-survival (PFS) at 6 and 12 months, median PFS, quality of life, and toxicity. The analysis of 1p/19q status by FISH and MGMT promoter methylation by PCR were mandatory.
The overall response rate (PR+MR) was 52% (13/25) with PR 32% and MR 20%, and disease control rate (PR+MR+SD) was 92%. Median time to maximum radiographic response was 6 months. Clinically, 65% (11/17) of patients improved, particularly those with uncontrolled seizures. 20/25 patients are still free from tumor progression with a median follow-up of 14 months (range 4–39). Combined 1p/19q or isolated 1p deletion was not associated with response on FLAIR images, whereas there was a trend toward an association between response of the enhancing tumor and 1p/19q codeletion. Grade 3 lymphopenia was observed in 45% of patients.
TMZ 1 week on/1 week off, as initial treatment for progressive low-grade oligodendroglial tumors, seems more efficacious than standard dose TMZ and is relatively safe. When patients develop cumulative lymphopenia, a conversion to standard dosing is feasible. Updated results, correlation between MGMT promoter methylation, 1p/19q loss and response, and monitoring with PET will be presented.
To investigate the prognostic value of growth rate in adult patients with low-grade gliomas (LGG).
An observational cohort study was conducted on 50 adult patients with LGG. Patients received no treatment except anti-epileptic drugs until malignant transformation was suspected. Two magnetic resonance imaging (MRI) studies, approximately 6 months apart, were used to determine growth rate based upon calculation of tumor volume on FLAIR MRI. Primary endpoints were death from any cause or clinical/radiological transformation of the tumor.
Growth rates measured over a 6-month period from study entry were highly variable between patients (range: −4.9 to 38.5 ml/yr; mean = 16.8 ml/yr; median = 17.5 ml/year). Patients with faster growing tumors reached the endpoint earlier (log-rank test: Chi21 = 13.4; p = 0.0002; relative risk of endpoint = 3.5; 95% confidence interval: 1.6–7.8). Cox regression analysis showed that faster growth rate was an independent prognostic factor (allowing for confounding variables including tumor size, patient age and gender, tumor histology, and whether the tumor crossed the midline). For every 10 ml/year increase in growth rate, the increased hazard of the endpoint (risk of malignant transformation or death) was 2.41 (95% CI: 1.30–4.46; p = 0.004).
Growth rate is a readily measured predictor of early transformation in LGG. The data suggest that LGG growth is best considered a continuous variable, with higher growth rates associated with earlier transformation.
Astrocytomas are the most frequent brain tumors in children, mainly low-grade neoplasias, but represent a very heterogeneous group of tumors.
To review the clinical experience with low-grade astrocytomas (LGA), their histopathological and clinical features, therapeutic approach, and prognostic features.
A retrospective study of children (0–18 y), diagnosed with LGA within a 30-year period (1974–2004) was performed. Histology variables included histological pattern, tumor classification (edited WHO version), and immunohistochemistry (IHC) in a tissue microarray for GFAP, WT1, p53, and ki67 expression.
A total of 143 patients managed at our institution for LGA were included. Average age at diagnosis was 7.4 y (SD 4.3). Seventy-three (51%) patients were male. Neurofibromatosis was documented in 11 cases (8.1%). Initial symptoms were headache (35.8%), seizures (20.9%), motor deficit (16.4%), visual (10.4%), and ataxia (9.7%). 84 tumors (58.7%) were infratentorial. Primary sites were cerebellum (41.2%), hemispheres (21.3), optic pathway (14%), brainstem (14%), spinal tumors (6.6), or mesencephalic (2.9%). Fifty tumors (39.7%) sited at midline and 30% each were right and left. Hydrocephalus was present in 61 patients (42.7%) and a mixed solid-cystic pattern was seen in 80% of cases. Only 4 patients had metastases (2.9%). Surgery was complete in 60 patients (45.1%). 27 patients had two or more resections (18.9%). A derivative procedure for hydrocephalus was required in 37 patients (27.4%), 31 of them currently device-dependent (23.1%). Pathological review classified 93 tumors as WHO grade-9 (73.8%), and 33 as grade II tumors (13 as fibrillary and 3 as pleomorhic xantoastrocytomas). IHQ for p53 was negative, save for 5 patients, all grade II tumors; Ki67 expression was >5% in 92.6% of cases, and WT1 (nuclear) 59.7%. No significative differences in survival were found among histological subtypes or IHQ pattern, but a trend for higher mortality in WT1 positive tumors was appreciated. Neurological complications were the most frequent in the immediate postoperative period (41 [31.3%]). Adjuvant therapy was given in 53 (37%) cases: radiation therapy in 36 patients (25%) and chemotherapy in 22 patients (15.3%). Currently, 15 (11.3%) patients have died of disease, 67 (50.4%) patients are alive with no evidence of disease, and 51 (38.3%) patients are alive with disease (median follow-up of 200 months). An age younger than 12 months was a prognosis factor, as 4 of 5 infants have died. Long-term secondary effects have been documented in 54 (37.8%) patients, 17 with more than one sequel. Forty-nine (35.5%) cases are lost to follow-up when discharged to adult care.
LGA in infants have an unfavorable prognosis. We are not aware of the current status of the eldest patients; long-term follow-up is necessary, but unfeasible in the actual health setting.
The optimal management for patients with primary gliomatosis cerebri (GC) is still unknown, and some studies suggest that temozolomide could be useful as an upfront treatment.
Between 2000 and 2006, 53 patients with biopsy-proven primary GC were treated with temozolomide (200 mg/m2/day for 5 days every 4 weeks until tumor progression or unacceptable toxicity) either upfront (31) or at relapse after prior radiotherapy/chemotherapy with nitrosoureas (22). Histological diagnoses were as follows: astrocytoma in 17 patients, oligodendroglioma in 6, oligoastrocytoma in 4, gemistocytic astrocytoma in 2, anaplastic astrocytoma in 9, anaplastic oligoastrocytoma in 2, anaplastic glioma in 6, glioblastoma (focal area) in 1, and glial proliferation consistent with GC in 6. Median age was 51 years (14–81), with 31 males and 22 females, and a median Karnofsky score of 80. Thirty-three out of 53 patients (62%) had minimal contrast enhancement on MRI. Response was evaluated based on measurable changes in tumor area using T2-weighted and Flair images.
Median number of cycles was 6 (range 2–26). Patients receiving upfront TMZ had a response rate (PR+MR) of 29%, a time to tumor progression (TTP) of 9.9 mos (1–69) and a progression-free survival at 6 months (PFS6) of 61%. Patients receiving TMZ at relapse after previous treatments had a response rate (PR+MR) of 23%, a TTP of 5.3 mos, and a PFS6 of 50%. Neurologic improvement (more commonly reduction of seizures) was identical (31%) in both groups. Among the 25 patients for whom 1p/19q status was available, response to TMZ was significantly associated with 1p–19q codeletion (p<0.05);conversely, response was not associated with histologic type or tumor grade. Overall survival (from diagnosis) was 14.9 months. After multivariate analysis (Cox regression model), Karnofsky status and 1p/19q codeletion retained an independent prognostic significance. Toxicity of TMZ was mild.
About one third of patients with primary gliomatosis cerebri benefit from treatment with temozolomide, and response is higher among patients treated upfront and with 1p/19q codeletion. Dose-dense schedules of temozolomide could improve the results. A multicenter phase II study of the AINO with upfront dose-dense TMZ is ongoing to delay large-field radiation, and material for 1p/19q status and MGMT analysis is being collected.
The effectiveness of many chemotherapeutic agents for malignant brain tumors is limited by the blood-brain barrier. Approaches such as blood-brain barrier disruption (BBBD) improve delivery of these agents to the brain. This report describes the degree of disruption for BBBD procedures and the impact of a cumulative degree of disruption score (as a measure of total dose intensity) on survival in a large case series of patients with newly diagnosed primary central nervous system lymphoma (PCNSL).
A previously reported case series of 149 PCNSL patients treated at four centers with a methotrexate-based regimen in conjunction with BBBD, without up-front whole brain radiotherapy, was evaluated for degree of disruption for each procedure. Individual disruptions are rated nil, moderate, good, or excellent or intra-arterial only (i.e., no disruption performed). These degree-of-disruption data are summarized and used, along with number of disruptions, as time-dependent covariates to predict overall survival. Analyses adjust for potential confounders including performance status, age, gender, surgical resection, discontinuation due to complications, and intraocular disease.
The majority of 2,080 procedures were rated moderate (35%) or good (36%), while fewer were rated nil (11%) or excellent (10%). Mean cumulative degree of disruption score was 44 (SD 28). The correlation between degree of disruption on two consecutive days of therapy was relatively low (Spearman r = 0.39, p<0.0001). Cumulative degree of disruption score (hazard ratio 0.97, 95% CI: 0.96, 0.98) was a significant predictor of survival. Controlling for potential confounders did not substantially change the hazard estimate or its p-value.
These results support the association between degree of disruption and survival in patients with PCNSL. The low correlation between disruptions on consecutive days suggests the degree of disruption depends on more than patient factors. Future research will explore associations of degree of disruption with complications.
There is a statistically significant association between degree of disruption, total dose intensity, and survival supporting the efficacy and importance of blood-brain barrier disruption for treating PCNSL patients.
Methotrexate (MTX)-based chemotherapy followed by whole brain RT improves survival in primary CNS lymphoma (PCNSL) patients but is associated with a high risk of RT-related neurotoxicity (NT). Deferring RT does not seem to compromise survival in the elderly (>60 years), but utilizing such an approach in younger patients remains controversial. In this study we report outcomes of an innovative MTX-based regimen for PCNSL patients younger than age 60 and seek to investigate whether withholding RT is an acceptable option for selected patients.
We reviewed records of all patients <60 years with confirmed PCNSL seen in 3 French institutions from 1994 to 2003. Through prior agreement among institutions, patients were offered a standardized treatment consisting of an induction chemotherapy with MTX (3 g/m2 on days 1, 10, and 20), CCNU (40 mg/m2 on day 1), procarbazine (60 mg/m2 on days 1–7), methylprednisolone (120 mg/m2 on days 1–20) and IT MTX (15 mg), cytarabine (40 mg) and methylprednisolone (40 mg on days 1, 5, 10 and 15). Patients with a complete response (CR) proceeded to a maintenance chemotherapy consisting of 5 monthly cycles with same doses of MTX on day 1, CCNU on day 1, procarbazine days 1–7, and same IT chemotherapy on day 1; those patients received no further treatment. Patients with less than a CR to induction were treated on an individual basis and typically received consolidation with RT or high-dose chemotherapy (HDC) with stem cell rescue.
Among the 64 included patients, median age was 47 (19–60), median KPS was 70 (20–100), and 41 (64%) were men. For the entire population, median progression-free survival (PFS) was 13 months and median overall survival (OS) has not been reached (median follow-up: 54 months). Ninety percent of patients responded to induction treatment (CR: 52%; PR: 38%). Consolidation or salvage treatment with RT has been given to a total of 27 patients, and HDC to 23. For the 26 patients who completed maintenance chemotherapy and received no further treatment, median PFS was 36 months. Grades 3/4 hematotoxicity developed in 18 patients, nephrotoxicity in 2, and NT in 6 (none in the chemotherapy-only treated patients). Two toxic deaths occurred. KPS≥60 was associated with increased OS (p = 0.017).
Deferring RT in chemosensitive patients seems to decrease PFS but not OS, with minimal NT. HDC and RT are effective salvage treatments. Further studies on the upfront use of RT versus HDC in younger patients are warranted.
The identification of prognostic factors in patients with primary CNS lymphoma (PCNSL) and the validation of prognostic scores is useful for stratification of patients in randomized trials and essential for a critical comparison of treatment results of nonrandomized clinical trials. Moreover, the attribution of cancer patients to different risk groups may lead to more sophisticated and risk tailored therapeutic strategies. The prognostic factors identified for PCNSL in immunocompetent individuals as defined by the International Extranodal Lymphoma Study Group are age, performance status, elevated lactate dehydrogenase serum levels, high protein concentration of the cerebrospinal fluid, and involvement of deep structures of the brain. The objective of this study was to evaluate the prognostic impact of early tumor response assessed by magnetic resonance imaging (MRI) in immunocompetent patients with PCNSL receiving systemic and intraventricular chemotherapy.
From 09/1995 to 12/2002, 88 patients with PCNSL (median age 62 years) were enrolled in a pilot/phase II study evaluating MTX-based high-dose chemotherapy without radiotherapy (“Bonn-Protocol”). MRI scans were performed after two and after six courses of chemotherapy, and radiographic response was assessed according to the Macdonald criteria. Overall survival (OS) and time-to-treatment failure (TTF) were measured.
Patients achieving a complete radiographic response already after two courses of chemotherapy (n = 18) had a significantly longer median OS (not reached) and median TTF (not reached) than patients with complete response only after termination of treatment (n = 24) (OS: 55 months; TTF: 32 months) (p = 0.01).
In addition to the prognostic factors defined by the International Extranodal Lymphoma Study Group, early tumor response assessed by MRI already in the early stage of treatment was shown to be a highly predictive prognostic factor for both OS and TTF in patients with PCNSL and treated with combined systemic and intraventricular chemotherapy. Patients without rapid response to chemotherapy possibly should be switched to alternative treatment strategies in future.
Primary pineal tumors (PPT) are uncommon in the context of pineal gland neoplasms. Outcome and results of therapy could depend on several factors including histological subtypes, staging at diagnosis, type of surgery (S), doses of radiotherapy (RT), and administration of chemotherapy (CT). The purpose of this analysis has been to better define of outcome and possible prognostic factors in the era of modern imagery, surgery and adjuvant treatment.
This multicenter retrospective study from the RCN recruited 35 patients during a period between 1994 and 2006. Median age of patients was 36 (2–76), and the male/female ratio was 1/2. The mean size of tumors was 19 mm (9–55). Type of surgical resection was as follows: biopsy in 12 cases, partial resection in 13, and total resection in 8 (2 cases were unknown). In 15 cases postoperative imaging within the first 72 hours was achieved. In 9 cases RT doses to the tumor bed (CTV) were lower than 54 Gy, and in 25 >54 Gy. CSI was delivered in 13 patients. CT was administered in 12 patients (platin-based in 8).
At the time of diagnosis, one patient presented with intracranial metastases, and 7 patients presented with spinal cord dissemination. Histological subtypes were pineoblastoma (PNB) in 21 patients, pineocytoma (PIN) in 8 patients, and pineocytoma with intermediate differentiation (PID) in 6 patients (9 samples were reviewed by an independent pathologist). In 15 cases, immuno-histochemical analysis was available (10 PNB, 2 PIN, 3 PID). All samples showed positivity for synaptophysin, 5 samples (3 PNB, 1 PIN, 1 PID) for neuron specific enolase (NSE), and 3 samples for GFAP, cd5leu7, and ulex europeus lectin, respectively. MIB-1/Ki67 was high in 3 cases (all PNB), moderate in 2 (1 PNB, 1 PIN), and low in one PNB. Fifteen patients relapsed (14 PNB and 1 PIN). Sites of relapse were the following: inside RT volumes (16 locations), outside in 6 locations, and both in 3 locations. Overall median survival time for all the series was not reached. Median of disease-free survival was 82 months. In univariate analysis age was the only significant variable (p < 0.003) in favor of patients older than 36 y. A trend for poorer survival was observed for PNB patients. Neither surgical procedures, nor RT doses or CT influenced OS. Late side effects of RT were dementia, leukoencephalopaty, or memory loss in 7 cases, occipital ischemia in 1, G3 seizures in 2 cases, long-term leucopenia 1 case, and G4 thrombopenia in 1. Side effects of CT were G3 neutropenia in 4 cases, G2 anemia in 2, G4 pancitopenia in 1, G4 vomiting in 1, renal failure in 1 and G2 thrombopenia in another case.
(1) Combination treatments for PPT achieved good overall survival. (2) Age and probably histological subtypes influenced survival in our series. (3) And the prevalence of chronic toxicity suggests that new strategies in RT and CT could be advisable.
Medulloblastoma is a highly malignant primary brain tumor with tendency to metastasize. For evaluating metastases, Chang’s staging system takes into consideration their locations but not their radiological aspects. The aim of this study was to correlate metastasis phenotypes and their evolution to patient’s outcome.
The authors reviewed outcomes of 99 consecutive patients (median age, 4.5 y; range, 0.5–14.5) affected by disseminated medulloblastoma treated at Institute Gustave Roussy, between 1989 and 2005. Metastatic disease was assessed using Chang’s staging system: M1, CSF cytology positive for tumor cells at least 8 days after initial surgery (n = 19); M2, evidence of intracranial seeding on gadolinium-enhanced MRI scan (n = 22); M3, evidence of tumor seeding down the spinal cord on gadolinium-enhanced MRI of spine (n = 58). Metastases were also evaluated by imaging (nodular, 32 pts; linear, 13; mixed, 35; invasion CSF, 19) and site extension (local, 26 pts; extensive, 54; invasion CSF, 19). During the study period, 33 patients were treated with conventional chemotherapy followed by standard craniospinal irradiation, 15 with high-dose chemotherapy (HDCT) with stem cell rescue (ASCT) followed by local radiation, and 51 with HDCT with ASCT followed by age-adapted craniospinal irradiation. All patients received induction chemotherapy with one or two cycles of carboplatin-etoposide.
After induction chemotherapy, 28 pts were in CR and 53 in PR. At the end of chemotherapy, 41 children were in CR and 33 in PR. The median follow-up was 7.8 y (range, 2.2–14.7). At 5 y, EFS and OS were 38% and 44%, respectively. 5-year OS for patients with M1, M2, and M3 tumors were 45%, 54%, and 40%, respectively (p = ns). 5-year OS was 58% for patients with isolated metastatic disease and 33% for patients with extensive disease (p = 0.032). 5-year OS for patients with positive CSF was 45%, linear metastases 39%, nodular 58%, and mixed 58%. 5-year OS was 54% for patients achieving CR at end of chemotherapy compared to 27% for patients failing to achieve CR (p = 0.0003).
Concerning this study, Chang’s staging system does not play a predictive role for prognosis. Metastasis radiological aspects may have an impact on survival in pediatric disseminated medulloblastoma.
Atypical teratoid/rhabdoid tumors (AT/RTs) are highly malignant embryonal brain tumors. They were first described as a distinct tumor entity in 1987 and included in the WHO CNS tumor classification in 1996. Histopathologically, AT/RTs display a complex morphology of rhabdoid, primitive neuroectodermal, epithelial, and mesenchymal components which may render the differential diagnosis from other malignant brain tumors, especially embryonal tumors difficult. AT/RTs occur predominantly in children <3 years. In the literature a prevalence of 1%–2% of pediatric brain tumors has been described. However, so far epidemiological data are scarce.
We conducted a population-based study on primary malignant, high-grade (WHO grade III+IV) pediatric brain tumors for identification of AT/RTs and generation of epidemiological data on this tumor type in the pediatric population of Austria.
All 7 centers (16 departments of pathology and pediatrics) engaged in the diagnosis and management of brain tumor patients in Austria participated in a nation-wide survey. All primary malignant CNS tumors (WHO grade III+IV) occurring in children <15 years, newly diagnosed from 1996 to 2006 were collected. In cases where tumor tissue was available a central histopathological review was performed. Additionally, immunohistochemical INI1 protein expression status, which has been recently described as valuable tool to delineate AT/RTs from other malignant brain tumors, was systematically analyzed.
A total of 268 cases have been identified so far. 93% underwent surgery. Tumor tissue was available in 211 cases and 18/211 of the cases were diagnosed on review as AT/RT. Among these, 10/18 (56%) were primarily reported as AT/RTs, while the others, 8/18 (44%) were originally classified as embryonal CNS tumors (medulloblastoma, sPNET, ependymoblastoma), Ewing sarcoma, or sarcoma. This would refer to an age-standardized incidence rate of AT/RTs in children of 0.4/1,000,000 person-years. Median age of AT/RT patients at diagnosis was 1.4 years, and 72.2% occurred in children <3 years. In this age group AT/RTs constituted the most common tumor (21.3%) followed by medulloblastoma (18.0%) and sPNET (14.8%).
AT/RTs mainly occur in very young children (<3 years). In this age group they represent the most common highly malignant brain tumor type. Accurate histopathological diagnosis including INI1 immunohistochemistry is important for identification of AT/RTs.
Imaging changes in malignant brain tumor patients may represent treatment effects rather than progression of disease [deWit MC 2004]. Pathological material from such patients generally reveals regions of bland acellular necrosis, admixed with a minority of active tumor. Such changes have been called early treatment related necrosis (ETRN) by some investigators [Chamberlain MC 2007]. ETRN has important ramifications for patient care and clinical trial design. When ETRN occurs, effective regimens may be discontinued, investigative agents may appear more effective than they are, and clinical trials designed around progression free survival endpoints may prematurely terminate. ETRN’s association with overall survival is unknown.
We report a retrospective chart review of 65 newly diagnosed, high-grade gliomas treated at a single institution with concurrent radiotherapy and temozolomide. We identify eight patients with ETRN, defined as pathological findings inconsistent with tumor growth (n = 6) and >25% radiographic improvement despite therapy discontinuation (n = 1) or no therapy change (n = 1). ETRN occurred in all eight patients within six months of completing radiotherapy. We compared ETRN overall survivals with the entire cohort (n = 57); those who progressed within six months of completing radiation therapy (n = 23), but did not meet the definition of ETRN; and patients who progressed during radiation therapy (n = 7).
Patients who progressed during radiotherapy had the shortest median overall survival (OS) (n = 7, OS = 4.6 months, 95%CI = 3.0–8.8) when compared with all other patients (n = 58, OS = 20.1 months, 95%CI = 15.1–23.9). This result was highly statistically significant (log-rank p<0.0001). ETRN patients had a trend to longer survival(n = 8, OS = 21.5, 95%CI = 15.9–NR) when compared with all other patients (n = 57, OS = 16, 95%CI = 12.1–22), but this result was not statistically significant (log rank p<0.2713). However, patients with ETRN had longer median overall survivals when compared to other patients who progressed within six months of radiotherapy (n = 23, median OS = 9, 95%CI = 5.5–12) even when patients who progressed during radiotherapy (i.e., those known to have particularly short survival) were excluded from the comparison group (n = 16, median OS = 11.6, 95%CI = 8.9–16.1). Both results were significant, log-rank p<0.0016 and p<0.0067, respectively.
To our knowledge, this is the first demonstration of a statistically significant difference in survival in early treatment related necrosis patients. The data also provide valuable information that progression during chemoradiotherapy is associated with a particularly poor outcome, previous data examined progression during radiation therapy alone [Gaspar LE 1993].
Extent of resection is an important prognostic factor in most trials in high-grade glioma. Theoretically, a decreased tumor load prior to the start of adjuvant chemotherapy may influence efficacy of adjuvant treatment. We explored within three randomized trials if a resection impacted the outcome to adjuvant chemotherapy.
Exploratory and pooled subgroup analysis in three prospective randomized trials (EORTC 26882, adjuvant dibromodulciterol [DBD] and BCNU in high-grade glioma; EORTC 26951, adjuvant PCV in anaplastic oligodendroglioma; EORTC 26981, concurrent and adjuvant temozolomide in glioblastoma; n = 1356) on the influence of the extent of resection on overall outcome and on outcome to adjuvant treatment. In all trials extent of resection was recorded as reported by the operating neurosurgeon. Primary endpoint of the analysis was overall survival. The study compared resection vs. biopsy and biopsy vs. partial resection vs. complete resection. A p value of <0.05 was considered statistically signficant. A Cox proportional analysis was used to control for other major prognostic factors (performance status and age).
. In two trials a statistically significant risk reduction was observed after adjuvant chemotherapy in the resected patients, but in none of the trials in the biopsied patients (table, adjusted analysis). Although in the pooled analysis adjusted for main prognostic factors the risk reduction of adjuvant chemotherapy was significant in the resected patients (n = 1,138, HR [95% CI] 0.71 [0.62–0.81]) but not in the biopsied patients (n = 218; HR [95% CI] 0.89 [0.67–1.19]), the adjusted interaction test was negative (p = 0.24). In younger patients (<50 yrs) the interaction test was positive (p = 0.03), but not in the group with good performance status (PS 0 and 1, p = 0.28). Differentiating between partially and completely resected patients did not impact upon the treatment effect.
Although efficacy of adjuvant chemotherapy appears to be more pronounced in patients who have undergone a resection, this is also influenced by other prognostic factors (age, PS) and statistical significance was not reached. Heterogeneity between the trials and the limited numbers of biopsy patients may have influenced the present findings.
Multimodality treatments with surgical resection followed by standard radiotherapy (60 Gy/2Gy fr) and adjuvant chemotherapy in patients with glioblastoma multiforme (GBM) have produced significant results even though the prognosis remains poor. Therefore exploring hypofractionated regimen using new radiotherapy technique is an attractive strategy. Recently the advance of technology has permitted to plan radiation treatment with high comformal index and important sparing of normal tissue. The rationale of this study was to combine the potential radiobiologic advantage of hypofractionation to GBM with highly conformal radiotherapeutic IMAT technique allowing to perform radiation treatment with a sharp dose gradient such as stereotactic radiotherapy treatments (SRT) even on PTV over 3 cm diameter. The authors present their results on hypofractionated IMAT radiotherapy regimen in the treatment of patients with GBM.
Thirty eligible patients were accrued between October 2002 and September 2005. All patients underwent surgery (craniotomy and cytoreduction of the mass). Planning CT-MRI was obtained for every patient in treatment position with immobilization device (head mask). PTV1: surgery field +/− residual disease (CTV1) was confirmed with MRI T1-weighted (with gadolinium) plus 3-mm margin for geometric uncertainty; PTV2: perilesional edema by MRI T2-weighted images (CTV2) plus 3-mm margin for geometric uncertainty. Dose delivered was 25 Gy/5 fr to PTV1 and 20 Gy/5 fr to PTV2, dose was prescribed at 70% isodose surface with internal dose gradient of about 30% and the hot-spot was localized on the eventual residual disease. Treatment planning was obtained by inverse planning module (Eclipse Varian).
Four weeks after radiotherapy, treatment with temozolomide was started (200 mg/m2; for 5 days every 4 weeks) and administered until disease progression.
Out of 30 patients, 28 were evaluated for outcome (two refused temozolomide). No G3 radiation toxicities were observed and hematologic toxicity with temozolomide G3 was reported in only 3 patients. Median survival was 14 months (range 4–48) and global survival at 1 and 2 years was respectively 53% and 30%.
Hypofractionated radiotherapy with IMAT technique and sequential temozolomide is safe and well tolerated. Toxicities and survival results, median and at 2 years, in agreement with literature data, suggest that hypofractionated radiotherapy could represent an alternative to standard treatment. Furthermore it is shorter then traditional treatment (60 Gy in 30 fr) and may improve patients, quality of life. Obviously, further studies are needed to compare this technique with standard treatment modalities.
To compare time to progression in patients with non resectable glioblastomas (GBL) treated with either radiotherapy (RT) or a nitrosourea based chemotherapy (CT) with platinium salts.
Phase III multicentric prospective randomized study comparing time to progression (TTP) in patients diagnosed for GBL confirmed with histological diagnosis according to WHO classification. All patients had nonresectable tumor because of location or callosal extension. CT (F/CDDP/VP16) consisted of 4 monthly cycles of fotemustine (100 mg/m2, d1), cisplatinium (33 mg/m2/d, d1–3), etoposide (75 mg/m2/d1–3) followed by conventional fractionated radiotherapy (55–60 Gy). MRI was performed at baseline before and after surgery and every 8 weeks. Clinical and QOL evaluation was monthly reported.
50 patients were randomized (26 CT-RT and 24 RT alone) with mean age : 58 yrs (33–71); mean Karnofsky score : 86.4 +/− 11; mean time between biopsy and NU-CT or RT was less than 6 weeks. 70% (35/50) had only stereotaxic biopsy and others had partial resection. Time to progression was 242 days (IC95: 99–314) for CT-RT versus 131 d (IC95: 92–208) for RT alone; p = 0.0085. Median survival was 375.5 d (IC95: 323–427) for CT-RT vs. 325 (IC95: 200–504) p = 0.36 for RT alone. The best tumoral response was 23% for CT-RT vs. 8% with RT. Patients had the following grade III or IV side effects: CT-RT, 62% hematological toxicity; RT, 17% neurological toxicity. Nine patients died during the follow-up (6 tumor progression, 2 sepsis, 1 unknown).
Time to progression and median survival are increased with the association CT followed by brain RT.
ENZ, an oral serine/threonine kinase inhibitor, suppresses signaling through PKCβ and the PI3K/AKT pathways to induce apoptosis, reduce cell proliferation, and suppress angiogenesis. PKCβ is activated by vascular endothelial growth factor, an angiogenic factor upregulated in most GBM. Phase I assessed maximum tolerated dose of ENZ given concurrently with RT plus TMZ and then adjuvantly with TMZ in pts with GBM/GS to define a phase II ENZ dose. Phase I secondary objectives assessed safety, pharmacokinetics, antitumor activity, pharmacogenomics, quality of life (QoL), and utility of magnetic resonance perfusion imaging. The phase II primary objective was overall survival, and secondary objectives continued most phase I endpoints plus progression-free survival.
Pts with newly diagnosed GBM/GS and KPS ≥60 were enrolled. Enzyme-inducing antiepileptic drugs were not permitted. Treatment began <5 weeks after diagnosis with RT 60 Gy given over 6 weeks with TMZ 75 mg/m2 given daily during RT and then adjuvantly at 200 mg/m2, from days 1–5 of a 28-day cycle. ENZ was given orally, once daily, during RT at a starting dose of 250 mg/day (cohort 1) and continued at the same dose in the adjuvant setting. ENZ was escalated to 500 mg/day (cohort 2), if, during RT and the first adjuvant cycle of TMZ, no more than 1/6 pts had a dose limiting toxicity (DLT). DLTs were ≥G3 thrombocytopenia (Tp), G4 anemia/neutropenia, or ≥G3 nonhematologic toxicity. Planned duration of adjuvant TMZ/ENZ therapy was 12 cycles. QoL was assessed using FACT-Br and MDASI-BT.
From 9/06–6/07, 12 pts enrolled in Phase I. There were no DLTs in the 6 pts of cohort 1. One pt withdrew and later progressed. As of 2/1/08, 5 pts in cohort 1 continue on ENZ 250 mg/day at ≥14 cycles. In cohort 2, 2/6 pts experienced a DLT; one G4 and one G3 Tp. One pt recovered to ≤G1 in <28 days and reinitiated adjuvant TMZ and reduced-dose ENZ. The other pt recovered to ≤G1 after 28 days, but discontinued. Three pts discontinued due to progressive disease, and one pt continues on ENZ 500 mg/day. Phase II began accrual in 9/07.
ENZ 250 mg/day with RT+TMZ was well tolerated and is the recommended phase II dose. In addition to phase I, available phase II, and QoL data, we shall present preliminary pharmacogenomic results evaluating the predictive value of potential ENZ molecular targets on treatment outcomes as well as initial results with the use of physiological imaging for pt evaluations.
To evaluate toxicity and efficacy of an intensified schedule of temozolomide plus indomethacin in addition to standard radiotherapy plus concomitant temozolomide in patients with newly diagnosed glioblastoma.
This trial (UKT-05) enrolled 41 patients treated with pre-irradiation temozolomide at 150 mg/m2 (one week on/one week off), radiotherapy with concomitant temozolomide (50 mg/m2), maintenance temozolomide starting at 150 mg/m2 according to an intensified one week on/one week off schedule, and maintenance indomethacin (25 mg bid) treatment. The primary endpoint was median progression-free survival (PFS). O6-methylguanine-DNA methyltransferase (MGMT) gene promoter methylation was assessed by methylation-specific PCR in 39 (95.1%) patients.
The median follow-up interval is 15.3 months (minimum, 9.4 months). Hematologic grade 3/4 toxicities according to the Common Terminology Criteria for Adverse Events (CTCAE; v3.0) were: anemia (3/41, 7%), leukopenia (9/41, 22%), lymphopenia (26/41, 63%), neutropenia (6/41, 15%), and thrombopenia (8/41, 20%). Treatment-related non-hematologic CTCAE grade 3 through 5 toxicities were reported for n = 6/41 (15%) patients. Median PFS was 7.6 months [95% CI, 6.2–10.4 months]. PFS rate at 6 months (PFS-6) was 70.7% [95% CI, 54.3–82.2%]. The censored one year survival rate was estimated at 71.8% [95% CI, 54.7–83.4%]. Median PFS, PFS-6 and median survival time (MST) in 16 patients with MGMT gene promoter methylation were significantly larger than in 23 patients without epigenetic MGMT gene silencing (median PFS: 15.8 vs. 6.2 months, P = 0.001; PFS-6: 81.3% vs. 65.2%; MST: 24.1+ vs. 12.9+ months, P = 0.027).
The dose-intensified regimen of temozolomide administered in a one week on/one week off schedule resulted in relevant but acceptable toxicity. Compared with data from the EORTC/NCIC trial 26981–22981/CE.3, patients with an unmethylated MGMT gene promoter in the tumor did not appear to benefit from intensifying the temozolomide schedule regarding median PFS or MST whereas the efficacy in the methylated cohort looks very promising. At the time of EANO, updated results and survival estimates after a minimum follow-up of at least 15 months will be available.
The major driving force of mortality in human glioblastoma (GBM) is in most cases the recurrence of tumor growth, likely caused by a heterogeneous population of rapidly dividing tumor cells that exhibit hitherto unrecognized resistance mechanisms to standard therapy. To assess the heterogeneous nature of GBM tissue, we have isolated and characterized paired biopsy samples from a total of 26 patients in direct comparison. In each case, the samples derived from distinct regions of the tumor mass and microscopic evaluation revealed dissimilar histological characteristics. For controlled analysis, we used defined in vitro conditions for the isolation and the expansion of cells from these paired biopsy samples. Our data indicate a distinct topography of the key features of cellular malignancy, i.e., the content of stem cells, and the respective migratory and proliferative potential of cells vary significantly between individual tumor regions. Because FACS and qPCR analysis furthermore revealed a regionally confined expression profile of therapeutically relevant surface antigens, we would like to suggest that the investigation of cells from different tumor regions should necessarily be considered for future diagnosis and treatment of GBM.
*MG was supported by the BONFOR program of the University of Bonn Med Ctr.
ΔBS is supported by the VW Foundation.
Diffusely infiltrative gliomas can be histopathologically classified as astrocytic tumors, pure oligodendroglial tumors, and mixed oligoastrocytic tumors. However, the classification and grading of these tumors can be usually difficult. The knowledge of the molecular genetic background of gliomas is helpful in classifying diffuse gliomas with regard to prognosis and response to chemotherapy. In particular, loss of heterozygosity on chromosome arms 1p and 19q is a frequent event in oligodendroglial tumors, significantly correlated with better prognosis and chemosensitivity. In this work, we aimed to identify molecular markers helpful in the diagnosis and grading of gliomas. For this purpose, we performed expression profiling analyses in 37 gliomas: 19 oligodendrogliomas (OG), 9 mixed oligoastrocytomas (OA), and 9 astrocytomas (AC). Functional analyses were carried out by using GEPAS (CIPF, Valencia, Spain) and GSEA (Broad Institute, Massachusetts, Cambridge). A comparative analysis of tumors of WHO grades II and III revealed that genes significantly up-regulated in grade III were involved in cell cycle, cell proliferation, and metabolism processes. Likewise, comparision of the gene expression profiles of OG, OA, and AC subtypes showed that mixed oligoastrocytomas presented gene expression patterns more similar to astrocytomas than to oligodendrogliomas. Significant up-regulation of CD44, C4B, or ANXA1 were detected in astrocytomas and mixed oligoastrocytomas, while they were underexpressed in oligodendrogliomas. Our findings support the molecular relevance of the astrocytic component in oligoastrocytomas and propose a set of candidate genes which may be useful for the differential diagnosis of gliomas and tumor grading.
Astrocytes may give raise to glioma, the most frequent primitive CNS tumors. TGF-alpha, an EGF family member, is trophic for both astrocytes and gliomas cells, but its sole deregulation is insufficient to lead to cancerous transformation of astrocytes. It however triggers their conversion into neural progenitor-like cells capable to give birth to neurons (Sharif et al, Oncogene, 2007). Neural progenitors have been reported to be more sensitive than astrocytes to oncogenic manipulations in transgenic mice. These data, and the frequent overexpression of TGF-alpha in early steps of gliomas progression, led us to envisage a participation of this factor to tumor genesis through the destabilization of the mature astrocyte phenotype. Irradiation, a classic mutagen, was used to determine whether astrocytes converted into neural progenitor-like cells are more sensitive than their mature counterpart to cancerous transformation. Mouse astrocyte cultures were maintained 7 days in serum-free medium without (control) or with TGF-alpha (50 ng/ml), prior to be either sham-irradiated or irradiated (137Cesium source, 5 Grays once fraction). Control cultures survived 2 weeks post sham-irradiation or irradiation. On the opposite, TGF-alpha-treated astrocytes were immortalized upon irradiation. Their cancerous transformation was ascertained by their ability to form colonies in methylcellulose medium, their karyotype anomalies, and the formation of sub-cutaneous tumors in NOD-SCID mice and high-grade glioma-like tumors after grafting into nude mice CNS. Sham-irradiated cells failed to show any evidence of transformation. These results demonstrate that instability of the mature astrocyte phenotype due to a single epigenetic change is sufficient to sensitize them to cancerous transformation. They allow proposing that TGF-alpha does not only favor growth and survival of established gliomas but participates also to the earliest stages of their genesis.
Investigation of the effect of temozolomide (TMZ) alone or combined with irradiation in human glioblastoma multiforme (GBM) cell lines.
A panel of four established (U251, D384, TG98 and HS682) and seven long term primary GBM cell lines were analyzed for the promotor methylation status of the methylguanine-DNA methyltransferase (MGMT) gene by a methylation specific multiplex ligation-dependent probe amplification assay. MGMT protein expression was studied by Western blotting. TMZ sensitivity of the cells was tested in a clonogenic assay by single exposure to TMZ (0–500 μM) for 24 h. Three cell lines that were sensitive to TMZ in the clinically feasible dose range were selected for combined treatment with irradiation. Cells were irradiated with five daily 2 Gy fractions with TMZ (3–10 μM) administered 1 h prior to each irradiation fraction. Treatment response was evaluated on cell proliferation and by clonogenic assays.
TMZ sensitivity was highly variable between the GBM cell lines: four of the eleven GBM cell lines expressed the MGMT protein and were TMZ resistant. The other seven cell lines had absent or diminished MGMT protein expression and were TMZ sensitive: the fraction of surviving clonogenic cells was reduced to <0.5 after 24 h exposure to <50 μM TMZ. The TMZ sensitivity correlated with an absent expression of the MGMT protein. However, no correlation was found between the sensitivity of cells to TMZ and the MGMT promotor methylation status. Two out of four partly MGMT promotor methylated cell lines still showed expression of the protein. In the combination experiments with fractionated irradiation and concomitant TMZ, one out of three (TMZ sensitive) cell lines showed a radiosensitizing effect of TMZ.
Expression of the MGMT protein, more than the MGMT-promotor methylation status, predicts the sensitivity of human glioma cells to TMZ. The effect of TMZ is additive to irradiation, but not necessarily synergistic. Neither the MGMT promotor methylation status, nor the absence of the protein predicts for the radiosensitizing effect of TMZ.
Supported by the Dutch Cancer Society project #VU 2000-2149.
Glioblastomas (GBM) are among the most aggressive human cancers with a median survival of 14.6 months despite multimodal therapy. Although today’s standard of care—temozolomide—significantly increases the proportion of patients surviving for more than two years, long-term survivors are still a rare exception. The biological properties of cancer stem cells (CSC) maintaining GBM provide a possible explanation for the failure of chemotherapy in the long term: CSC from several tumor entities overexpress multidrug resistance proteins which may protect them against cytotoxic drugs that kill progenitor and differentiated cells. CSC would then give rise to recurrent tumors. However, although multiple authors have postulated this model, it has not yet been convincingly confirmed experimentally for GBM. We therefore investigated the effect of temozolomide on CD133+ and CD133− CSC lines (n = 5) using an established in vitro model of GBM stem cell lines. Surprisingly, temozolomide induced a dose- and time-dependent decline of the rapidly proliferating stem cell subpopulation while differentiated tumor cells constituting the bulk of all tumor cells were resistant to the cytotoxic effects of the substance. Incubation with sublethal concentrations of temozolomide for 2 days eliminated clonogenic tumor cell growth in vitro and reduced the tumorigenicity in vivo. The susceptibility of CSC to temozolomide was meditated by O6-methylguanine-DNA-methyltransferase (MGMT), the MGMT promoter methylation status did not differ between CSC and differentiated tumor cells. MGMT-expressing CSC showed a 10 fold higher resistance to temozolomide which was reversible by the MGMT inhibitor 6-benzylguanidin. Despite the selective elimination of CSC in GBM, we found that CSC in all stem cell lines investigated were completely resistant toward the cytotoxic effects of 5 μM temozolomide, which corresponds to the peak concentration measured in the cerebrospinal fluid of patients. While concentrations required to eliminate MGMT positive CSC (500 μM) can not be reached in the patient, concentrations up to 50 μM have been measured in the plasma of patients after standard treatment and sufficed to eliminate CSC in MGMT negative tumors. While these findings explain the poor outcome of MGMT positive tumors, it is unclear why patients with MGMT negative tumors die despite the susceptibility of CSC to temozolomide. Considering the impaired blood-brain barrier in the tumor and infiltration of CSC in otherwise healthy brain tissue with much lower temozolomide concentrations, we postulate that recurrent GBM in MGMT negative tumor derive from invading CSC protected against temozolomide by the intact blood-brain barrier. Together, this study will fundamentally change the understanding of CSC and bears important implications for the conception and interpretation of future and past clinical trials.
Genetic and epigenetic aberrations, like loss of chromosomal heterozygosity (LOH) and hypermethylation of gene promoters, are frequently observed in glial tumors. Detection of these aberrations, which is performed on tumor DNA, carries a potential diagnostic and prognostic relevance. As sample materials for diagnosis should be easily accessible by a minimally invasive procedure, there has been much interest in the potential use of nucleic acid markers in the blood of patients with cancer.
To find whether genetic aberrations in serum DNA of glial tumors can be detected at a time distant from surgical intervention and to evaluate its concordance with tumor DNA.
DNA was extracted from the serum and the paraffin embedded tumor sections of 35 patients with GBM and 23 patients with oligodendrogliomas (18 grade II; 5 grade III). The methylation status of O(6) methylguanine-methyl-transferase (MGMT) promoter and chromosomal 10q LOH were evaluated in DNA samples obtained from both the serum and tumor sections.
In GBM, the incidence of MGMT promoter methylation was 43% in the tumor and 31% in serum while 10q LOH was found in 64% of tissue samples and 43% of serum DNA. In oligodendrogliomas, the incidence of MGMT promoter methylation was 70% in the tissue and only 17% in serum. The concordance between the results obtained from tumor DNA and serum DNA for both MGMT methylation status and 10q LOH was 71% for GBM. In oligodendrogliomas the concordance of the results for MGMT was 48%.
These results imply that serum DNA, which is obtained at various times following tissue diagnosis or initial treatment, often represents the tumor DNA. The concordance between serum and tumor DNA is better for GBM than for oligodendrogliomas. It is possible that higher proportion of GBM patients had an active tumor at time of serum sampling as compared to patients treated for oligodendrogliomas. It cannot be excluded that high-grade tumors shed more DNA into the bloodstream, explaining the better concordance observed in GBM. We are currently correlating results of paired serum-tumor samples with clinical and imaging details of each patient. In addition, we are in the process of comparing the rate of detection of genetic aberration following extraction of serum DNA by two different techniques in order to refine the methodology. These results will be presented at the conference.
The incidence of self-reported stroke in survivors of childhood Hodgkin’s lymphoma (HL) is increased compared to their siblings. It is not clear whether the increased risk is also present in adults and when analyses are restricted to clinically verified strokes and transient ischaemic attacks (TIAs).
We performed a cohort study among 2,202 5-year survivors treated for HL in the period 1965–1995 before age 51. We compared the incidence of stroke and TIA with the general population and calculated standardized incidence ratios (SIRs) and absolute excess risks (AERs per 10,000 patients per year) for different age, treatment, and sex strata.
During follow-up (median 17.8 years), 61 (SIR 1.9 [1.5–2.4], AER 9) and 47 (SIR 3.4 [2.5–4.5]) patients developed stroke and TIA, respectively. Nine strokes (SIR 4.0 [1.8–7.6], AER 8) and seven TIA (SIR 10.0 [4.0–20.6], AER 8) occurred among patients treated for HL before age 21. SIRs were increased among those treated with radiotherapy (RT) alone (especially neck RT), or with RT+chemotherapy (CT). Females experienced higher risks than males, especially those treated before age 21. The 30-year cumulative risk of CVA was 10.6%.
Survivors of HL are at increased risk for clinically verified stroke and TIA. The risks are treatment and gender related and substantially higher among childhood and adolescent HL survivors compared to adults.
Bevacizumab, a monoclonal antibody targeting vascular endothelial growth factor, has recently been studied for both initial treatment and recurrence of malignant gliomas. Initial data indicate that patients show an increased response rate as well as an increase in progression free survival as compared to historical controls. Despite this promising data, isolated reports have emerged of acute optic neuropathy in a small subset of patients with malignant gliomas treated with bevacizumab.
Using the clinical records from 2005 to 2008, we identified adult patients with glioblastoma treated with bevacizumab who developed acute optic neuropathy. Four institutions participated in this study including the University of Virginia, UCLA, Columbia University, and Rush University. Age at diagnosis, gender, radiation therapy data, the chemotherapeutic regimen including the bevacizumab dosing schedule, ophthalmologic records, laboratory values, and MR imaging were assessed.
Five patients were identified, including one male and four females. Mean age at diagnosis was 54 years (range 37 to 68). Following surgical resection of a glioblastoma all patients received fractionated radiation therapy with concomitant temozolomide. One patient received bevacizumab at initial diagnosis, while the other four received the drug at recurrence. The mean radiation dose to the optic chiasm, left optic nerve, and right optic nerve was 5,509.7 cGy, 4,344.3 cGy, and 3,679.8 cGy, respectively. The mean time from the end of radiation therapy to the onset of visual symptoms was 11.5 months (range 8 to 15.5). The patients completed a mean of 10.4 doses (range 3 to 18) of bevacizumab prior to onset of visual symptoms. Visual loss subsequently developed unilaterally in three patients and bilaterally in two patients. MR imaging from one of the former patients displayed left optic nerve enhancement while another patient displayed parenchymal gliomatosis. MR imaging from one of the latter patients displayed bilateral optic nerve enhancement as well as evidence of parenchymal gliomatosis. Post-mortem pathologic specimen of the patient with left optic nerve enhancement displayed no evidence of tumor infiltration. CSF analysis was performed in four patients and showed no evidence of demyelinating disease or meningeal gliomatosis. The mean time from visual symptom onset to no light perception in at least one eye was 1.7 months (range 0.5 months to 4.5 months).
A small subset of glioblastoma patients treated with standard chemoradiation therapy has developed optic nerve dysfunction following treatment with bevacizumab. Patients receiving bevacizumab should be followed closely in order to clarify whether this complication represents drug-related optic neuropathy, coincidental radiation optic neuropathy, or an unusual bevacizumab-related pattern of tumor failure.
We proposed a few years ago a new system classification called Basic Score for Brain Metastases (BSBM) built on a retrospective study concerning patients with brain metastases (BM) treated by radiosurgery. The three significant prognostic factors were (i) a Karnosky Performance Status (KPS) equal or superior to 80 (no = 0; yes = 1 point), (ii) an absence of a systemic disease (no = 0; yes = 1 point), and (iii) a primary disease controlled (no = 0; yes = 1 point). This classification varied from 0 to 3 points. Since January 2003, we followed prospectively the new cohort of patients. This study analyzed the survival of both (retrospective and prospective) populations.
267 evaluable patients with more than 700 brain metastases were treated with a Leksell Gamma Knife model C. Median marginal dose was 20 Gy at a median 50% isodose. There were 45% (119 pts) lung cancer, 22% (59 pts) breast cancer, 19% (49 pts) melanoma, and 15% (40 pts) other primary tumors. The retrospective study concerned 110 patients (102 dead = 92%) treated from December 1999 to December 2002. The prospective study concerned 157 patients (95 dead = 61%) treated from January 2003 to December 2005.
median survival (MS) for the entire population is 11 months and 3, 6, 12, and 23 months for respectively BSBM 0, 1, 2, and 3 points subgroups. For the selective categories BSBM 0, 1, 2, and 3 points, the MS were respectively 3, 3.5, 11, and 20 months for the retrospective study versus 3.5, 7, 13, and 27 months for the prospective study. The multivariate analysis identifies as significant factors : the KPS, the SIR and the BSBM models. On the validation set, the discrimination power of the BSBM score assessed by the area under the ROC curve. The difference between the retrospective and the prospective subgroups is not statistically different.
The BSBM system’s classification is comparable in the retrospective and prospective studies suggesting the value of this new original classification.
We previously reported that piecemeal resection (PMR) of a posterior fossa or supratentorial brain metastasis increased the risk of LMD development compared to enbloc resection (EBR). In this study, we assess the association between resection methodology and local recurrence in patients with a previously untreated SBM and no AWBRT.
An institutional review board approved data review was performed. This report summarizes data from 570 patients with an SBM treated at M.D. Anderson between June 1993 and December 2006 identified from the prospectively compiled Brain and Spine Center Database. The primary outcome was incidence of recurrence at the resected site (local recurrence). Of note, in most cases, the method of resection was prospectively noted by the neurosurgeons and entered into the database at the time of the procedure. In determining the incidence of local recurrence, patients were censored at the time of WBRT; a biopsy, resection, or stereotactic radiosurgery to the index site after the index procedure, whichever occurred first. The Cox proportional hazards method was used to assess local recurrence following PMR and EBR.
The median patient age was 58 years; 55% were males; 95% had a Karnofsky score >70. The most common primary type was lung (28%), melanoma (21%), renal (19%), and breast (10%). PMR was performed in 202 patients (35%) and EBR in 368 (65%). The two groups were comparable with respect to gender, age at surgery, KPS score, and tumor location with respect to eloquent grade. They were also comparable with respect to status of systemic cancer and primary cancer type, except that PMR was used significantly more often in patients with renal cell carcinoma. Tumors resected following the PM approach were significantly larger (median volume 15.8 cc versus 7.6 cc in the group resected following the EB approach). Local recurrence was observed in 84 patients (15%). In preliminary analyses, undergoing PMR resection carried a significantly higher local recurrence risk than EBR [hazard ratio (HR) for PM = 1.7; 95% confidence interval (CI), 1.1–2.6; p = 0.03]. The effect of method of resection in patients with different prognostic characteristics is being assessed.
In preliminary analyses, there was an increased risk of local recurrence with piecemeal resection of a SBM compared to EBR in patients not treated with adjuvant whole brain radiation. Piecemeal resection was previously found to be independently associated with an increased risk of LMD in patients with posterior fossa and supratentorial metastasis at our center. Prophylactic measures such as early institution of AWBRT may be warranted in some situations in which PMR of the tumor is unavoidable.
The authors investigated outcome of stereotactic radiosurgery (SRS) alone for the patients with brain metastases, compared with SRS plus whole-brain radiotherapy (WBRT).
261 patients with 1 to 10 metastatic lesions were retrospectively analyzed. 180 patients were treated with SRS alone (marginal dose 13–26 Gr), and 81 patients received SRS (marginal dose 12–19 Gy) plus WBRT (30 Gy). Actuarial curves were estimated using the Kaplan-Meier method regarding overall survival (OS), distant brain control (DC), and local brain control (LC) stratified by RPA class. Analyses for known prognostic factors including age, sex, performance status, primary tumor origin, number of brain metastases, extracranial metastases, treatment modality, and RPA class were performed using the Cox proportional hazards model.
In RPA class 1, survival time with SRS plus WBRT was significantly longer than SRS alone (390 vs. 931 days, p = 0.017). However, in RPA class 2, no significance in OS was observed between SRS alone and SRS plus WBRT (275 vs. 346 days, p = 0.311). In RPA class 1, DC was higher in SRS plus WBRT group (p = 0.047), while there was no difference in LC between two groups (p = 0.718). However, in RPA classes 2, DC and LC revealed no difference between two groups.
For the patients in RPA class 1, SRS plus WBRT appeared to be more effective than SRS alone, whereas addition of WBRT did not have significant impact on survival, DC, and LC in class 2. The result suggests that SRS alone may be recommended for the patients with limited life expectancy due to extracranial disease and addition of WBRT for the patients in whom long term survival is expected based on extracranial disease status.
Brain metastases from breast cancer (BCBM) have become more prevalent with improved systemic therapies for this malignancy. As a result, recurrent brain metastases after whole brain radiation therapy (WBRT) are becoming more common. Patupilone is an epothilone which crosses the blood brain barrier and is many-fold more potent than paclitaxel or docetaxel. This agent has activity against systemic breast cancer as well as brain metastases from non–small cell lung cancer. This trial evaluates patupilone in patients with BCBM.
This multicenter, phase II study will enroll 45 pts with progressive, measurable BCBM after WBRT with an exploratory group of 10 additional pts with leptomeningeal disease or asymptomatic, unirradiated brain metastases. Patients may continue trastuzumab and hormonal therapy but no other chemotherapy. Patients must be at least 4 weeks past WBRT, 3 weeks past cytotoxic therapy (6 weeks for nitrosoureas), 2 weeks past noncytotoxic agents, and on a stable dose of steroids. Exclusions are prior treatment with epothilones and NCI CTC grade > 1 diarrhea or neuropathy. Patients receive patupilone 10 mg/m2 i.v. over 20 minutes every 3 weeks with MRI brain and CT chest/abdomen every 6 weeks. The primary endpoint is 3-month central nervous system (CNS) progression free survival (PFS). Secondary endpoints include toxicity, systemic disease response rate, and overall survival.
18 pts have enrolled since February 2007 including 14 in the primary study group. The median age is 55 (43–68 yrs). All have had prior chemotherapy: 94% taxanes and 71% anthracyclines. Baseline KPS is 60–70 in 28% (5 pts) and 80–100 in 72% (13 pts). Toxicity data are available for 17 pts. The most common toxicity has been diarrhea—grade 1: 8 pts (47%); grade 2: 4 pts (24%) and grade 3: 5 pts (29%). Four grade 4 events (hypokalemia , upper GI bleeding, and anemia) have occurred. Three pts have had 25% dose reductions (myalgia, diarrhea/hypokalemia, and GI ulcer). Three patients have died, 1 possibly related to therapy. To date, best responses have been 3 partial responses (1 confirmed), 3 stable disease, 9 progressive disease, and 1 too early. One pt withdrew before assessment, and 1 pt died before assessment. The 3-month PFS in the CNS is 3/18 (17%). One pt suspected of radiographic progression on MRI underwent resection and had only necrotic tissue without viable tumor.
Patupilone has activity against recurrent breast cancer brain metastases. The three-month CNS progression-free survival is 17%. Toxicity is acceptable in this heavily pretreated patient population. The finding of necrosis in one patient with apparent radiographic progression suggests the need for careful evaluation of previously irradiated lesions in patients treated with chemotherapy for BCBM. The role for agents like patupilone will be come more important as the ability to control systemic disease improves
Information about the MGMT (O6-methyl-guanine DNA methyltransferase) promoter methylation status and LOH (loss of heterozygosity) on 1p and/or 19q becomes increasingly important for prognostic evaluation and treatment decision in glioma patients. The aim of this prospective study (10/2006–10/2007) was to evaluate whether (1) small samples obtained from stereotactic biopsy are sufficient for the determination of the MGMT and LOH status, and (2) extended biopsy procedures (which are necessary for additional molecular-genetic characterization) are associated with an increased risk.
Molecular stereotactic biopsy was indicated in case of highly eloquent tumor location and/or significant co-morbidity. Multiplanar trajectory planning with the BrainLab navigation software was based on co-registrated CT, MRI, and 18F-FET PET data. A serial biopsy was taken along a trajectory representative of the solid tumor including the area with the highest FET uptake. The maximum amount of tissue per specimen was 1 mm3. At least two samples collected from different sites were used for molecular-genetic analysis. Methylation-specific PCR for determination of promoter methylation and microsatellite analysis for LOH1p19q were specifically adopted for small sample sizes. Any adverse sequel potentially attributable to the biopsy procedure was considered as morbidity.
216 patients were included. Histopathological evaluation revealed 41 astrocytomas WHO grade II, 47 astrocytomas WHO grade III, 23 oligodendrogliomas/oligoastrocytomas (grade II/III), and 105 glioblastomas. The MGMT and LOH 1p/19q status could be determined in 99.5%. The overall promoter methylation rate was 50% and LOH on 1p/19q was seen in 28%. An isolated allelic loss on 1p or 19q was found in 3% and 8% of patients, respectively. 12–24 tissue samples per tumor were considered necessary by the neurosurgeon and the intra-operatively attending neuropathologist to achieve a sufficient tumor characterization. Tissue samples taken from distinct sites throughout the tumor always revealed the same MGMT and LOH 1p/19q pattern. The transient morbidity rate was 0.48%. There was no permanent morbidity and no mortality.
Because of its homogeneous distribution within each tumor, information about the MGMT promoter methylation and the LOH pattern on 1p and 19q can be reliably obtained from small sized stereotactic biopsies. Extensive but still minimally invasive biopsy procedures can be done with minimal risk using a guided multimodal imgaging approach.
Modern treatment strategies of gliomas have to consider recent advances in metabolic imaging and molecular markers with respect to individualized multimodal therapeutic regimes. Aim of this study was to correlate 18F-FET-PET based tumor grading, LOH1p/19q, and MGMT promoter methylation with histological classification of gliomas WHO° II and III from serial stereotatic biopsies in patients with eloquently located tumors.
86 patients have been included in this prospective, single-center study. Dynamic 18F-FET PET studies have been matched with MRI scans prior to stereotactic surgery to mark regions of interest. Oligodendroglial tumor characterization by 18F-FET PET was performed using max standard uptake value (mSUV) combined with kinetic uptake analysis. Biopsies have been classified according to WHO criteria. LOH1p/19q and MGMT methylation status has been detected using small tissue samples adjusted fluorescence in situ hybridization and methylation-specific PCR methods.
8 oligodendrogliomas (O), 24 mixed oligoastrocytomas (OA), and 54 astrocytomas (A) (WHO° II and III) have been included. 1p/19q codeletion (33/78 = 42.3%) was significantly more frequent in tumors containing an oligodendroglial component (25/31 O+OA [80.7%] versus 8/47 A [17.0%]; p<0.05). 55/75 gliomas (73.3%) exhibited methylation at the MGMT gene (26/28 O+OA [92.9%] versus 29/46 A [63%]; p<0.05). Oligodendroglial tumors showed significantly elevated LOH 1p/19q and MGMT methylation coexpression status compared to astroglial tumors (24/28 O+OA [85.7%] versus 6/45 A [13.3%]; p<0.05). Oligodendroglial tumors (24/26) have been reliably identified by analyzing 18F-FET-PET mSUV and uptake kinetics (22/26 O+OA [84.6%]). 43 out of 79 patients underwent multimodal treatments prior to stereotactic procedure. In case of LOH 1p/19q and MGMT methylation coexpression status 11/11 newly diagnosed patients with oligodendroglial tumors were initially treated with chemotherapy. 2/2 oligoastrocytomas without LOH/MGMT methylation underwent primary radiotherapy.
Combination of metabolic 18F-FET-PET analysis and molecular genetic marker profiling with histopathological classification of serial stereotactic biopsies was sufficient to gather more information on the difficult discrimination of oligodendroglial tumors and mixed oligoastrocytoma from astrogliomas. Since MGMT and LOH are correlated with favourable prognosis in certain glioma entities this diagnostic strategy can evaluate the role of molecular markers in a more general concept of glioma therapy.
The aim of the present study was to evaluate the effects of BNCT in the treatment of glioblastoma multiforme using a novel protocol for the boronophenylalanine-fructose (BPA-F) infusion.
The study included thirty patients, 26 to 69 years old with a good performance status, of which 27 have undergone debulking surgery. BPA-F (900 mg BPA/kg body weight) was given i.v. over 6 hours. Neutron irradiation started 2 hours after the completion of the infusion. We also used microdialysis to investigate the extracellular in vivo kinetics of boron in three intracerebral compartments—solid tumor, brain adjacent to tumor (BAT), and the normal brain, as well as the subcutaneous tissue before, during, and after BNCT treatment.
In tumor tissue the extracellular concentration of BPA followed that of blood with a maximal concentration of 31.2 ppm and a maximal ratio vs. blood of 1.07. In BAT, the peak level delayed for 4–6 h compared to the peak in blood with a maximal ratio of 1.2. The uptake of BPA in the normal brain was lower compared to the blood and tumor tissue. The boron-blood concentration during irradiation was 15.2–33.7 μg/g. The average weighted absorbed dose to normal brain was 3.2–6.1 Gy (W). The minimum dose to the tumor volume ranged from 15.4 to 54.3 Gy (W). Seven patients suffered from seizures, 8 from skin/mucous problems; 5 patients were stricken by thromboembolism and 4 from abdominal disturbances in close relation to BNCT. Four patients displayed 9 episodes of grade 3–4 events (WHO). At the time for follow-up, minimum 10 months, 23 out of the 29 evaluable patients were dead. The median time from BNCT treatment to tumor progression was 5.8 months and the median survival time after BNCT was 14.2 months.
The efficacy of BNCT in the present protocol seems to be comparable with conventional radiotherapy. However, the observed side effects and the requirement of complex infrastructure emphasize the need of further evaluation, especially directed to improve the accumulation of 10B in tumor cells.
Since 2002, we have applied BNCT for over 50 cases of malignant gliomas with epithermal neutron and two different boron compounds (BSH and BPA). BNCT is a binary cancer therapy involving the administration of B10-containing compounds which concentrate to tumor and subsequent irradiation with a neutron beam. The neutrons interact with the boron to cause the boron atom to split into an alpha particle and lithium nucleus. Both of these particles have short ranges equivalent to one cellular diameter, and therefore, BNCT can be thought of as a tumor-selective particle radiation. Here we examined the survival benefit of BNCT for 22 cases of newly diagnosed glioblastoma (NDGB) patients.
We have treated 22 cases of NDGB with surgical removal followed by BNCT. The early 11 cases were treated with BNCT alone, and in the recent 11 cases, total 20 to 30 Gy fractionated external beam X-ray radiation therapy (EBRT) was additionally applied after BNCT mainly to the deep part of the tumor, where insufficient neutron fluence was simulated. No chemotherapy was applied until tumor progression was observed. Survival benefit of BNCT was analyzed with reference to the RTOG Recursive Partitioning Analysis (RPA) database and our institutional historical controls.
Treatments were well tolerated without any kinds of acute systemic or local severe toxicity in all patients. Mean overall survival of all NDGB patients treated by BNCT (N = 22) was 22.8 months and the median survival time (MST) was 16.7 months with 2-year survival rate of 39.4%. Especially, patients treated with BNCT followed by EBRT (N = 11) achieved longer MST, 21.3 months with 2-year survival rate of 57.1%, without the significant increasing in frequency of radiation necrosis after treatment. Stratification by RPA criteria showed 8, 10, and 4 patients in class III to V, respectively. Four patients of 8 in class III and 2 of 12 in class IV are still alive, and all the patients in class V have died at the time of analysis. MST for the BNCT group compared to the RTOG database was as follows: 19.8 months vs. 17.9 months for class III; 15.1 months vs. 11.1 months for class IV; 15.9 months vs. 8.9 months for class V.
Our modified BNCT protocol showed a survival benefit in all of the RPA classes of the RTOG database not only for the good prognosis group, and the additional EBRT after BNCT might enhance the therapeutic benefit of BNCT and contribute to improve the prognosis of NDGB patients.
Dendritic cell (DC)–based immunotherapy is a promising approach in glioblastoma. Patients are vaccinated with tumor-antigen loaded DCs. After vaccination, the DCs have to migrate to lymph nodes and there to induce specific anti-tumoral T-cell responses. We have established such a protocol and report here on initial results.
Autologous tumor material is obtained from surgery. After minimal mechanical dissection, necrosis is induced and lysates are prepared. Monocytes enriched from leukapheresis products serve as the starting population for DC cultures. Successive generation of immature and mature DC for tumor lysate-loading and vaccination, respectively, are generated in a two-step culture system. Five vaccines are prepared and applied weekly by intradermal injection.
22/22 tumor samples processed yielded a preparation of sterile, avital tumor cell lysates. For 12/22 of these preparations, GBM and a tumor content of at least 75% was confirmed neuropathologically. For 5 patients, vaccines were produced. From leukapheresis products containing 1.5±0.1×1010 leukocytes with a monocyte frequency of 25.5±1.1%, monocytes were enriched to 97.6±0.9% purity. Immature and mature DC obtained after 6 and 9 days of culture showed typical phenotypic and functional properties of bona fide DC, including induction of anti-tumoral responses in vitro. Overall, for each patient 5 vaccines containing 3.1±1.0×107 cells each were produced with a final purity of CD83+ mature DC of 87.9±2.9%. Vaccination was well tolerated and severe adverse events have not been observed. In one patient there was clear evidence of anti-tumoral immunity induced by the vaccines. Two of the patients progressed and died early. One patient vaccinated after third recurrence died 260d after the last resection. One patient vaccinated between radiochemotherapy and temozolomide chemotherapy was progression free for 529d but died 663d after resection.
Vaccine production in glioblastoma patients is feasible and early results of vaccination, including induction of anti-tumoral immunity, are promising.
Cotara (131I-chTNT-1/B Mab) is a radioiodinated chimeric monoclonal antibody specific for histone H1 and DNA. Histones are found in the cell nucleus and upon cell death they are found tightly bound to DNA and thus can be exploited for TNT-1 targeting of degenerating or necrotic cells to selectively deliver 131I to the glioma cells. A clinical trial of patients with recurrent Glioblastoma is being undertaken to evaluate the maximum tolerated dose given as a single 25 hour intratumoral infusion.
Eligible patients with recurrent glioblastoma tumor volume between 5–60 cc were treated first with a convection-enhanced intratumoral infusion dose of 3 mCi for imaging of Cotara to assess biodistribution and calculate radiation dose. They were then treated 2–4 weeks later with a second therapeutic intratumoral infusion dose of 1.5 mCi/cc gross tumor volume.
Three patients have been treated with tumors of 13.5 to 59.9 ml gross total volume. The second infusion delivered a therapeutic dose of Cotara ranging from 17.11 to 73.67 mCi per tumor/patient. All patients tolerated the infusions well. No dose limiting toxicities (DLT) were seen in any of the patients. There was no significant normal tissue uptake observed beyond background. The tumor had a concentration >100 times the normal brain in the contralateral hemisphere. The mean tumor-to-body dose ratio was 551 (range 223–732). The maximal non-CNS uptake was seen in the stomach ranging from 1.1% to 3.3% of the delivered dose (radiation dose 1.8–3.3 cGy/mCi). Post-treatment MRI showed marked peri-infusional edema at the 3 month post-treatment MRI. This edema responded to dexamethasone therapy.
Cotara has been delivered safely via an intratumoral infusion directly into the brain with no DLT identified. We are presently enrolling patients at the 2.0 mCi/cc dose and expect to progress to the final dose of 2.5 mCi/cc.
We evaluated the survival probabilities in newly diagnosed WHO grade III glioma patient risk groups classified by the prognostic factors using recursive partitioning analysis (RPA).
A total of 132 patients of anaplastic astrocytoma (AA, n = 55), anaplastic oligodendroglioma (AO, n = 67) and anaplastic oligoastrocytoma (AOA, n = 10) were included in this study. All patients were treated with either radiotherapy or radiotherapy followed by PCV chemotherapy. Five possible prognostic factors such as histological diagnosis, age, performance status, extent of resection, and treatment were incorporated into RPA. Risk groups were defined by RPA and survival analysis performed to validate their prognostic values.
Four risk groups were defined by RPA. The lowest risk group was patients who were treated with radiotherapy followed by PCV after gross total resection (GTR) or patients who were treated with radiotherapy followed by PCV after non-GTR but had performance status ECOG grade 0. The low risk group was AO/AOA patients with ECOG grade > 0 who were treated with radiotherapy followed by PCV after non-GTR or patients with ECOG grade < 2 who were treated with radiotherapy only after GTR. The high risk group was AA patients with ECOG grade > 0 who were treated with radiotherapy followed by PCV after non-GTR or AO/AOA patients with ECOG grade < 2 who were treated with radiotherapy only after non-GTR. The highest risk group was AA patients with ECOG grade < 2 who were treated with radiotherapy only after non-GTR or patients with ECOG grade ≥ 2 who were treated with radiotherapy only. Survival analysis showed significant differences in mean survival between groups: 163.4 months (95% CI: 144.9–182.0), 109.5 months (86.7–132.4), 66.6 months (50.8–82.4), and 27.7 months (16.3–39.0) respectively from lowest risk to highest risk groups (p = 0.00).
The present study shows that the RPA grouping can successfully predict the survival probabilities in patients with WHO grade III glioma.
Tumor-methylation of the MGMT-promoter (MGMT-meth) is associated with a better prognosis in patients with glioblastoma multiforme (GBM) receiving standard treatment. Whether the degree of DNA methylation in the tumor tissue correlates with survival is however unclear. We developed a methylation-specific, primer extension based, PCR method to quantitatively assess the MGMT-meth status in a consecutive series of patients with GBM diagnosed and treated until end of 2007. We analyzed the impact of age, surgery (versus biopsy), MGMT-meth and LOH 1p/19q status on overall survival. To date, the data are available for 57 out of 85 patients (analysis ongoing). After a mean F/U of 14.9 months, 61% of patients had died. Only three patients had evidence of LOH 1p/19q (5%); 58% of patients did not show any evidence of MGMT-meth, 12% showed a >90% methylation, and the remaining 30% of the cases showed intermediate proportions of MGMT-meth (range: 18%–89%). Age (<50 versus >50 years), surgery (versus biopsy), and MGMT-meth status (any MGMT-meth versus none) were associated with overall survival in the Kaplan-Meyer analysis. There was no linear correlation between the amount of MGMT-methylation and survival in these preliminary data. Patients with GBM with detectable levels of MGMT-meth as low as 18% have a survival advantage compared to patients without detectable MGMT-promoter methylation.
The DNA repair enzyme O6-methylguanine-DNA-methyltransferase (MGMT) is known to be a key factor in limiting therapy success with temozolomide (TMZ) in glioblastoma (GBM) patients. In this study we have analyzed MGMT promoter methylation and protein expression in GBM primary cell cultures with respect to their prognostic and predictive quality. One hundred and six patients who underwent surgery at our department were included in this study. Patients had GBM (n = 101) and gliosarcoma (n = 5) according to WHO criteria. Subsequent to surgery patients received combined radiochemotherapy with TMZ (n = 68) or radiotherapy alone (n = 17). Twenty-one patients did not receive further therapy due to low performance status (11/21) or rapid worsening of performance status (10/21). Tumor-derived primary cell cultures were analyzed with respect to MGMT promoter methylation (n = 89) by methylation specific PCR following DNA modification and protein expression (n = 71) by Western blot. Prognostic and predictive significance of these two MGMT parameters was assessed using Cox models. Methylated MGMT gene promoter sequences were detected in 67% (60/89) of the investigated GBM samples. Out of the 60 cases with methylated sequences, 14 (23%) lacked any unmethylated DNA while the remaining 46 samples contained a mixed profile with varying proportions of methylated and unmethylated sequences. In 33% (29/89) GBM patients the MGMT promoter was completely unmethylated. Despite a significant association between MGMT promoter methylation status and protein expression (chi-square test; p = 0.002) protein expression levels according to promoter methylation were strongly overlapping. With respect to overall survival (OS), patients with lack of MGMT expression lived significantly longer while promoter methylation showed no significant association. Moreover, MGMT expression, but not promoter methylation, was found to be strongly associated with TMZ therapy response predicting significantly shorter OS in TMZ-treated (p = 0.003) versus untreated patient group. Summarizing the data we demonstrate, that epigenetic silencing via promoter methylation is an important but not exclusive mechanism regulating MGMT expression. Lack of MGMT protein expression but not promoter methylation is of superior quality as predictive factor for TMZ response of GBM patients. We conclude that MGMT expression, additionally to promoter methylation, should be considered as biomarker for GBM patients best-placed to benefit from TMZ-therapy.
Epigenetic silencing of O6-methylguanine-DNA methyltransferase (MGMT) protein expression due to MGMT gene promoter methylation is considered to associate with a benefit of postoperative chemoradiotherapy in glioblastoma multiforme (GBM) patients. The aim of this study was to test the usability of MGMT immunohistochemistry (IHC) as clinical biomarker.
We performed anti-MGMT immunostaining of a tissue microarray containing tissue samples of 164 GBM patients from the prospective, randomized EORTC/NCIC trial 26981/22981. Two commercial anti-MGMT antibodies (clones MT3.1 and MT23.2) were used. Immunostaining results were semiquantitatively evaluated by 4 observers from 3 laboratories using a predefined algorithm. We analyzed (a) inter- and intraobserver agreement on MGMT protein expression (kappa statistics), (b) correlation of MGMT protein expression with MGMT gene promoter methylation status (kappa statistics), and (c) association of MGMT protein expression with patient survival (log rank test).
Interobserver agreement on MGMT immunohisochemistry varied from slight to almost perfect and interobserver agreement ranged from substantial to almost perfect. MGMT protein expression showed poor to moderate correlation with MGMT gene promoter methylation status. We found no significant correlation of MGMT protein expression with patient surrvival.
In glioblastoma, observer variability, lack of association with MGMT promoter methylation status and patient survival limit the usefulness of anti-MGMT immunohistochemistry as clinical biomarker for routine diagnostic purposes.
Brain Pathol. 2008
A high TIMP-1 level has earlier been shown to be related to shorter survival in colorectal cancer and breast cancer suggesting that TIMP-1 might be an important prognostic biomarker. It has been suggested that TIMP-1 is involved in tumor invasion, proliferation, and apoptosis in different types of cancers. Whether TIMP-1 is a prognostic marker in gliomas has not yet been investigated, although TIMP-1 is known to be expressed in gliomas. The aim of the present study was to investigate the immunohistochemical expression of TIMP-1 in gliomas and relate this expression to patient survival. We used a new TIMP-1 monoclonal antibody (VT-7) on formalin-fixed paraffin-embedded tumor tissue from 23 diffuse astrocytomas, 17 anaplastic astrocytomas, and 72 glioblastomas. The staining of tumor cells and vessels were scored and compared with patient overall survival. Moreover, TIMP-1 in situ hybridization was performed on selected glioblastomas. The results showed that TIMP-1 protein was expressed in the vast majority of tumors being expressed in both tumor cells and vessels and in both central and invasive areas of the tumors. In situ hybridization for TIMP-1 mRNA on glioblastomas confirmed the expression of TIMP-1 protein. The fraction of stained tumor cells and vessels as well as the staining intensity varied between tumors of the same grade, but the mean staining score increased with tumor grade. The overall survival of glioblastoma patients was divided into four groups based on TIMP-1 score (0–5, 6–7, 8–9 or 10–12). The multivariate Cox regression test showed that patients with the lowest TIMP-1 expression (score 0–5; n = 12) had a significantly longer overall survival (p = 0.004). There was no significant correlation between TIMP-1 expression and overall survival for diffuse and anaplastic astrocytomas. In conclusion, the study shows that TIMP-1 immunohistochemical score is significantly associated with overall survival in glioblastoma patients, suggesting that TIMP-1 could be used as a new biomarker for this type of cancer. The shorter survival of patients with high TIMP-1 protein expression may be explained by the anti-apoptotic effect of TIMP-1 preventing apoptosis induced by radiation and chemotherapy. Accordingly, TIMP-1 or TIMP-1 interacting proteins could represent new targets in future anti-cancer therapy.
As many as 20% of meningiomas display clinically aggressive features, leading to increased patient morbidity or mortality. NDRG2 promoter hypermethylation has been associated with aggressive behavior in patients with meningioma.
The aim of the present study was to evaluate the prognostic significance of NDRG2 promoter hypermethylation in patients with atypical and anaplastic meningioma.
Thirty-two patients with the diagnosis of WHO grade II and III meningioma (20 atypical and 12 anaplastic meningiomas) were included. Tumor DNA was extracted from microdissected formalin-fixed, paraffin-embedded sections (n = 21) or frozen tissue when available (n = 11). NDRG2 promoter methylation analysis was performed using sequencing of sodium bisulfite-modified DNA covering the 15 CpG sites frequently methylated in advanced meningiomas. Each CpG site was scored as strongly (3), moderately (2), weakly (1), and not methylated (blank) on the basis of the bisulfite sequencing results. For each tumor, a methylation score was calculated by adding the figures determined at the individual CpG sites. A second score was calculated by adding the figures obtained for the six CpG sites (F5/R7 region) immediately 5′ to the promoter core region of the NDRG2 gene.
In the univariate analysis, factors influencing relapse-free survival were methylation of the F5/R7 region of the NDRG2 promoter and KPS. The multivariate analysis revealed that KPS ≥70 (HR: 0.27; 95% CI: 0.083–0.913; p = 0.035) was the only variable independently associated with longer relapse-free survival. NDRG2 promoter methylation (p = 0.038), methylation of the F5/R7 region of the NDRG2 promoter (p = 0.023) and KPS (p = 0.033) were identified as variables with prognostic influence on overall survival in the univariate analysis. Cox’s regression analysis showed that F5/R7 region of the NDRG2 promoter hypermethylation was the only variable independently associated with overall survival (HR: 0.238; 95% CI: 0.286–0.94; p = 0.040).
Hypermethylation of NDRG2 promoter, especially within the F5/R7 region, was independently associated with shorter overall survival in patients with atypical and anaplastic meningioma.
(1) to compare the health-related quality of life (HRQoL) of long-term to short-term high-grade glioma (HGG) survivors, (2) to determine the prognostic value of HRQoL for overall survival, and (3) to determine the effect of tumor recurrence on HRQoL of long-term survivors.
Following baseline assessment (after surgery, before radiotherapy), self-perceived HRQoL (MOS SF-36) and brain tumor-specific symptoms (BCM-20) were assessed every 4 months up till 16 months after histological diagnosis. Kaplan-Meier survival analysis and the Cox proportional hazards model were performed to estimate overall survival of patients with impaired scores on the aggregated SF-36 higher-order summary scores measuring physical functioning (PCS) and mental functioning (MCS).
Sixteen patients with a short-term survival (baseline and 4-month follow-up) and 16 with a long-term survival (follow-up up till 16 months after diagnosis) were selected out of 68 initially recruited HGG patients. At baseline, the short-term and long-term survivors did not differ in their HRQoL. Between baseline and 4-month follow-up, HRQoL of short-term survivors deteriorated, whereas the long-term survivors improved to a level comparable to healthy controls. Patients with impaired mental functioning (MCS) at baseline had a shorter median survival than patients with a normal functioning. Accounting for differences in patient and tumor characteristics, it was not independently related to poorer overall survival. Not surprisingly, in the group of long-term survivors, the 5 patients with recurrence had a more compromised HRQoL at 16-month follow-up compared to the 11 patients without recurrence.
(1) Baseline HRQoL is not related to duration of survival. (2) Long-term survivors show improvement of HRQoL to a level comparable to that of the healthy.
Patients with (low-grade) gliomas not only suffer from epileptic seizures and neurological deficits, but also from cognitive deficits, including memory loss and problems with concentration. These cognitive deficits can have a negative impact on daily functioning and quality of life. This randomized controlled study evaluated the effects of a multifaceted cognitive rehabilitation program on cognitive functioning and aspects of quality of life in glioma patients. 140 adult patients with gliomas and favorable prognostic factors, whose disease was in remission, were recruited from 10 hospitals in the Netherlands. Patients with both subjective cognitive symptoms and objective cognitive deficits were randomized to either an intervention group or to a waiting-list control group. The intervention incorporated both computer-based attention retraining and individual teaching of compensatory strategies directed toward attention, memory, and executive functioning. The program consisted of six weekly individual two-hour sessions plus several hours of homework per week. Both study groups completed neuropsychological (NP) tests and questionnaires on cognitive functioning, fatigue, and quality of life at baseline, immediately following completion of cognitive rehabilitation (or at an equivalent time point for the control group), and at six months follow-up. Preliminary results (repeated measures [M]ANCOVA’s) indicate significant effects at the post-treatment assessment on measures of self-reported cognitive functioning, burden of cognitive dysfunction, and mental fatigue favoring the intervention group. In objective NP test scores, no significant group differences at post-treatment were observed. At 6-month follow-up, however, significant intervention effects were found on NP tests of attention and verbal learning. Changes in executive functioning did not reach statistical significance. The effect on self-reported outcomes was less pronounced at this 6-month assessment. Cognitive rehabilitation of glioma patients has immediate positive effects on self-reported cognitive outcomes and positive effects on objectively assessed NP functioning over a longer period of time.
Treatment of gliomas WHO III and IV is based on either radiation (RT) or chemotherapy (CT) and may cause an endocrine deficiency in women of reproductive age. We tried to evaluate hormonal changes and their impact on quality of life in women with malignant gliomas.
We examined the hormonal status of 22 female patients with high-grade gliomas (aged 28 to 48 years) after whole brain RT and during CT. The cumulative dose of radiation ranged from 50 to 66 Gy and all received first line CT, most frequently temozolomide or the combination of fotemustine with dacarbazine, nine of them had second line therapy and two were under third line treatment. Serum concentrations of gonadal, pituitary, and thyroid hormones were measured at the start and the end of chemotherapy. Vaginal ultrasound was performed within the fist cycle of CT and patient quality of life was assessed by questionnaires EORTC QLQ-30.
19 patients (86%) showed change in menstrual pattern. Ten had an amenorrhoea with a mean duration of 26.1 months (ranged from 4 to 96 months), and 6 presented with oligomenorrhea or hypomenorrhea. Four females showed low hemoglobin levels, without relation to CT, and all of them reported hypermenorrhea at variable intervals. Serum concentration of prolactin was elevated in 12 patients, and 11 of them presented with amenorrhoea. Sixteen of the pre-menopausal women (72%) had low level of estradiol and 8 (36%) had low progesterone serum concentrations. Overall, 10 patients had evidence of hypogonadism and a postmenopausal hormone pattern. Two presented with a latent hypothyroidism, and four had lower serum concentrations of TSH, total thyroxine, or total triiodothyronine than in normal subjects. The QLQ evaluation showed that patients frequently reported about fatigue and described feelings of postmenopausal women, like flushes, weakness, and gain of weight. Among 22 patients, 12 had uneventful pregnancies (ranged 1 to 3) before diagnosis of a malignant glioma, and none of them expressed a wish to preserve their childbearing potential or to become pregnant. The decrease of libido combined with loss of the feeling to be attractive as a female, with a need for tender care and security on the other hand, led to problems in their partnerships.
The hormonal deficits caused by RT and chemotherapy need a careful evaluation as well as individualized measures of diagnostics and treatment.
Depression is often under-diagnosed and under-treated in cancer. Depressed glioma patients may present both a diagnostic challenge (due to uncertainty over the significance of understandable symptoms) and a treatment challenge (due to a risk of antidepressant-induced seizures). Little is known of the current attitudes of clinicians in this area.
To explore attitudes among five specialty groups towards (1) diagnosing and (2) treating depression, in a patient with glioma and a history of seizures.
We created a clinical case study of a patient with a glioma and previous seizures, which met 7/9 criteria for DSM-IV Major Depressive Disorder. We sent it, with a questionnaire, to consultants in neurology, clinical oncology, neurosurgery, and liaison psychiatry in Scotland (UK) and to the GPs of all glioma patients on our local database (total n = 228).
Responses were received from 105 doctors (46% response rate). Sixty-five diagnosed depression (62%); the remainder considered the presentation an “understandable reaction.” Forty of those identifying depression chose an antidepressant as first-line treatment (62%). No statistically significant differences were found between specialties in respect of diagnosing depression (p = 0.16) or readiness to prescribe an antidepressant (p = 0.11). GPs and psychiatrists were more likely to consider themselves best-placed to lead the management of depression (both p = <0.00001). Among respondents identifying depression, concern about seizures was higher in both GPs (p = 0.006) and psychiatrists (p = 0.03) than neurologists. Across the whole sample, higher seizure anxiety was significantly associated with lower anti-depressant prescribing rates (p = 0.04).
Almost four in ten respondents did not identify major depression in a case study of a depressed patient with a glioma. There was also poor agreement on how to treat it, with almost four in ten of those diagnosing depression opting for first-line management other than an anti-depressant. In the whole sample, only 38% of respondents would have diagnosed and appropriately treated the depression. Since GPs and psychiatrists felt significantly more confident about leading management of depression than neuroscientific specialties, and significantly less confident about managing seizures than neuroscientists, multidisciplinary management may be beneficial for these patients. However as some analyses were exploratory, results should be interpreted with caution.
Craniospinal radiation (CSI) is the cornerstone of the treatment of medulloblastoma, but may damage hypothalamic pituitary function. There are many publications reporting endocrinological toxicity of CSI in children, but little is published on adults.
To study the prevalence of pituitary deficits, following CSI in children (<15 years at radiotherapy) and adults.
A series of 30 patients (16 children, 14 adults) treated by surgery, radiotherapy, and chemotherapy for a medulloblastoma was retrospectively studied. The children group was composed of 5 pubescent girls and 11 boys (3 pubescent)). Age at radiotherapy ranged between 5 and 12 years. Delay between CSI and evaluation was 1–9 years. Adult patients were 4 women (3 menopausal) and 10 men. Age at radiotherapy ranged between 15 and 40 years. Delay between CSI and evaluation was was 1–24 years. GH and adrenal axes were assessed using stimulation tests (insuline tolerance, glucagon-propanolol, and synacthene test). Gonadotropin and thyrotropin functions were tested using baseline blood measurements.
Hypopituitarism was present in 57.2% of adult patients and 87.5% of children. The most frequent deficit was GH deficiency, respectively observed in 35.7% and 62.5 % and significantly associated with the young age of radiotherapy (p< 0.05). ACTH deficiency was rare, observed in 7.1% of adults and 12.5% of children. Hypothyroidism from central or peripheral origin was recorded in 21.4% of adults and 31.2% of children. Finally, gonadotropin deficiency was observed in 35.7% of adults and 6.2% of children. Two children have developed a peripheral gonadal insufficiency.
Both adult and children treated with CSI for a medulloblastoma are at high risk of hypopituitarism (especially GH deficiency) which depends on age at time of RT. Long term surveillance and periodic evaluation are needed in children and adults.
Meningioma patients are the most prevalent brain tumor survivors. Despite their good prognosis, surprisingly little is known about the long-term health-related quality of life (HRQOL) sequelae. In this study we investigated HRQOL in WHO grade I meningioma patients following treatment and the impact of cognitive deficits, epileptic seizures, and anti-epileptic drug (AED) use.
Eighty-nine WHO grade I meningioma patients who underwent surgery with or without adjuvant radiotherapy were individually matched to 89 healthy controls for age, sex, and educational level. We assessed HRQOL by means of the SF-36 questionnaire. We measured cognitive functioning through a neuropsychological test battery and assessed epileptic seizure frequency and AED use of each patient. Student’s t-tests were performed to test for differences in HRQOL between patients and healthy controls, between patients with or without AED use, and between patients with or without seizures. By using stepwise linear regression analysis, we additionally correlated HRQOL with cognition, corrected for age, sex, and educational level.
Except for role limitations caused by physical problems (p = 0.015) meningioma patients did not differ from healthy controls in any of the 8 HRQOL scales. However, the 23 patients using AEDs had significantly lower HRQOL scores in 5 out of the 8 SF-36 scales when compared with healthy controls. Analysis of the patients using AEDs showed no significant differences in HRQOL between patients with or without epileptic seizures. Impaired cognition had also a negative impact on HRQOL; regression analysis demonstrated impaired executive functioning to be correlated with lower scores on 7 out of 8 HRQOL scales (p<0.05). Furthermore, deficits in other cognitive domains, except for working memory, also negatively affected quality of life, albeit to a lesser extent.
Health-related quality of life is unaffected in most meningioma patients. Subgroups of patients with cognitive deficits and/or patients using AEDs, however, have a decreased HRQOL. Incomplete seizure control among patients using AEDs does not further deteriorate their HRQOL.
MB is a rapidly proliferating tumor arising in the posterior fossa and seen mainly in children. Standard therapy includes craniospinal radiotherapy (CSRT) and adjuvant chemotherapy. Patients presenting with leptomeningeal metastases (Chang Stage M1–3) have a poor outcome and current research is aimed at improving treatment for these patients, including the investigation of novel therapies. A HART regimen could improve the therapeutic ratio for RT by the use of acceleration in addition to hyperfractionation and thus minimize tumor cell repopulation during RT. The purpose of this study was to evaluate the feasibility and toxicity of a regimen comprising HART (1.24 Gy twice daily) with concomitant and adjuvant chemotherapy for metastatic MB. Between February 2002 and October 2007, 25 eligible patients (16 male, 9 female) aged 3–13 (median 7, mean 7.8) with M1–3 medulloblastoma were entered into the study. Initial management was with maximum surgical resection of the primary, followed by HART. Treatment then continued with 8 cycles of adjuvant chemotherapy. The CSRT dose was 39.68 Gy in 32 fractions, followed by a boost to the whole posterior fossa of 22.32 Gy in 18 fractions. Where appropriate, boosts to metastases of 9.92 Gy in 8 fractions were given. Adjuvant chemotherapy consisted of Vincristine (VCR) 1.5 mg/m2, CCNU 75 mg/m2, and Cisplatin 70 mg/m2. The first 7 patients did not receive VCR concurrent with RT. The eighth and subsequent patients received VCR 1.5 mg/m2 weekly × 8 doses starting during the first week of RT. Of 25 patients entered into the study, 23 have completed HART. One patient had an interruption to HART because of an intercurrent infection. HART has been completed with a median duration of 34 days (range: 32–38). The numbers of patients experiencing grade 3–4 toxicity include anaemia: 3, thrombocytopenia: 2, leucopenia: 14, mucositis: 3, radiation dermatitis: 3, nausea: 5, and vomiting: 2. Of the 14 patients who have completed all treatment, 8 received all 8 cycles of chemotherapy, 4 have received 7 cycles, 1 has received 5 cycles and one received no chemotherapy. To date, 6/25 patients have relapsed and 5 have died following relapse. Data on event-free survival will be presented. In conclusion the HART regimen is tolerable and could be delivered in all patients without any major delay, and can be incorporated into multimodality management including chemotherapy.
Prognosis of disseminated medulloblastomas (DMB) is poor and treatment remains a major challenge in young children because of the late toxicity of standard dose craniospinal irradiation (CSI). We developed therefore a new strategy based on sequential high-dose chemotherapy (HDC) followed by autologous stem cell transplantation (ASCT) and reduced dose CSI.
Thirty-four children were enrolled in this phase II study. Median age at diagnosis was 3.2 years (range, 0.8 to 7). Chang stages were M1 (6 pts) and M3 (28 pts). After initial surgery, children received two courses of carboplatin-etoposide; then peripheral blood stem cells were harvested. In absence of disease progression, five sequential courses of chemotherapy followed by ASCT were performed: melphalan (100 mg/m2) at day 43, cisplatin (100 mg/m2) at day 64, melphalan at day 85, cisplatin at day 106 and thiotepa (720 mg/m2) at day 127. Children with resectable tumor residue were operated at the end of chemotherapy. Treatment was completed by age-adapted CSI.
At the end of the chemotherapy 11 pts were in CR, 7 in PR. Out of the 19 pts who completed the treatment, 13 children are alive in CR1 (median follow-up, 26 months). Delivered doses of CSI were 50 Gy to the posterior fossa and 18 Gy to the brain and spinal canal in 3 pts, 50 Gy and 24 Gy in 6, and 50 Gy and 35 Gy in 1. The overall survival is 50% at 30 months. The haematological toxicity was high but manageable; no toxic death was observed.
This chemotherapy regimen combined with reduced dose CSI produced a survival that compares favorably with published data. A longer follow-up is necessary to assess long-term quality of life.
To review clinical and histological characteristics of patients diagnosed with meduloblastoma and evaluate their prognostic strength.
Retrospective study, including 51 patients diagnosed with medulloblastoma from 1982 to 2004. Clinical and Histological variables were analyzed, including histological variants, mitotic activity, microvascular proliferation, pseudopalisading necrosis, and neuroblastic rosettes. Additionally, a tissue microarray including synaptophysin, neuronal specific enolase (NSE), glial fibrilar acidic protein (GFAP), neurofilaments-68 (NF68), epitelial membrane antigen (EMA), vimentin, Wilms tumor protein 1 (WT1), c-erbB2, p53 protein sobreexpression (p53), INI-1 and ki67 expression was performed.
The average age at diagnosis was 5.6 years old. The overall survival (OS) was 37.5%, although survival improved steadily in recent cohorts. According to the histological parameters, the desmoplastic and the extensive nodularity were variants with better OS, whereas tumors with more than 20 mitosis/10HPF were associated with a worse prognosis and low OS. IHC analysis showed synaptophysin expression in 76.6% of cases, NSE 100%, GFAP 31.9%, NF68 36.2%, EMA 8%, vimentin 34%, WT1 (cytoplasmic) 25.6%, cerbB-2 4.4%, p53 6.7%, and INI-1 93.5%. Ki67 expression was >50% in 23.5% of cases, 20–50% in 41.2% and <20% in 35.3% of cases. It is remarkable the lack of INI-1 expression in 3 tumors previously diagnosed as meduloblastoma, which suggested the diagnosis of Atypical Teratoid/Rhabdoid tumor (AT/RT). This was further confirmed by hSNF5/INI1 gene mutation.
Desmoplastic and extensive nodularity subtypes correlate with better prognosis; likewise, less than 20 mitosis/10HPF. AT/RT show similar histology and IHC markers as medulloblastoma, but INI-1 negative expression identify AT/RT, being essential to perform IHC for INI-1 in pediatric embryonal CNS tumors to disclose a possible AT/RT.
To evaluate the clinical features and outcome (symptomatology of the onset and long term sequelae) in intracranial germ cell tumors (GCTs) and the results of the multimodal treatment.
We retrospectively analyzed 24 cases treated in our institute between 1995– 2005. Patients’ characteristics: overt predominance of male cases (80%); age limits: 8 and 30 years old; the most frequent sites: sellar and pineal region (70%). The therapeutic approach was varied in the studied group. The multidisciplinary treatment consisted in: surgery (S)+ radiation therapy (RT)+ chemotherapy (ChT) in 40% of patients (pts); S+ RT in 31% of pts; RT+ ChT in 4 pts. Surgery was performed in 20 pts: subtotal and cvasitotal resections (66 %) and biopsies (33%). 70% of pts underwent RT on different fields (cranial, localized, or craniospinal) and in varied doses (2,000–5,600 cGy). Starting with 2002, the therapy approach was unitary and adapted to the international standard SIOP-CNS-GCT 96: limited S with diagnostic purpose + ChT (2–4 cycles) + RT (cranial field, 2,400–3,000 cGy in complete response or 3,000–3,600 cGy in partial response, + spinal irradiation in cases with malignant cells in cerebrospinal fluid). According to the same international standard, the definitive diagnosis was accepted without biopsy in pts with characteristic radiological aspects correlated with high-levels of tumor markers (CEA, AFP) in serum and/or cerebrospinal fluid. ChT comprised two types of protocols: BEP and ETO+ IFO+ DDP.
70% of cases were registered in patients aged >10 years, 50% of pts were between 15–25 years old. The predominant symptoms: headaches: 75% of pts, visual disturbances (diplopya, strabismus, poor vision): 75% of pts, and diabetes insipidus: 40%. Increased intracranial pressure urged the installing of ventriculo-peritoneal shunts in 11 pts. The long-term sequelae were endocrine disturbances: diabetes insipidus (7 pts), global hypophysis impairment (1 pt), hypothyroidism (1 pt); ophthalmologic impairments (8 pts); equilibrium deficiencies (2 pts); and hearing disturbances (1 pt).
Overall survival evaluated by Kaplan-Meier method: 84%. Median survival: 137 months. We registered recurrences in 5 pts and cerebral and/or bone marrow metastases in 3 pts. DFS: 20 months. In 4 cases complete remission of tumors after 4 cycles of pre-irradiation chemotherapy was registred. In 6 cases partial remission of tumors (>75%) was registred after pre-irradiation ChT.
(1) The current multidisciplinary therapeutic approach entails a minimize of the late side-effects by reducing the cranial irradiation efficient dose. (2) The surgical approach limited to a diagnostic purpose obviate important endocrine and ophthalmologic long-term sequelae. (3) Pre-irradiation ChT have in important impact for long term complete remission of tumors.
Bevacizumab (BV), a monoclonal antibody against vascular endothelial growth factor (VEGF), has proven anti-cancer activity in a number of solid tumors. Regarding intracranial tumors concern exists with respect to intratumoral hemorrhage associated with BV. We report the tolerability and feasibility of a compassionate use BV treatment in 16 patients with intracranial tumors.
We retrospectively analyzed 16 children (11 male, 5 female) with intracranial tumors. The diagnoses were glioblastoma WHO°IV (n = 1), diffuse intrinsic pons glioma (n = 1), medulloblastoma (n = 2), supratentorial primitive neuroectodermal tumor (n = 2), anaplastic astrocytoma WHO°III (n = 1), anaplastic ependymoma WHO°III (n = 1), atypical teratoid rhabdoid tumor (n = 1), CNS germ cell tumor (n = 1), pilomyxoid astrocytoma (n = 2), pilocytic astrocytoma (n = 2), clear cell meningeoma (n = 1), and metastatic alveolar soft part sarcoma (n = 1). Except for one patient (no parental consent for chemotherapy) all patients received BV for recurrence/progression of their disease. The median age at start of BV treatment was 9.6 years (range 1.5–18 years). The dose ranged from 5–10 mg/kg (median 8.18 mg) and was administered intravenously every 2–4 weeks. A total number of 255 BV courses was administered (median per patient: n = 16, range 2–46). Except for one patient all patients received additional chemotherapy or oral four-drug antiangiogenic metronomic therapy concomitant to BV.
BV was well tolerated in most of our patients. We observed mild BV related side effects in 5/16 children (mild proteinuria: n = 2, mild hypertension: n = 2, fatigue: n = 1, epistaxis: n = 1). No thrombosis or hemorrhage occurred, although all patients presented with a large tumor ranging in size from 2.5–8 cm in diameter. In one patient treatment had to be stopped due to hypertension, requiring antihypertensive treatment for 4 months. Another patient showed moderate proteinuria, and in one child wound healing was impaired.
In contrast to reports in adult patients with high-grade gliomas no tumor hemorrhage was observed in any of our 16 pediatric patients. However, further prospective studies including more patients are warranted.
To present outcome data and response evaluation in a prospective study of concurrent and adjuvant temozolomide (TMZ) with radiotherapy (RT) in children with diffuse intrinsic brain stem gliomas (DIBSG).
Pediatric patients were prospectively treated with RT to a dose of 54 Gy/30 fractions and concurrent daily temozolomide TMZ (75 mg/m2). After RT, adjuvant TMZ (250 mg/m2) was given every 28 days up to a maximum of 12 cycles. Response was evaluated clinically and with MRI and FDG-PET scans.
Between March 2005 and November 2006, 20 newly diagnosed DIBSG patients (15 boys, 5 girls, median age 8 years) were accrued. At baseline pre-RT evaluation, 6 patients had no contrast enhancement on MRI. MR perfusion scan showed hyperperfusion on 13 and hypoperfusion on 7 patients. On MR Spectroscopy (MRS), 10 patients each had features of low-grade and high-grade gliomas. Among 11 patients with baseline PET scans, increased uptake was seen in 6. All 20 patients completed concurrent chemo-RT. One patient died at completion of RT. 8 patients completed 6, and 3 patients 12 cycles of adjuvant TMZ. Mean concurrent and adjuvant dose of TMZ was 65 mg and 174 mg, respectively. In concurrent RT+TMZ, thrombocytopenia of Gr-III/IV was seen in 3 and leukopenia in 2 patients. Gr-III/IV vomiting was seen in 07 patients requiring medication. Most patients (18) had grade-II skin toxicity. In adjuvant TMZ, 2 patients developed Gr-III/IV leukopenia and 1 patient Gr-IV thrombocytopenia. Mean overall survival (OS) was 10.39 months (range 1.7–22 months) and progression free survival (PFS) was 9 months (range 1.7–22.6 months). 16 patients lived for more than 6 months, 7 1 year, and 2 for 18 months. At last follow up, 17 patients (85%) have died due to progression and 3 patients (15%) are alive with stable disease. MRI diagnosis of HGG (p = 0.001) had worse survival than low-grade glioma. MR perfusion showed significant correlation with survival (p = 0.043). However, MRS (p = 0.258) and PET scan uptake (p = 0.911) of high-grade tumor did not show correlation with survival. Patients with neurological improvement after RT had significantly better survival (p = 0.048) than those who did not.
Temozolomide with RT has not yielded any improvement in outcome of DIBSG than already known. Patients diagnosed with hyperperfusion on MR have significantly worse survival. FDG-PET was poor in diagnosis of the grade of the tumor and response to treatment.
The frequency of pediatric brain tumors is different from that of adults regarding the histopathology of tumors and their management. This study was conducted to reveal the most frequent pediatric tumors that were surgically treated in our hospital and discusses their management.
The authors retrospectively investigated 55 child patients with histologically confirmed intracranial tumors (except one case of pons glioma which was diagnosed radiologically), who were operated on at Göztepe Education and Research Hospital in Istanbul between 1998 and 2008.
The mean age of the patients (21 males and 34 females) was 8 years (range: 3–16 years). Of the total 55 cases, 31 cases were supratentorial and 24 were infratentorial. The most frequent tumors were astrocytomas (21 patients), medulloblastomas (15 patients) and pituitary adenomas (7 patients). The less frequent tumors were meningioma, ependymoma, craniopharyngioma, dermoid cyst, and mesenchymal sarcoma. Surgery was performed in 54 patients. Early post-operative mortality was seen in 2 cases. In 7 patients, there were focal motor and/or cranial nerve deficits after surgery. However, they recovered functionally although tracheotomy and gastrostomy were necessary temporarily in these patients.. Twelve patients received adjuvant chemotherapy. One patient with a pons glioma received radiotherapy only. 20 out of total 55 cases were treated between 1998 and 2002, whereas 35 cases out of total 55 cases were treated between 2003 and 2008.
The most frequent pediatric intracranial tumor was astrocytoma in our series. There was an increase in the number of patients in the second half of the 10-year period in this study. Postoperative morbidity included cranial nerve deficits and focal motor deficits.
Despite standard treatment, malignant glioma has a very poor prognosis. Newer strategies incorporating local or systemic chemotherapy have been developed, and increasing evidence in the literature shows that these multimodal approaches could be of the increased benefit in the management of these patients. We present a case of long-term survival in a pediatric patient with GBM treated with a multimodal strategy including exeresis, local and subsequent oral chemotherapy, and re-treatment after relapse.
A 6-year-old male patient with a posterior temporal lobe lesion was admitted to our institution. Posterior temporal craniotomy and subtotal resection were performed, and a diagnosis of GBM was made. Two weeks later the patient was re-operated and intrasurgical, corticographically-guided, macroscopically complete exeresis was performed, during which 8 BCNU wafers were implanted. There were no relevant postsurgical complications and the patient began receiving fractionated radiotherapy (total, 50 Gy) with concomitant TMZ (60 mg/m2/day, days 1–5, every 28 days). The patient continued TMZ treatment for 18 months without incident, until a scheduled MRI scan revealed a local relapse. The patient underwent complete resection and histopathology confirmed GBM relapse; 4 BCNU wafers were implanted into the resection cavity. There were no relevant postsurgical complications, and the patient re-initiated TMZ for 24 months without disease progression. After 18 months of follow-up post-TMZ, and 5 years after the initial diagnosis the patient remains disease-free.
Multimodal therapy is an option in the management of GBM. Our experience in this case shows that multimodal treatment with macroscopically complete resection, BCNU wafers, TMZ, and radiotherapy (for primary disease) both in first surgery and local relapse appears to be well-tolerated. There were no clinically significant complications associated with the multimodal approach. The use of BCNU wafers in combination with TMZ provides continuous anti-tumor activity immediately following resection. This may improve tumor control and provide a survival benefit.
The Ommaya reservoir facilitates repetitive delivery of drugs into the CSF and is a pharmacologically rational system for administering intrathecal chemotherapy. However, previous studies have found a rate of infection and other complications ranging from 4.8 to 19 %.
Between 1992 and 2007, 58 patients aged 4 months to 18 years with various poor prognosis brain tumors received intrathecal chemotherapy via an Ommaya reservoir. All patients received perioperative antibiotic therapy. 2,213 chemotherapy administrations via the reservoir were performed amounting to a median of 35 deliveries per reservoir. Only personnel undergoing a special training were allowed to administer the intrathecal therapy. CSF cell counts and bacterial cultures were obtained at each use of the reservoir. Intravenous antibiotics were used generously in patients with increase of C-reactive protein (CRP) levels to avoid seeding of the reservoir during bacteremia.
There were no cases requiring reoperation for infection or malpositioning of the ventricular catheter. In one patient the trajectory of the intraventricular catheter required reassembly due to kinking at the burr hole edge. 35 (60%) of the patients are surviving a median of 4 ½ years (range 3 mo–13 ½ years) after reservoir placement. No late onset Ommaya reservoir infections occurred in any of the long-term survivors.
Use of Ommaya reservoirs is safe in immunocompromised children with brain tumors if neurosurgeons experienced with the technique are implanting the device, physicians and nurses administering the drug undergo special training, and systemic antibiotics are used perioperatively and when CRP levels increase.
To evaluate survival, quality of life and neurotoxic sequelae in long-term survivors with primary central nervous system lymphoma (PCNSL) after systemic and intraventricular chemotherapy with deferred radiotherapy.
From 09/1995 to 12/2001, 65 patients with PCNSL (median age 62 years) were enrolled in a pilot/phase II study evaluating chemotherapy without radiotherapy; thirty patietns were younger than 60 years. A high-dose methotrexate (MTX) (cycles 1,2,4,5) and cytarabine (ara-C) (cycles 3,6) based systemic therapy (including dexamethasone, vinca-alkaloids, ifosfamide, and cyclophosphamide) was combined with intraventricular MTX, prednisolone, and ara-C. All living patients were contacted, a neurological examination, comprehensive neuropsychological testing, Quality of Life (QoL) Assessment (EORTC QLQ-C30 Version 3.0 and EORTC QLQ-BN20) and imaging were performed.
Twenty-one of 65 patients (32%) are alive. Of these, 17 patients were younger than 60 years at diagnosis, resulting in long-term survival rate of 17 of 30 patients (57%). Median follow up is 99.5 months (range, 70 to 147 months) in surviving patients. Seventeen/21 patients completed all investigations, in three patients test results are pending, one is lost to follow up. One patient is currently undergoing therapy for tumor relapse; one other patient developed a second malignoma (liposarcoma). In three patients, a pure extraneural relapse was diagnosed after 9, 31, and 40 months and showed complete remission to high-dose chemotherapy in all. There was no evidence of late onset neurotoxity nor compromise of therapy- or tumor-related QoL measures in any of the examined patients.
Primary polychemotherapy based on high-dose MTX and ara-C is highly efficient in PCNSL. More than half of patients < 60 years can obviously be cured with this regimen without suffering from long term neurotoxic sequelae.
High-dose methotrexate (HD-MTX)–based chemotherapy with whole brain irradiation improves the prognosis of PCNSL, however up to 30% of patients are refractory to primary therapy and 60% relapse. Fewer than 50% of patients enter a second remission. Thus, novel treatment regimens for relapsed PCNSL are needed. As the majority of PCNSL are B-cell type expressing the CD20 antigen, treatment with the monoclonal antibody (mAb) rituximab might be reasonable. However, delivery of mAb to the brain would be limited by the blood-brain barrier (BBB). Osmotic BBB disruption (BBBD) is currently being used to increase the delivery of chemotherapeutic agents for the treatment of brain tumors. We hypothesize that BBBD would enhance deribery of mAb to the CNS. The activity of rituximab plus ifosfamide, cisplatin, and etopside (R-ICE) as salvage therapy in systemic lymphomas provides the rationale for our use of the R-ICE regimen in conjunction with BBBD. Here we report our experience with i.v. rituximab with BBBD followed by ICE-based chemotherapy in patients with relapsed or refractory PCNSL. Five patients who failed therapy with HD-MTX–based chemotherapy were treated with a regimen of rituximab with BBBD followed by ICE in a 28-day cycle. BBBD was performed with 20% mannitol infused through catheter placed into the artery of the tumor site selectively at a rate of 3ml per second for 30–60 seconds. ICE included ifosfamide 900 mg/m2/d, CDDP 20 mg/m2/d, and etoposide 60 mg/m2/d on days 1–5. Four patients had prior RT and had been pretreated heavily. The mean age was 50 years. Three complete remissions, one partial remission, and one stable disease were achieved. The median progression-free survival from first R-ICE/BBBD was 4 months. The major toxicity was grade 3–4 neutropenia (5/5) and grade 3–4 thrombocytopenia (2/5). Hematotoxicity was prolonged in an elderly patient. The preliminary data presented here suggest that R-ICE with BBBD is an effective second-line treatment in PCNSL and clinical trials are necessary to evaluate both the efficacy and long-term safety of this treatment.
The addition of high-dose methotrexate (MTX) to radiation has improved disease control and survival in patients with primary central nervous system lymphoma (PCNSL). However, more than one half of patients eventually relapse, and uncontrolled PCNSL remains the primary cause of death. We designed a maintenance therapy using rituximab (anti-CD20 antibody) after induction of complete remission to suppress recurrence.
Preradiation chemotherapy consisted of MTX 3.5 g/m2 for a total of three to five doses. Cyclophosphamide, doxorubicin, vincristine, and predonisone (CHOP) were given concomitantly with each cycle of systemic MTX. Chemotherapy was followed by 36–40 Gy of whole brain radiation. Older patients were given the option of deferring radiation after a discussion of potential risks and benefits. Within 3 months after induction of complete remission, patients received multiple courses of rituximab; each course consisted of rituximab 375 mg/m2 given weekly for 4 consecutive weeks. Each course was repeated every 3 months.
Nine patients with a median age of 65 years (range, 55 to 72 years) were included in this study between 2004 and 2007. Six patients were older than 65 years-old. Median follow-up period was 21 months (range, 18 to 38 months). All patients had histologic diagnosis of diffuse large B-cell lymphoma and positive immunohistochemical staining for anti-CD20 antibody. One patient died of recurrent disease, but 8 patients were alive (range, 18 to 38 months) and their Karnofsky performance status (KPS) was kept higher than 80. Eight patients experienced recurrence. The median time-to-recurrence was 15 months (range 6–34 months). No toxicity was seen in association with rituximab.
In recurrent cases, tumor growth is often explosive, which may result in a delay of treatment. In this study, complete remission was achieved in all patients who had recurrence. Maintenance therapy with rituximab did not suppress recurrence rate, however, it might prolong time-to-recurrence and suppress tumor growth rate and eventually contribute to prolong overall survival. In older patients, deferring whole brain radiation by retreatment with high-dose MTX and maintenance with rituximab did not compromise overall survival but did reduce neurotoxicity.
We sought to determine whether there is any differentiating imaging feature between primary (G1) and secondary (G2) central nervous system (CNS) lymphoma. CT, MR imaging including Gd enhanced series, and diffusion-weighted imaging (DWI) findings of 16 patients, of which 10 were primary and 6 were secondary CNS lymphoma were re-evaluated. Pathological subtype of the tumors were diffuse large B cell in all primary and systemic tumors except 2 (1 lymphoblastic and 1 Burkitt type). In all patients, the mass lesions were mildly hyperintense to white matter and hypointense in the background of profound edema on T2W imaging. In 7/16 patients lesions had focal T1 hyperintensities due to hemorrhage (6 in G1; 1 in G2) and the rest were hypointense to white matter and could not be differentiated from accompanying edema. An infiltrating lymphoma of butterfly pattern, i.e., transcallosal bifrontal involvement, were observed in two patients of G2 but none in G1. Two patients were of intravascular type of CNS lymphoma. One had focal hemorraghic masses with profound edema and another patient had enhancing periventricular lesions with stripes of T2-hypointense oriented as perivascular spaces. The lymphomatous lesions were hyperintense on DWI and of low ADC values in 10 of 16 patients (6 in G1, 4 in G2). There was no correlation between CT density values of the lesions and ADC values. Ten of 14 patients who had CT examinations (6/8 in G1, 3/5 in G2) showed increased HU values. In 5 patients a very infiltrative pattern was observed. In a background of profound vasogenic edema scattered T2-hypointense nodules and stripes oriented parallel to medullary veins suggesting perivascular infiltration was present. Four of these cases including one with intravascular subtype were diagnosed as primary CNS lymphoma. These T2-hypointense nodules and stripes enhanced markedly on postcontrast T1-weighted series. Lobulated masses with marked enhancement and fuzzy borders were observed in 5 patients, patchy and perivascular enhancement with small nodules in 5 patients, and enhancement with well-defined borders in 6 patients were determined. In all four patients with infratentorial involvement either brain stem or cerebellum was of primary CNS lymphoma group. The basal ganglia were involved in three patients in G1 and none in G2. As a conclusion, involvement of infratentorial structures and the basal ganglia, hemorrhagic lesions appreciated as T1 hyperintensity and T2-hypointense nodules and stripes probably showing perivenular infiltration and Gd enhancement along the venules have been shown in primary CNS lymphomas in this study.
To evaluate the activity of nitrosourea-based (NU) chemotherapy as salvage treatment in progressive low-grade glioma (LGG) previously treated with temozolomide (TMZ).
Retrospective review of patients treated from 1999 to 2005 in a single institution and who met the following criteria: adults with histologically proven LGG, progressive disease after TMZ treatment, NU based chemotherapy as second line treatment, and absence of prior radiotherapy (RT). Response was assessed by MRI. Progression-free survival (PFS) and overall survival (OS) were measured by Kaplan-Meier methodology.
30 out of 148 patients treated with TMZ as primary treatment during this period fulfilled the criteria. In most cases, NU (PCV or BCNU) as second line treatment was preferred to RT, because of the large size of the tumor and/or the initial chemosensitivity to TMZ. Median age was 46 years (range 30–81). Histological subtypes included 21 pure oligodendrogliomas and 9 diffuse astrocytomas and mixed gliomas. Seventeen tumors demonstrated a mild contrast enhancement on MRI. Three patients achieved an objective minor response (10%); 20 patients had a stable disease (67%) and 7 patients progressed (23%). Toxicity was mainly hematologic with a 17% rate of grade III–IV myelosuppression. Median PFS was 6.5 months (95% CI 3–11.months). Median OS from start of salvage treatment was 23.4 months (95% CI, 13–29 months). Tumors without contrast enhancement were significantly associated with a better prognosis both in terms of PFS (p = 0.0003) and OS from start of NU (p = 0.0006). The 1p/19q codeletion available for 15 patients was not predictive for objective response to salvage treatment but correlated with a better PFS (p = 0.02).
Salvage NU chemotherapy provide disappointing results in TMZ pretreated LGG, which should be treated in priority by conventional radiotherapy.
The role of chemotherapy in low-grade gliomas (LGG) is still under investigation. However, recent studies have reported an interesting activity of temozolomide chemotherapy with standard schedule in patients affected by LGG of different histology treated in different stage of disease. Low-grade gliomas present a peculiar pattern of growth due to their biology: LGG have only a small fraction of proliferating cells, and many authors suggested that a chronic administration of a low dose of chemotherapeutic agents may be more suitable for slow growing tumors respect a high cytotoxic dose. The aim of our protocol of study is to explore the activity of a low dose schedule of temozolomide (50 mg/mq) given for 1 week on and 1 week off, in newly diagnosed LGG requiring treatment for the presence of negative prognostic factors such as residual tumor after surgery or biopsy, age higher than 40 years, neurological deficits, or uncontrolled epilepsy. Previous chemotherapy or signs of anaplastic components, with the presence of enhancing areas at MRI, are considered exclusion criteria. Main objectives include evaluation of activity with therapeutic response rate, PFS at 12 and 24 months, and toxicity. At present, fourteen patients have been enrolled (6 grade II astrocytomas, and 8 oligos or mixed), and 167 cycles of chemotherapy have been delivered (median = 12 courses). 1p–19q co-deletion was present in 9 cases. No hematological or gastro-digestive toxicity more than grade 1 has been reported. 14 patients are evaluable for objective response after first follow-up MRI examination with 4 minor responses (28%), and 10 stable disease. 5 patients (38%) presented partial or complete seizure control during treatment. 7 patients presented early progression (median PFS = 7 months). Considering the high progression rate the study was stopped.
Continuous administration of a low dose of temozolomide shows interesting activity with objective response and clinical benefit, but does not seem to prevent early progression in the treatment of newly diagnosed requiring treatment LGG.
The role of chemotherapy in the management of intracranial ependymomas is still controversial, and few data are available in the literature regarding adult patients. Preclinical studies have shown activity of temozolomide against tumor xenografts derived from ependymoma, and a case is reported of a recurrent intracranial ependymoma in remission 10 years after completing temozolomide chemotherapy.
In this prospective study we evaluated the efficacy and toxicity of temozolomide in patients with a histological diagnosis of intracranial ependymoma (grade II or III according WHO) who recurred after standard treatments. Fourteen adult patients were treated between 1999 and 2008. All patients had a recurrence after surgery and radiotherapy, while 4 received also a first line chemotherapy without benefit. Histological diagnosis was anaplastic ependymoma in 11 patients and ependymoma in 3. All lesions were enhancing on pre-treatment MRI. Tumor location was supratentorial in 8, infratentorial in 5, and supra and infratentorial in 1 patient. Two patients had an associated leptomeningeal spread. Median age was 36 years (range 18–60), median KPS was 75 (range 50–90), and 12 patients had neurological symptoms. Temozolomide was given orally at 150–200 mg/m2 daily on day 1–5 in cycles of 28 days up to 18 cycles. A patient was shifted to an extended schedule (150 mg/m2 1 week on/1 week off) after two cycles. Response was evaluated according to Macdonald’s criteria based on enhanced MRI of both brain and spine performed every 3 cycles.
We observed 2 complete responses, 4 partial responses, 2 minor responses, 4 stable disease, and 2 progressive disease, with a 43% response rate. PFS was 71% at 6 months and 36% at 12 months. Median survival was 19 months (range 3–90). All responding patients had a histological diagnosis of anaplastic ependymoma. In 7/8 responding patients, we observed a neurological improvement. Myelotoxicity was moderate and not cumulative.
Temozolomide has some activity in ependymomas, particularly in the anaplastic variant, and even patients unresponsive to previous chemotherapeutic regimens can respond. These data need to be confirmed in larger series.
Tanycytic ependymoma (TE) is one of rarest histopathological variants of ependymoma. These tumors are usually found in the spinal cord. Supratentorial location is extremely rare. We report five cases of TE: three spinal cases, another one from the fourth ventricle, and an unusual one arising from the lateral ventricle.
From the three spinal cases, two were located on the most frequent site, affecting cervico-medullary junction and high cervical spinal cord. The third one affected the conus medullaris. A complete gross resection was performed in all of them, and no adjuvant treatment was added. One of the patients died 2 years after surgery because of a previously diagnosed leukemia. The other two patients are free of illness one and six years, respectively, after surgery. The patient with the fourth ventricle TE underwent surgery, getting macroscopical complete resection. After eight years, a medullary image suggesting recidiva appeared on MRI. She declined a new surgery, and one year later she keeps without major neurological deficit and the lesion has not grown on control MRIs. The fifth case was a TE located in the left lateral ventricle. A complete transcortical resection was performed and adjuvant radiotherapy was added.
TE usually have a more benign course than other ependymomas. They have special affinity for the cervical spinal cord, being extremely rare the supratentorial location. Microscopically, they are characterized because of their spindly bipolar cells resembling tanycytes. As ependymal rosettes are typically absent and pseudorosettes only vaguely delineated, these lesions can be misdiagnosed as astrocytomas or even schwanomas. TE are low-grade gliomas, and that is why surgical resection and follow up are considered as optimal therapeutic management. Adjuvant radiotherapy in residual lesions remains controversial.
The advances in the understanding of the natural history of WHO grade II gliomas, and consequently the optimization of their therapeutic management, has allowed an improvement of their prognosis. Therefore, more and more young women harboring a WHO grade II glioma now envision a pregnancy. However, relationships between gliomas and pregnancy are poorly known. We report 8 consecutives cases of pregnancies observed in the supra-tentorial hemispheric WHO grade II glioma database of our French glioma study group (REG/Réseau d’Etude des Gliomes). In these cases, the clinical behavior of the tumor clearly worsened in 6 of the 8 observed pregnancies. The clinical implications will be considered and the mechanisms underlying such interactions will be discussed. In such tumor, a pregnancy may reasonably be envisioned, but the benefit-to-risk ratio should be carefully weighted and argumented: the will to have a child versus the possible risks for the fetus and the mother, which depend on the oncological status. If a pregnancy is decided, we advise to perform a very close neurological follow-up with repeated control MRIs in addition to a rigorous obstetrical monitoring.
Infections with neurotropic viruses such as simian virus 40 (SV40), JC virus (JCV), BK virus (BKV), human papilloma virus (HPV) and members of the herpes virus family including human cytomegalovirus (HCMV), herpes simplex virus (HSV), and Epstein-Barr virus (EBV) have been accused to play a role in growth of primary brain tumors. In immunodeficient patients, EBV-carrying B-lymphocytes seemed to be a possible source, but the possibility of direct infection of brain resident cells has not been ruled out yet. Here, we tested the appearance of neurotropic viruses. In one oligodendroglioma tissue we could verify an EBV infection. We describe results obtained from our study sample, the clinical course of the EBV-positive patient, and our attempt to analyze the real source of EBV-positive cells. DNA of 55 frozen brain tumor samples (gliomas, meningiomas) was isolated and tested by nested PCR analysis for the integration of viral sequences from the above mentioned neurotropic viruses. In one patient with an EBV infection of his low-grade oligodendroglioma sample, DNA derived from the tumor tissue, the corresponding tumor cell culture and the patient’s blood were analyzed for an EBV sequence. Immunohistochemical staining was performed using antibodies directed against the EBV-specific protein LMP1. While HSV, HCMV, HPV, BKV, JCV, and SV40 were not detected in any of the tumor samples, we amplified a viral DNA sequence of the EBV-specific protein EBNA-2 in the tumor tissue of a 53-year-old patient suffering from low-grade oligodendroglioma who underwent neurosurgical tumor resection. High IgG titers against EBV were detected in the patient’s serum, and the EBNA-2 sequence was amplified from the tumor tissue and blood cells but not from the corresponding cell culture. In line with these results, we did not detect any EBV-specific LMP1 protein expression in the patient’s tumor tissue by immunohistochemical staining. Our data suggest that infection of brain tissue by neurotropic viruses is an extremely rare event and does not seem to play a significant role in the growth of primary brain tumors. For the only patient in our study sample with a confirmed EBV infection in his tumor tissue, we were the first to identify blood cells to be the real source of EBV while tumor cells were undoubtedly negative. Therefore, future studies should include an analysis of corresponding blood DNA in order to answer the question whether viral infections are involved in tumorigenesis.
Primary leptomeningeal oligodendroglioma has very rarely been reported. The tumor is thought to be derived from heterotopic nests of neuroglial tissue within the leptomeninges. The disease has always a fatal outcome within a few years even if in some rare cases a partial response has been observed to radiotherapy and chemotherapy. Diagnosis during life may be difficult, and sometimes definite diagnosis is made only at autopsy. In the literature only one case has been reported with a deletion of chromosome 1p. We report a 60-y-old man presenting with seizures. Brain MRI showed a diffuse right parieto-occipital subarachnoidal enhancing lesion without intraparenchymal extension. Pathological examination of the surgical specimen obtained by open biopsy revealed an anaplastic oligodendroglioma (WHO grade 3). Radiotherapy was planned but during the preparatory work-up a marked progression was observed on MRI with diffuse ipsilateral and contralateral subarachnoidal tumoral extension. MRI of the spine revealed multiple enhancing nodular intradural lesions spreading over the entire spinal cord. Because of the diffuse meningeal involvement no radiotherapy was considered. As molecular analysis using the FISH technique showed 1p36 and 19q13 deletions, we decided to treat the patient with temozolomide 150 mg/m2 for 5 days every 4 weeks. We observed a dramatic regression of the lesions on serial MRIs with no residual enhancement left after 6 cycles of chemotherapy. The patient showed a remarkable neurological recovery. To our knowledge this is the first report of a patient with a primary leptomeningeal anaplastic oligodendroglioma with diffuse spinal seeding bearing a 1p36/19q13 deletion. Our patient showed a remarkable and sustained clinical and radiological remission following temozolomide treatment. Our case illustrates that molecular analysis with a search for 1p/19q deletions should be performed in all cases of leptomeningeal gliomas to select those patients who might benefit from temozolomide chemotherapy.
Malignant gliomas of the spinal cord are very rare neoplasms with a dismal prognosis. Although temozolomide (TMZ) is a novel oral alkylating agent with demonstrated efficacy as first-line and second-line therapy for patients with intracranial high-grade astrocytic gliomas including glioblastoma (GBL), there is no evidence or report that it is also effective for primary malignant gliomas of the spinal cord. Authors experienced six cases of primary spinal cord malignant gliomas which have been treated with multidisciplinary treatment using TMZ, and this retrospective study has been performed to determine the efficacy of and to document the role of TMZ in these patients. This may be the first report concerning the results of TMZ with a relatively large number of primary malignant gliomas of the spinal cord.
Six patients (21–54 years, median 40) were enrolled into this retrospective study from Dec 2003 to Jan 2008. Three patients were anaplastic astrocytoma (AA), 2 were anaplastic ependymoma (EPN), and one was primary spinal cord GBL. TMZ (75 mg/m2/d × 7 d/wk for the period of radiation therapy) was administered orally concomitant with fractionated radiotherapy followed by TMZ monotherapy (150–200 mg/m2/d × 5 days, every 28 days for more than six cycles) in one GBL patient, and TMZ was used as adjuvant therapy (150–200 mg/m2/d × 5 days, every 28 days for more than six cycles) in the other five cases after radiation therapy.
One patient with GBL died after concurrent chemoradiotherapy using TMZ and 6 cycles of TMZ chemotherapy at 17 months of diagnosis. In three AA patients, one died after 39 months, one who had recurrent tumor after 12 years after radiation therapy showed partial response state after 6 cycles of TMZ (10 months after recurrence), and one was stable now with ongoing TMZ chemotherapy. In one anaplastic EPN patient 12 cycles of TMZ chemotherapy was done after radiation therapy. The patient showed partial response and was clinically stable 18 months after diagnosis. In the other case it was progressed after radiation therapy followed by 2 cycles of TMZ and although second line chemotherapy was done he was dead 9 months after diagnosis. These results were relatively excellent ones compared with historical data or data on published papers. Non-hematologic toxicities such as fatigue and anorexia were mild in severity. The concomitant treatment phase or adjuvant chemotherapy showed grade 1 neutropenia and thrombocytopenia in 2 patients only.
TMZ may be one promising chemotherapeutic agent for primary malignant gliomas of spinal cord even for anaplastic EPN as primary or second line therapy. Concurrent chemoradiotherapy using TMZ may be also positively taken into account for the initial treatment of spinal cord AA or GBL. Further investigation of a larger number of patients and randomized clinical trial are mandatory.
Myxopapillary ependymoma is a WHO Grade I variant of ependymoma which occurs in the lumbosacral region and is characterized by perivascular cuboidal cells with myxoid degeneration. Acute spinal subarachnoid hemorrhage (SAH) due to myxopapillary ependymoma has been described as a presenting feature. We present cases illustrating acute, recurrent, and chronic hemorrhage due to myxopapillary ependymoma and discuss the implications for management.
Case 1: A 32-year-old man presented following thrombolysis for a myocardial infarction with paraparesis due to acute hemorrhage from a lumbosacral myxopapillary ependymoma. Following complete surgical resection he recovered good neurological function. Case 2: A 42-year-old man who had suffered episodes of acute spinal SAH from presumed neurofibromata over 20 years underwent surgery for progressive symptoms and was found to have multiple sites of metastatic myxopapillary ependymoma throughout the spinal canal. He was initially managed with surgery alone but subsequently received radiotherapy to an area of progressive disease with no further subsequent episodes of acute SAH. Case 3: A 36-year-old man underwent surgery and radiotherapy for a lumbosacral myxopapillary ependymoma. Despite no evidence of macroscopic residual disease he subsequently developed deafness, ataxia, and cognitive impairment due to superficial siderosis from chronic SAH.
Myxopapillary ependymoma is generally regarded as a slow-growing tumor with an excellent prognosis. Treatment is by complete resection—where possible—with adjuvant radiotherapy given to areas of macroscopic residual disease. Hemorrhage is recognized as a presenting feature but is less well known as a cause of chronic morbidity. The cases illustrate the hemorrhagic potential of myxopapillary ependymoma. Chronic subclinical hemorrhage from residual disease can lead to superficial siderosis and permanent neurocognitive impairment with substantial morbidity. Radiotherapy is effective at reducing vascularity and hemorrhage in other clinical contexts and should be considered following incomplete resection to both improve local control and minimize the risk of future hemorrhage.
Patients with myxopapillary ependymoma are at risk of hemorrhage and long-term morbidity. Radiotherapy to residual or metastatic disease should be considered to minimize this risk and all patients should be monitored for signs of chronic hemorrhage.
Clinical outcomes of patients with intramedullary tumors seem to depend on histology and extent of surgical resection. The aim of our study was to assess surgical results and the neurological outcome of these patients.
A total of 43 tumor resections was performed in 38 patients (27 ependymoma, 7 astrocytoma WHO °I, 4 astrocytoma WHO °II) between 1/2002 and 5/2006 under intraoperative electrophysiological monitoring, 29 in the ependymoma, and 14 in the astrocytoma group. For ependymomas mean tumor extension was 2.9 levels (range 1–10), and in astrocytomas was 3.7 levels (range 1–9). In case of multilevel approaches, we performed a laminoplasty with refixation of the incised lamina(e). Routine intraoperative ultrasound and postoperative MRT was done to evaluate complete tumor removal. Clinical outcome using the McCormick scale (McC 0, no deficit −5, completely dependent) was assessed before operation, one week and 6 months after surgery.
Complete tumor removal was achieved in 22/29 ependymomas and in 6/14 astocytomas, 76% and 43%, respectively. Recurrent tumor growth during follow-up (range 6–40 months) was noted in 2 astrocytoma patients and no ependymoma patients. Ependymoma patients showed an improvement of neurological function during follow-up: Median McC improved from 2 (95%CI: 0.29) to 1 (0.38) after 1 week and 1 (0.40) after 6 months. In astrocytomas, in only 4/14 cases patients (McC 1–2) improved slightly. However, McC scores remained unchanged: 2 (0.36) preoperatively, 2 (0.39), and 2 (0.43) postoperatively. No neurological deterioration was seen in these patients.
In contrast to generally high resection rates and encouraging outcomes in ependymomas, favorable neurological outcomes in astrocytomas are achieved only by (1) using cautious intraoperatibe strategies and (2) in patients with good preoperative McC. Therefore, early admission for surgery instead of watchful waiting should be the strategy in these patients.
Ganglioglioma is a rare type of central nervous system (CNS) tumors occurring most frequently in children and young adults. In this report we analyze thirteen cases of CNS ganglioglioma treated in our institution during the period 2003–2007. During the last five years, 699 patients having astro- and/or oligodendroglial CNS tumors were operated on in the Department of Neurosurgery at University Hospital “St. Ivan Rilski,” Sofia. Ganglioglioma occurred in 13 (1.86%) of them. The group consisted of 7 (53.8%) male and 6 (46.2%) female patients. The mean age of the patients was 23 years (from 4 to 74 years). The histological grade of gangliogliomas was as follow: grade I in 2 (15.4%) cases, grade II in 8 cases (61.5%), and grade III in 3 cases (23.1%). During the study period, only one patient necessitated second intervention because of tumor recurrence and anaplastic progression—a 36-year-old woman with coexisted Turner syndrome. Despite the fact that most gangliogliomas show benign clinical behavior and long-term survival rate, in some cases anaplastic progression may be seen. Further studies on a larger number of patients are needed to find the appropriate treatment strategies in these cases.
Peripheral primitive neuroectodermal tumor (pPNET) is a kind of highly malignant tumor, and researches on the tumors are rare. This study was to investigate the clinical characteristics, treatment, and prognostic factors of pPNET, and thus to explore the treatment principle and efficacy of the treatment.
Clinical data of 40 pPNET patients, treated in Cancer Center, Sun Yat-Sen University, from 1997 to 2007, were analyzed retrospectively. Cumulative survival rate was analyzed by the Kaplan-Meier method. The log-rank test was applied to compare survival rates.
Of 40 patients, 23 received multimodality treatment, 14 received only surgery or chemotherapy, and 3 did not have any anticancer therapy. The median survival time was 23 months. The 1-, 2-, and 5-year overall survival (OS) rates of the 40 patients were 67.7%, 49.9%, 41.6%, respectively. Five-year OS was 57.9% in patients who received multimodality treatment, 13.1% for the rest (P = 0.000); 25.1% in those with tumor size ≥10cm, 48.5% in those with tumor size <10cm (P = 0.028); 50.8% in those who received complete resection, 41.7% in those who underwent incomplete resection, 27.5% in those who received no surgery (P = 0.048); and 45.9% in those with nonmetastasis, 35.7% in those with metastasis(P = 0.424).
Based on our experience and a review of the literature, we concluded that multimodality treatment is the main treatment of peripheral primitive neuroectodermal tumors. Tumor size greater than 10 cm, complete resection, and metastasis are prognostic factors for survival.
Astroblastoma is one of the rarest tumors of the CNS whose classification, histogenesis, diagnosis, and therapeutic management are still being debated. The typical histopathological appearance is the perivascular, astroblastic pseudorosette, which is, however, present in other CNS tumors. Although it has intermediate features between astrocytomas and ependymomas, observations of a particular cytogenetic profile and clinical behaviour make it seem more and more likely that astroblastoma is a single entity. To clarify the clinical, radiological, histopathological, prognostic, and therapeutic characteristics, which have been treated only recently and not well established yet, six cases of histologically proven astroblastoma were retrospectively analyzed playing special attention to therapeutic remarks.
Between 1996 and 2005, six patients with histologically proven cerebral astroblastoma were surgically treated at the Department of Neurosciences-Neurosurgery of Rome Sapienza University. In three cases the lesion was a low-grade astroblastoma, high-grade in the other three.
Median age of the six patients was 36 years. The time to diagnosis ranged from one week to 18 months. Surgical removal was total in four cases, subtotal in two. All patients received radiotherapy; two also had chemotherapy with temozolomide (TMZ). The three patients with low-grade astroblastoma are still alive today after a follow-up of 2, 5, and 19 years respectively. Of the three patients with high-grade lesions, one is still alive after a 7-year follow-up, while the other two survived for 17 months (progression time 15 months) and 35 months (progression-reoperation time 23 months) respectively.
Radical surgical resection appears to be the treatment of choice for astroblastoma. Radiotherapy may play an adjuvant role in the treatment of high-grade lesions. The role of chemotherapy is very debatable, and we advocate the use of a safe adjuvant chemotherapeutic regimen with TMZ, especially for the high-grade astroblastoma cases.
Paragangliomas represent rare tumors of neural crest origin. They are highly vascular neoplasms and in the majority of cases are benign. Traditionally they have been treated by surgery, embolization, and/or external beam radiotherapy. Unfortunately these therapies are charged by the risk of high morbidity and mortality. Stereotactic radiosurgery could play a relevant role as a therapeutic approach: the aim of this study is to evaluate the acute effects and efficacy in term of local control using radiosurgery.
Between August 2004 and December 2007, 9 patients, 8 with jugular glomus paraganglioma and 1 with carotid paraganglioma, have been referred to our Cyberknife center: in four cases patients were undergone to partial excision; no patients have had previous radiation therapy. The most common presenting symptom was conductive hearing loss, pulsatile tinnitus, and dysphagia, except for one patient who presented hypertension crisis. The patient with carotid glomus tumor had postoperative weakness of cranial nerves X, IX. Tumor volume ranged from 1602 mm3 to 12782 mm3 (median 5848 mm3) The delineation of the target was based on the integration of images from TC scan fused with RMN: the planning target volume was defined as the radiographic tumor volume with no margin. The tumor dose was at 11 to 13 Gy (mean 12.5 Gy) in single fraction for 8 patients (reference isodose 72–83%), and one patient received staged radiosurgery (total dose 24 Gy in 3 fractions).
The mean follow-up was 20 months. One patient died and thus was lost to follow-up. All patients have showed tumor stabilization. In two patients transient vertigo occurred, and one patient had headache with complete resolution three months after treatment. All patients demonstrated neurological stability. No cranial nerve palsies arose or deteriorated.
Stereotactic radiosurgery appears to be both safe and effective in the treatment of glomus tumors. Long-term complications remain to be studied, and despite our preliminary results it is prudent to be cautious about tumor local control. Important prerequisites for good results are CT and MRI image fusions to correctly identify the PTV, and inverse planning to obtain the best conforming dose with minimal treatment delivery errors.
The skull base reconstruction is a very important procedure in the cases of radical resection of skull base tumors.
We presented 72 patients (male 40, female 32), aged 3.5 to 76 years, with skull base tumors extending into the orbits and paranasal sinuses (benign 45, malignant 27). All these lesions were divided into 3 groups. I group: the midline lesions (42 cases) included defects of ethmoid and sphenoid sinuses, frontal sinuses, and medial parts of maxillary sinuses. II group: the lateral lesions involved lateral parts of frontal sinus, upper-lateral parts of maxillary sinus as lateral skull base defects (19 cases). III group: combined skull base defects included both medial and lateral defects with widely opened paranasal sinuses and nasopharynx. (11 cases).
It is important to emphasize that the reconstruction with a periosteum flap from the frontoparietal area should be preferred in midline defects, a temporalis muscle flap with adjusted periosteum- in lateral defects. The reconstruction with autograft using microsurgery technique (m. latissimus dorsi flap, m. pectoralis, combined flap using m. pectoralis and m. abdominalis rectus musculocutaneous flap, omentus, m. latissimus dorsi with split-rib grafts) is indicated in case of combined defects. It was observed that two patients had nasal cerebrospinal fluid leak which resolved after continuous lumbar drainage. Preoperative planning of optimal method of closure of skull base defect depends on location and expansion of skull base tumor.
The lacrymal gland tumors are relatively rare. Specific neurosurgical approaches are needed in some cases with intracranial expansion. Subjects of the study were patients having different type of lacrymal gland tumors surgically treated in the Department of Neurosurgery at University Hospital St. Ivan Rilski, Sofia during the period 1998–2007. The studied group consisted of 14 patients, 6 (42.9%) males and 8 (57.1%) females. The mean age of the patients was 41 years. The preoperative diagnosis was made using neuroophthalmological evaluation and neuroimaging procedures. All patients were operated on. Total excision of the tumor on the first stage was achieved in 12 (85.7%) cases, subtotal in 2 (14.3%) of them. The histological type of the tumor was as follows—adenoma in 3 (21.4%) cases, tumor mixtus in 4 (28.6%), carcinoma in 7 (50.0%). Five patients were reoperated on. In two of them orbital exenteration was performed. The lacrymal gland tumor treatment is mainly surgical with two main aims—radical excision of the lesion and vision preservation. The prognosis in these cases depends on histological type and tumor expansion at the time of diagnosis.
The aim of this study was to determine the surgical frequency of the intraorbital tumors, to analyze the most frequent clinical symptoms, to define the diagnostic role of neuroimaging methods, to specify the optimal neurosurgical treatment, and to analyze the results of operative management and most frequent postoperative complications.
The study included 38 patients with orbital tumors who were admitted in the period 2000–2007. Thirty-six patients with primary intraorbital tumors were operated on. In 2 patients reoperative interventions were done by reason of residual tumor or tumor recurrence. Different surgical approaches were used according to tumor localization.
The extraconal tumors were prevalent. The most frequent clinical symptom was exophthalmia (89%), followed by diplopia (45%), bulb dislocation (26%), and visual disturbances (15%). According to histopathological determination, mucocele was found in 29%, dermoid cysts in 11%, lacrimal gland tumors in 8%, neurinomas in 8%, and lymphomas in 8%, etc. Magnetic resonance imaging has some disadvantages as a diagnostic method. In 43% of operated patients the lateral orbitotomy was done. Transitory postoperative complications were assessed in 25% of patients and in 14% of them the complications were permanent.
The choice of surgical approach was defined according to tumor localization, size of the tumor, expected histopathological results, and results of neuro-ophtalmologic investigation. The right surgical approach and precise surgical technique decrease the frequency of postoperative complications.
Brain tumors located in the ventricles represent 1.4% from all brain tumors. From a surgical point of view, the profound location and the necessity to intersect normal brain tissue with minimal functional alteration, the abundant vascularisation that interferes with normal vasculature, and the large dimensions of tumors, altogether represent problems that must be solved by neurosurgeon.
We retrospectively present the indications, results, outcomes, and surgical treatment of 39 patients diagnosed with intraventricular supratentorial tumors.
Supratentorial intraventricular tumors represent 2.8% of 1,376 intracranial tumors operated on by the author between January 2000 and December 2007. From a surgical perspective, our strategy is a transcortical approach for tumors located in the lateral ventricle, for tumors located in the third ventricle transcallosal, and subfrontal trans-laminaterminalis approaches, if there is an intraventricular extension from a sellar tumor. VP-shunt is reserved only for cases of persistence of symptomatic hydrocephalus after tumor removal, or for cases in which the direct approach was denied or considered deleterious related to neurological status of the patient. Temporary EVD (4 to 5 postoperative days) is frequently used after transcallosal or transcortical approach. By this protocol we obtained a total removal in 84.6% of the cases. Recurrences occurred in 12% of the cases. Pathological examination revealed features such as 23% astrocitomas, 18% malignant gliomas, 12.8% ependimomas and craniopharingyomas, 5% cholloid cysts and neurocytomas, and 2.5% metastases, meningiomas, gangliolglyomas, and pituitary adenomas. Postoperative complications consisted of persistent symptomatic hydrocephalus (5 cases) and intraventricular bleeding and/or edema in four cases. The surgical results were GOS 5 and 4 in 30 cases, 7 cases with GOS 3, and 2 deceased (one case with multiple metastases and one case with pulmonary embolism).
Complete neuroimagistic evaluation, a proper surgical approach, progressive debulking of tumor with great attention to preserve normal vasculature, and EVD in order to prevent acute hydrocephalus are the key points for good surgical results.
Endoscopic resection of sella lesions has become more frequently performed in recent years. Our initial experience is evaluated.
Since April 2006 we started to perform endoscopic sella surgery. Till April 2007, altogether 39 patients with either pituitary adenoma or craniopharyngeoma were operated on with this technique. This group consisted of 24 women and 15 men. Three patients had STH hormone hypersecretion and 4 had ACTH hypersecretion. Altogether, 26 patients had bitemporal hemianopsia caused by compression of the chiasm. Headaches lead in six cases to MRI finding of pituitary adenoma which was reaching to chiasm but not yet causing visual field deficit.
Pituitary adenoma was resected in 37 cases and in two cases craniopharyngeoma. There was no 30-day morbidity and mortality in this small series. In one case redo surgery was performed due to refractery postoperative CSF leak. In one case, postoperative nasal packing was needed due to nasal bleeding. Normalization of hormone levels was reached in three of four cases of ACTH hypersecretion. In two cases, STH decrease was sufficient but was not in one other; residuum of adenoma was disclosed on postoprative MRI and the patient prefered LGK treatment as the second step procedure. Visual field deficit improved in 19 patients (73%), and remained stable in 7 patients. The mean postoperative in-hospital stay was 4.4 days. All patients found postoperative comfort as very good and suffered minimal pain.
Endoscopic resection of sella lesions is a feasible technique with a relative steep learning curve. In-hospital stay was shortened in this group of patients compared to in-hospital stay of patients who formerly underwent sublabial paraseptal resection. The postoperative comfort was enhanced, and all patients tolerated the surgery and postoperative course very well.
A retrospective analysis of the surgical treatment and results of 20 patients with benign and malignant craniospinal tumors operated on during the period 1992–2007 at the Department of Neurosurgery, Medical University-Sofia, Bulgaria was performed. The aim of the analysis was to assess the factors affecting choice of minimal invasive approach, operative radicality, and outcome. The mean age of the patients was 45.2 years. The male to female ratio was 1:1.5. The most frequent neurological symptoms were intracranial hypertension, qudriparesis, bulbar palsy, alternate syndromes, hydrocephalus, and spinal nerves palsy. Preoperative imaging revealed extradural tumors in 19 cases and intra/extradural tumor in 1 case. In 13 patients the tumor origin was cranial with secondary spreading into the spinal area, and 7 patients had primary spinal tumor localization followed by cranial spreading. In this group of patients we have observed 1 meningioma, 1 myeloma, 5 metastasis, 12 chordoma, and 1 sarcoma. In our group of 20 patients we made 34 operations. We performed 30 cytoreductive operations and because of craniocervical instability, we made occipitospinodesis in 4 cases. We used a transoral approach in 22 cytoreductive operations, and a transsphenoidal, endonasal approach was employed in 8 operations. Complete tumor removal was achieved in 1 case (3.33%), subtotal resection in 15 cases (50%), partial tumor resection in 10 of the cases (33.33%), and biopsy in 2 cases (6.66%). The most common operative complications were: CSF leak in 1 case (3.33%), brainstem edema in 3 cases (10%), and operative mortality in 3.33%. The 5-year follow-up showed good recovery in 42%, moderate disabling in 35%, severe disabling in 13%, vegetative state in 4%, and death of 6%. We conclude that early minimal invasive decompressive surgery facilitates neurological recovery by preserving the existing neurological function. Transoral and endonasal, transsphenoidal approaches are an effective surgical method for direct decompression in patients with irreducible ventral midline extradural compressive tumors of craniospinal area.
The tuberculum sellae meningiomas are usually removed using various transcranial approaches. The so-called extended transsphenoial approach is a modified transsphenoidal procedure for some lesions around the sella turcica, anteriorly to the tuberculum sellae and the planum sphenoidale, laterally to the cavernous sinus, and posteroinferiorly to the clivus. The author approaches the supradiaphragmatic tumors though the tuberculum sellae, termed “transsphenoidal-transtuberculum sellae approach,” even for some tuberculum sellae meningiomas. This paper describes some characteristics and indications between the transsphenoidal-transtuberculum sellae approach and the interhemispheric approach for the tuberculum sellae meningiomas.
The operative approach was selected by three indexes: (1) development of the sphenoid sinus, (2) anterior and lateral extension of the tumor, (3) tumor extension into the optic canal. Utilizing the transsphenoidal-transtuberculum sellae approach since 1994, 25 supradiaphragmatic tumors including 3 tuberculum sellae meningiomas have been removed, and two tuberculum sellae meningiomas were resected by the interhemispheric approach because of the broad base and the intraoptic canalicular extension of the tumor.
Utilizing the transsphenoidal-transtuberculum sellae approach, total resections were performed for two meningiomas and subtotal resection was achieved in one case. The other two meningiomas were totally removed using the interhemispheric approach. In the transsphenoidal-transtuberculum sellae approach, the operative complications were the injury of the anterior cerebral artery during resection of the tumor, which was wrapped for hemostasis using fibrin sealant, and the postoperative pneumocephalus. Whereas in the interhemispheric approach, unilateral olfactory nerve injury and asymptomatic unilateral frontal contusion occurred in both cases. The mean operative time was 5h 52m in the transsphenoidal-transtuberculum sellae approach and 8h 38m in the the interhemispheric approach, and the mean hospitalization was 32.7 days in the former and 20 days in the latter.
The transsphenoidal-transtuberculum sellae approach for accessing small tuberculum sellae meningiomas is a useful procedure as a less invasive technique with short operative time compared with the interhemispheric approach, although it requires care during incision between the tumor and the surrounding structures.
Traditionally, the tumors of the anterior cranial fossa and/or sellar area, except pituitary adenomas, are approached through a subfrontal or transylvian approach. The latter is the most popular for the sellar area, after the technique described by Yasargil in the ’70s.
We used the transylvian approach routinely over the years. In selected cases, since 4 years, we used the supraorbital mininvasive approach according to Perneczky. We operated on 18 patients with meningioma: 13 of the tuberculum sellae, 5 of the etmoidal planum. There were 3 males and 15 females. Age ranged from 39–81, average 57. In 4/5 of the patients with planum ethmoidalis tumor the diagnosis was done at MRI or CT made for other reasons; the same happened only in one of those with tuberculum sellae meningioma. 12/13 of these had severe preoperative visual impairment.
The anatomical operative view was excellent. We had some difficulties using large instruments. However, it was rather easy to become accustomed to work through a limited bone flap. Head position, CSF peroperative spinal drain, and excellent anesthesia are essential for the approach. The frontal lobe should not be retracted but gently reclined. One patient with tuberculum sellae meningioma died due to the sequelae of an internal carotid artery intraoperative lesion. Another had transient right hemiparesis. 4 had postoperative anosmia on the side of the approach. Out of the 12 patients with visual deficit, 11 were evaluated (1 died): 9 improved, 2 worsened: one had subclinoidal implant of the tumor, another was a recurrent case where adhesions between optic nerve and tumor were tenacious. All the patients with planum tumors did well. The only symptomatic one had her ataxia and confusional syndrome improved. Postoperative MRI, done at 3–6 months postoperatively, did not show any sign of brain contusion. None of our patients had seizures and at present needs any anticonvulsivant medication. Cosmetic results were excellent, therefore, there is no postoperative retraction of the temporal muscle.
In our minds, these are the major advantages of the technique. The major clinical disadvantage may be a postoperative olfactive disturbance, because in some cases it has been impossible for us to spare the olfactory tract of the side of the approach. This is an elegant and effective alternative to the subfrontal and pterional ones. Its limits may be the dimensions of the tumor if combined with fibrous or hard consistency.
Spinal intradural tumors are generally removed by uni- or multilevel laminectomy with midline dural incision. Delayed postoperative kyphosis and spinal instability, which occurs in 6% of the patients, may be reduced by a more conservative unilateral microsurgery, sparing bilateral damage.
49 patients, 34 with neurinoma, 8 meningioma, 3 caudal ependymoma, 2 conus medullaris dermoid tumor, and 1 each dorsal intramedullary melanoma and cervical intramedullary metastasis were operated on between June 2000 and June 2007. One had 3 lumbar tumors and required two operations because of MRI misinterpretation, one had 2 dorsal meningiomas removed during the same operation, and 2 patients had dumbbell neurinoma. 1 neurinoma and 2 meningiomas were located at the C1–C2 level. Surgery was performed in the prone position with a unilateral approach. The extension of the laminectomy was kept generally to one level, having care to remove all the ligamentum flavum. When necessary, it was extended cranio-caudally for 1.5–2 cm. The dura was opened paramedially and the tumor dissected and removed either “en bloc,” when smaller than 2 cm, or piecemeal. In one patient, it was deemed necessary to enlarge the exposure to a traditional approach because of concomitant severe stenosis and huge tumor. Dural closure is done with 5–0 or 6–0 stitches. Tumor dimensions ranged from 1.2 to 2.8 cm in maximum diameter. The patient that required the extension of the approach had a multilobulated neurofibroma of 6 cm.
All the patients were mobilized on day 2–3 and discharged on day 4–5. No important complications due to the technique were observed; postoperative pain was minimal. One patient had a pseudomeningocele under the muscular plane, that did not require reoperation. Plain X-ray films showed that none of them had kyphosis and/or instability 3 months post-op. Overall neurological results were good.
Hospital stay may be reduced and stability preserved with an appropriate microsurgical mini-invasive technique. In selected patients it can be used also for small intramedullary lesions, although in our cases it was done because of preoperative misdiagnosis. In case of operative difficulty it can be converted rapidly into a traditional approach: only one case in the present series.
Brain tumors can be a reason of medically intractable temporal lobe epilepsy (TLE). Intralobar localization of new growth influences epileptogenesis, features of diagnostics, and surgical treatment.
During 2004–2007, 7 adult patients were operated concerning intracerebral brain tumors, combining with pharmacoresisting TLE. In all cases TLE was the main or unique clinical display of disease on an extent from 6 months to 18 years. Used: EEG, MRI, ECoG, and EsCoG. In 5 supervisions gliomas were small, localized directly in temporomedial departments, and coincided with the hearth of epileptic activity. Operative intervention was limited to the resection of new growth and perifocal area within the limits of the temporomedial structures. In one supervision, large (5 × 6 cm) astrocytoma II was localized in the overhead and middle departments of the right temporal lobe and spread on the insula but had secondary influence on temporomedial structures with formation in hippocampus, the limited focus of destruction. All of the pathologically changed area of the temporal lobe was totally removed. 2 patients with TLE had large astrocytomas, without a germination in a temporal lobe. In these cases low-grade tumors were removed totally without surgical manipulations on temporal structures. Histological results: astrocytoma 4 cases, ganglioma/gangliocytoma 1, and oligodendroglioma 2.
Surgical mortality was zero. The attacks of epilepsy in all cases were stopped. Two patients undergoing surgical resection in insula may have transient postoperative motor and speech deficits, which recover during few weeks or months.
Intracerebral low-grade tumors of temporal localization can be a principal cause of pharmacoresistant focal temporal epilepsy. The mechanisms involving epileptogenesis temporal mesenchymal structures are different. In these cases radical open surgical interventions are highly effective ways of treating TLE.
The morbidity/mortality of glioma surgery varies between 5–30% according to different studies. Tumor localization, type of tumor, and performance status are the major factors influencing risk of glioma surgrey.
Restrospective analysis of all patients who underwent glioma surgery in year 2006. Short-term postoperative complications, 30-days morbidity/mortality, and relation between postoperative morbidity and patients’ outcome were studied.
Altogether 100 surgeries in 98 patients with glioma tumor were performed. Brain glioma was treated in 97 cases, spinal cord glioma in 3 cases. Histological findings were as follows: pilocytic astrocytoma 2, ganglioglioma 4, plexus papilloma 1, fibrillary astrocytoma 11, anaplastic astrocytoma 4, glioblastoma multiforme 51, gliosarcoma 2, medulloblastoma 2, oligodendroglioma 5, anaplastic oligodendroglioma 10, anaplastic oligoastrocytoma 2, ependymoma 2, and anaplastic ependymoma 1. Spinal cord tumors: fibrillary astrocytoma in cervical level 5–7, ependymoma in thoracic level 8–10, and myxopapillary ependymoma in lumabosacral level. Average preoperative Karnofsky score (KS) was 89, postoperative 87. Complications were observed in 17 cases (epidural hematoma 2, repeated epileptic seizures 2, bleeding into resection cavity 3, postoperative edema causing deterioration of neurological picture 3, peroperative air embolization 1, postoperative fever accompanied by confusion 1, and worsening of neurological picture 5). Preopertive KS in group of patients with complications was 85, postoperative 60. 30-day mortality in group of patients without complications was 0%, in complications group was 35%. By the end of 2007, 4 patients from complication group were alive (mortality 76.4%, postoperative median survival 4.5 months) and 42 patients from noncomplication group were alive (mortality 49.3%, postoperative median survival 9.4 months).
Postoperative survival in group patients with any postoperative complication was significantly shorter than in noncomplication group. Risk of surgery even in modern neurosurgical era is still significant (17%).
Surgical resection of tumors located in the insular region is challenging for neurosurgeons. There are four reasons. The first reason is a tumor resection between middle cerebral artery. The second reason is the opercular part is a barrier for resction of tumor. The third reason is the pyramidal tract and the arcuate fasciculus exist around the tumor. The fourth reason is the important small artery (lenticulostriate artery and long insular artery) exists in surroundings of the tumor. Coexisting illness of the insular glioma removal is generated by short odds. We report our experience with 40 intrinsic tumors of the insular between April 2000 and October 2006. The open MRI and navigation were used for checking the residual tumor. Neurological monitoring (SEP, MEP) was used for checking existence of nerve fiber. Awake surgery and cortical mapping were used for checking verbal function. If a correct removal method is used, even if gliomas are located in the insular region, it is possible to remove all safely.
Recently, glioblastoma therapy was advanced by the approval of ALA for enhancing resections and approval of temozolomide for adjuvant radiochemotherapy. It is unknown whether maximal resections combined with concomitant therapy are additive in their effects on outcome. To determine whether additive effects are present, a large observational study was conducted for which preliminary results are available.
Data of 102 of 169 glioblastoma patients recruited into an observational study by the ALA glioma study group were available for this analysis. All patients had early postoperative MRI within 72 hours after surgery and MRIs in 3-monthly intervals. Concomitant therapy was performed according to the EORTC 26981 protocol (75 mg/m2 oral/day during radiotherapy followed by adjuvant temozolomide 5 of 28 days at 150–200 mg/m2). Progression-free survival (PFS) was determined based on contrast-enhancement on gadolinium-enhanced t1 weighted MR images. Patients were directly compared to patients from the phase III ALA study.
Patients from the observational study where comparable to patients from the ALA study regarding median age (60.7 vs. 60 years), median KPS (90 vs. 90) and frequency of complete resections (66 vs. 65%). PFS was 8.6 (95% CI: 7.1–10.7) months compared to 5.1 (3.4–6.0) months in the ALA study. Within the observational study, PFS for patients with complete resections was 13.0 (9.9–16.2) compared to 5.1 (3.9–6.3) months for patients with incomplete resections (p = 0.000). In the ALA study PFS for complete vs. incomplete resections was 5.9 (3.4–6.8) vs. 3.5 (3.2–5.3) (p = 0.0014).
These preliminary results indicate that the efficacy of concomitant therapy is enhanced by optimal resections.
Tumors infiltrating highly functional structures like sensorimotor and language cortical areas and subcortical pathways cause a reorganization of cortico-subcortical functional maps. Mechanisms of cerebral plasticity take place as a result of surgical removal of lesions and determine a remodeling of neuronal circuitry, which correlates to clinical evolution. Privileged tools for studying mechanisms of cerebral plasticity are functional MRI (fMRI) and cerebral electrical stimulation mapping (ESM).
We retrospectively analyzed clinical data of 20 patients operated on with the aid of ESM. The preoperative clinical-radiological evaluation included determination of hemispheric dominance for speech function, neuropsychological assessment using Boston Diagnostic Aphasia Evaluation, and fMRI. Patients harboring lesions in or near language areas underwent awake craniotomy for language mapping, while the others displaying lesions involving motor areas were operated on in general anesthesia with the aid of ESM for locating motor pathways.
Eight patients harbored a mass lesion in or near language areas and twelve patients had a tumor involving sensorimotor pathways according to clinical presentation and radiological findings. All patients had immediate post-operative deficits. Recovery occurred within 3 months in all of them. Twelve lesions were either totally or subtotally resected on post-operative MRI.
There are mechanisms of cerebral plasticity taking place before treatment of the lesion and both in an acute stage and at distance from surgical intervention. Even if cortico-subcortical rearrangement of functional circuits seems to be more efficient in case of slow growing lesions, a brain plasticity potential is present also in case of malignant tumors, which may be vulnerable to tumor resection as well as to complementary chemo- and radiotherapy. Understanding how peri-tumoral tissue takes over the lost function and resists to further damage due to tumor relapse and complementary treatments and integrating this knowledge into a wider view of the modifications of cerebral connections in case of brain tumors may guide treatment of cerebral tumors in eloquent areas and define prognosis and rehabilitation programs.
Neuronavigation has become a standard technique procedure during brain surgery. However, the imaging used in conventional neuronavigation is acquired preoperatively and thus, cannot project the dynamic changes during surgery and result in significant brain shift and decreased accuracy. Intracranial navigation by using intraoperative magnetic resonance imaging (iMRI) allows the surgeon to reassess anatomical relationships in near-real time during brain tumor surgery. We report our experience with a neuronavigation system coupled to a low-field iMRI system.
Between September 2005 and August 2007, 148 neurosurgical procedures were performed using the mobile 0.15-tesla PoleStar N-20 iMRI system. The cases included 40 craniotomies and 107 transsphenoidal approaches for 22 gliomas, 105 pituitary adenomas, 5 metastatic tumors, and 15 other tumors.
Tumor resection was controlled with the use of image guidance until the final intraoperative images demonstrated that there was no residual tumor or that no critical brain tissue was at risk of compromise. Additional lesion, not otherwise apparent, could be removed and other unnecessary dissection was avoided when the new image confirmed that the surgical goals had been achieved. Additional time incurred with use of the iMRI diminished with increasing experience. There were no iMRI-related complications during surgery.
Even though low-field iMRI has certain limitations such as low resolution image or complicated procedures, it has definite advantage rather than conventional surgical technique. iMRI-based neuronavigation offers near-real time brain images and is very helpful, especially in the cases of localization of critical structures in considerable brain shifting situation. Moreover, this system is feasible for tumor resections that require multiple other surgical adjuncts including cortical mapping, monitoring of somatosensory evoked potentials, motor evoked potentials or electrocorticography. Standard neurosurgical drills, microscopes, and other equipment can be used safely in conjunction with this iMRI system.
Pleomorphic xanthoastocytoma (PXA) is a rare glial neoplasm involving mainly the supratentorial compartment of children and young adults. Despite many reports, there is little knowledge about long-term results of the treatment of this tumor. Between 1990 and 2006, 34 patients with diagnosis of PXA were operated on at the Department of Neurosurgery of Verona. The mean age at diagnosis was 34 year old (range 12–65). 18 patients were females and 16 males. Symptoms at presentation were seizures (25/34), headache (3/34), and neurological deficits (6/34). Locations of the tumor were: temporal (16/34), frontal (6/34), parietal (6/34), occipital (1/34), paratrigonal region (1/34), hypothalamus (1/34), quadrigeminal plate (2/34), and cerebellum (1/34). Neuroradiological appearance was of solid-cystic lesion in 23 cases, solid lesion in 8 cases, and purely cystic in 3 cases. All patients were treated surgically: in 27 cases the resection was radical, in 6 cases subtotal, and only in 1 case a simple biopsy was performed. In 28 patients diagnosis of grade II PXA was done, and in the remaining six anaplastic features were described. In those latter cases, or when the excision was incomplete, radiation therapy and/or chemotherapy was also accomplished. At a median follow-up of 67.5 months (range 6–216 months) the overall survival was 76%. Complete surgical resection is the treatment of choice for PXA. In cases of an incomplete excision, or for those cases defined as “with anaplastic features”, radiotherapy and/or chemotherapy should be administered. Nevertheless, also for grade II tumors we recommend a very long-term follow-up (>10 years) because the biological behaviour of PXA is relatively unpredictable.
In the United Kingdom, the National Institute for Clinical Excellence (NICE) has published extensive guidelines for the management of patients with brain tumors. They acknowledge the need for follow up imaging after surgical resection of supratentorial meningiomas but do not advise on imaging protocols. The frequency and duration of post operative imaging could be limited to the time period within which tumor recurrence is likely to occur. Previous studies have concluded that the likelihood of recurrence varies with the histological grade and the extent of resection of the tumor.
Retrospective audit of records of all patients that commenced routine follow up imaging after undergoing surgical resection of supratentorial meningiomas between January 1997 and December 2007.
To assess tumor recurrence and time to recurrence in relation to histological grade (WHO grade 1 or 2) and extent of resection (complete or subtotal).
Records of 143 patients (99 females, 44 males) were reviewed. Age range was 20–85 years. 19 tumor recurrences were identified. Recurrence was significantly more likely in patients with WHO grade 2 tumors (p<0.001) and in patients with subtotal resection (p = 0.032). Median time to recurrence was 28 months. 17 cases of recurrence occurred within 5 years of surgery. The remaining 2 cases were detected at 62 months and 99 months after surgery. The time to recurrence was significantly shorter with WHO grade 2 meningiomas (p<0.001) and with subtotal resection of the tumor (p = 0.003).
Almost all cases of recurrence in this cohort were detected on imaging within 5 years of surgery. It therefore seems reasonable to discontinue routine imaging after this time period. The timing of the recurrences in our study would suggest that imaging at 1, 2, 3, and 5 years post-surgery would be adequate. Patients with WHO grade 2 meningiomas or with subtotal resection have an increased risk of recurrence and consequently may require more prolonged imaging follow up at their supervising clinician’s discretion. However, as the time to recurrence in these patients is likely to be shorter, there is no reason to suggest that they would benefit from a longer duration of imaging follow up than other patients.
Meningiomas sometimes do not increase their volume further after attaining a significant size. This growth pattern does not correspond to constant linear or exponential growth. We attempted to simulate growth patterns of meningiomas with various growth models.
Forty-four patients with meningioma were followed up for 3.1 to 21.7 years (median: 6.8 years) with four or more CT or MR imaging studies. Twenty-five patients including 2 neurofibromatosis cases had 36 asymptomatic tumors. The remaining 19 patients were followed up after surgery with a residual or recurrent tumor. Tumor volume was plotted against time, and relative changes were analyzed with non-linear regression analyses using the Gompertz, logistic, power and exponential curves.
Seven asymptomatic tumors showed no growth. In other 48 tumors, non-linear regression analysis revealed that the time-volume curve corresponded to power growth (R2 = 0.934), exponential growth (R2 = 0.925), logistic growth (R2 = 0.964), and Gomperztian growth (R2 = 0.965). The coefficients of determination (R2) were higher in logistic and Gompertz curves than those of power or exponential growth (P<0.0003). Twenty-eight of 36 asymptomatic tumors had passed the inflection point before or during follow-up and showed deceleration or arrest of growth. Also, 8 of 13 symptomatic grade-1 tumors went over the inflection point post-surgically, while most of them initially showed quasiexponential growth. On the contrary, all 6 atypical meningiomas continued to grow exponentially.
The growth of meningiomas correlated highly with the sigmoid curve like Gompertzian or logistic growth. Although benign meningiomas might eventually decelerate their growth, atypical meningiomas were unlikely to pass the inflection point.
It is well known that intracranial meningiomas can recur in spite of their benign nature and seemingly complete removal. Survival and recurrence is also difficult to assess due to the long period of follow-up that is required.
To determine the prognostic factors for survival and recurrence after surgical treatment through an ambispective study and identify patients at higher risk of recurrence and death. A total of 208 patients with a mean age of 56.6 years old, and a female to male ratio 3:1 were operated with a minimum follow-up of 60 months. A database was created using patients clinical records with data that included among others: age, gender, location, Karnofsky status, and type of resection. Data was analyzed using a SPSS/PC v.12 (SPSS Inc, Chicago,Ill).
The overall survival rate was 90.4% with an estimated survival at 5 and 10 years of 96.1% and 95.6% respectively. Global recurrence was 29.8%. For recurrence, the free recurrence period was 68.3 months.
Univariate prognostic factors for increased survival include gender (p = 0.03) and complete surgical resection (p<0.001). Univariate prognostic factors for recurrence include gender (p = 0.03), grade of resection (p<0.001) and intratumoral hipodensity in imaging (p = 0.002). A multivariate analysis found that resection according to Simpson grades (HR = 4.5; CI 95% 2.1–9.5) and intratumoral hipodensity in imaging(HR = 2.9; CI 95% 1.3–6.5) are independent factors that correlate to a higher risk of recurrence.
Invasive skull base meningiomas are challenging from the surgical point of view. Although benign, meningiomas of the skull base are locally invasive and have a high recurrence rate.
A total of 136 patients (82 females and 54 males) with intra- and extracranial meningiomas aged 11–72 years (median age = 49 years) were operated on by different variants of bifrontal and orbito-zygomatic approaches between 1998 and 2005. Tumor locations include anterior cranial fossa with or without extension to cribriform plate, ethmoid sinus, sphenoid sinus, orbit, subtemporal fossa, and nasal cavity. Most patients had benign meningiomas. Skull base reconstruction was usually performed with a pericranial flap, and/or abdominal fat. A temporal muscle flap was also used in some patients. The clinical outcomes were graded as follows: good (complete recovery), moderate (improvement), and poor (unchanged or severely disabled).
Total tumor resections were achieved in 67 cases, subtotal or large resections in 56 cases, and partial resection in 13 cases. Of 136 treated patients, good and moderate outcomes were observed in 81 and 44 patients, respectively, while 11 patients had poor outcome and 2 patients died postoperatively from pulmonary embolism within 10–15 days after surgery. Postoperative complications included CSF leakage (8 cases), infection (4 cases), transient impaired vision (9 cases), oculomotor deficit (18 cases), anosmia (12 cases).
Total tumor removal at an early stage is the recommended therapy for the most patients but is not always feasible because of vital structures involvement at the skull base. Extensive tumor resections with intraoperative damage of cranial nerves lead to significant neurology deficiency, thus the optimal treatment of invasive skull base meningiomas is subtotal resection with subsequent radiotherapy.
Meningiomas are the second most common primary tumors of the brain after gliomas. Many therapies are currently used ranging from surgery to radiotherapy including stereotactic radiosurgery and stereotactic fractionated radiotherapy. The goal of the present study was to analyze our series of patients with meningiomas of any location except skull base, treated by either single or combined therapies and see whether implantation was a prognostic factor for tumoral control.
Retrospective observational study in a single center. Between January 1983 and December 2007, a total of 161 meningiomas were treated at our center. All consecutive patients with a meningioma, excluding skull base meningiomas, were enrolled in this study. 42 women and 34 men were treated and median age at diagnosis was 61.48 years. Of the 76 meningiomas treated, 55 underwent single treatment and 21 combined treatments. Tumor volumes ranged from 0.2 to 113.1 cm3 (median 13.2). We compared the group of parasagittal meningiomas (28/76) to the other non skull base meningiomas.
Median follow-up was 30 months (range 9.4 to 288.6) including surgical records, MRI imaging, discharge letters, and histological records. Parasagittal meningiomas had a median volume of 13.4 cm3. 53.57% had only surgery and 46.43% had combined treatments. Their grade was WHO grade 1 in 69.23%, grade 2 in 26.92%, and grade 3 in 3.85%. When surgical resection was performed 47.62% were Simpson 1, 28.81% Simpson 2, and 28.57% Simpson 4. There was no significant difference in sex, age, size, grade, type of treatment, and Simpson between parasagittal and other locations. Nevertheless tumoral control was lower for parasagittal meningiomas (62.96%) and time to recurrence was significantly lower as well: 61.9 months vs. 159.8 months (p = 0.022).
Our data have shown that although parasagittal meningiomas have the same epidemiologic, morphologic, histologic distribution, and treatment modality than meningiomas with other locations, their behavior is different. Parasagittal location was identified as a statistically significant prognosis factor for higher recurrence rate and lower tumoral control.
The purpose of this study is to evaluate the preoperative radiologic factors associated with tumor-brain adhesion in intracranial meningioma. In particular, we focused on usefulness of MRI and DSA in predicting tumor-brain adhesion during surgery.
The subjects were 79 patients with intracranial meningioma who underwent tumor excision at which time neurosurgeons examined the tumor-brain adhesion. The authors evaluated age, sex, the preoperative MRI finding including tumor volume, peritumoral rim, shape of tumor, signal intensity of tumor obtained T2WI, presence of peritumoral edema, and presence of pial blood supply in DSA.
Of the various factors, male sex (p = 0.012), the presence of peritumoral edema (p = 0.001), shape of tumor (p<0.001), tumor volume (p = 0.000), and pial blood supply (p = 0.000) were significantly correlated with tumor-brain adhesion. Age (p = 0.153), the signal intensity of tumor parenchyma in T2WI (p = 0.161), and peritumoral rim (p = 0.170) were not statistically significant with the tumor-brain adhesion.
We retrospectively examined the findings of MRI and DSA performed before meningioma removal and clarified the relationship between the findings and tumor-adhesion. Male, larger tumor, irregular margin, peritumoral edema, and pial blood supply should be considered high-risk group of meningioma removal, especially in eloquent area.
In a previous report we developed a grading system called Clinical-Radiological Grading System (CRGS), in order to standardize surgical indications in elderly patients (older than 70) affected by intracranial meningiomas. On the basis of the results we have demonstrated that patients with a score lower than 10 had a bad prognosis regardless of surgical treatment, those with a score between 10 and 12 had a prognosis positively influenced by surgery, and those with a score higher than 12 had a good prognosis regardless of surgical treatment. In the present study we analyzed, during 6 years (2000–2006), a series of 108 surgical patients older than 70 affected by intracranial meningioma in which the selection for surgery was made by the means of the CRGS as a predictor of outcome. The series is composed of 58 women and 50 men ranging from 70 to 86 (mean age 75). The lesion location was the following: convexity (20), falx (13), parasagittal (48), olfactory groove (3), presellar region (1), tuberculum sellae (1), sphenoid ridge (2), middle fossa (3), tentorial edge (6), cerebellar (3), ponto-cerebellar corner (6), and intraventricular (2). In three cases multiple meningiomas were demonstrated. Hystological examination revealed atypic meningioma in 15 cases. Neurological symptoms were the following: seizures (13), motor disturbances (19), cognitive and language problems (18), gait and balance disturbances (17), cranial nerves deficits (6), and visual impairment (6). Two patients were asymptomatic and the diagnosis was made by the presence of cranial hyperostosis. The associated diseases were the following: arterial hypertension (16), cardiac diseases (13), lung diseases (2), diabetes (5), thrombocytopenia (2), hypothiroidism (2), glaucoma (2), vascular diseases (3), and kidney disease (1). The mortality rate within 3 months after surgery was 7% and the 1-year mortality ratio was 15%. Post-operative complications and the presence of the above mentioned concomitant diseases influenced the duration of the hospitalization: the median length was 19.5 days (range from 5 to 41 days). Post-operative complications occurred in 19 cases: broncopneumonia (2), pulmonary embolism (5), worsening of neurological deficit (7), status epilepticus (2), partial seizures (2), diabetes decompensation (1), chronic subdural hematoma after 1 month (1), and acute subdural hematoma requiring evacuation (1). Eighteen patients were transferred in a rehabilitation unit. The presence of post-operative complications and the length of post-operative rehabilitation influenced the quality of life of the patients in term of self sufficiency (Karnofsky performance status) and cognitive status (Mini Mental State Examination) at follow-up examination.
The present study is aimed at finding radiological parameters that can provide indirect information on invasive growth of meningioma relevant enough to predict the likely risk of postoperative neurological deficit development.
The cohort is made up of 40 consecutive adult patients (from January 2004 till May 2005) of comparable general condition parameters (age 18–75 years, KRS 70–100, ASA 1–2) with meningiomas solely attacking brain tissue with the whole of their volume. The Pearson chi-square test was used for statistical evaluation.
Radical resection of the meningioma was attained in 33 (82.5%) patients, subtotal resection in seven (17.5%). Ten (25%) patients at 7 days after the operation had their neurological findings worse than before. Seven were found to have a new neurological deficit and there were three cases of progression of the existing neurological symptoms. Three patients (7.5%) were worse off neurologically than before the operation as long as 3 months after surgery while seven had their neurological condition restored ad integrum. All of the ten patients with postoperatively worsened neurological findings had their meningiomas localized in the eloquent area. A correlation was found between the eloquent area and neurological deficit, and also between the presence of peritumoral edema (small, medium, large) and neurological deficit. Interdependence was detected between a discernible tumor-brain interface and absence of edema, between a discernible tumor-brain interface and dural type of vascular supply, and between the dural type of vascularization and absence of edema.
As follows from the outcome, meningioma growth in the eloquent area and the presence of peritumoral edema are the two adverse parameters predicting the development of postoperative neurological deficits. In contrast, dural types of vascularization, visible tumor-brain interface, meningioma growing in a non-eloquent area, and the absence of peri-tumoral edema are favorable predictive parameters. To go by the results, in the presence of the last two parameters the patient need not be exposed to the risks of invasive selective angiography.
In meningioma WHO grade I, long term tumor control with minimal side effects is achievable with stereotactic radiotherapy (SRT) and radiosurgery (SRS). However, precise target volume (GTV) definition is critical due to the possible infiltration into venous sinus, bone, and along cranial nerves or vessels and dura. Meningiomas show a high 11C methionine uptake and can be visualized by PET (MET-PET). In this study we tested the hypothesis that the addition of MET-PET would result in a smaller GTV and to a radiotherapy dose reduction on critical brain structures.
20 patients were studied with pre SRT/SRS CT, GAD enhanced T1 3D MRI and MET-PET (ECAT HR+). GTV was defined by MR/CT only, MET-PET only, and by the combination (final GTV). MET-PET GTV was defined based on the window at the tumor to brain interface. GTV calculated from CT/MR only, MET-PET only and final GTV were compared using paired samples statistics. Discrepancies in GTV areas were analyzed. In 5 patients with skull base meningioma, SIMRT or SRS was performed using MR/CT GTV and final GTV. Mean dose reduction on hippocampus, chiasm, and brain stem was calculated. Patients were treated with NOVALIS and BRAIN LAB Iplan/Brainscan.
Meningiomas were visualized in 18 patients, in 2 patients MET-PET was negative possibly due to small tumor volume. MET-PET defined GTV were significantly smaller compared to CT/MR based GTV, mean 8.7 cm3 (range 0.83–27.9) and 15.9 cm3 (range 2.5–61.6) (p = 0.02). The addition of MET-PET did not led us to significantly reduce the final GTV compared to CT/MR (mean 12.1 cm3 range (1.75–36.2) (p = 0.06). MET-PET adequately identified the main tumor mass but not CT/MR suspected small extensions at the tumor margin below the PET camera resolution (5–6 mm) (along dural tails (89% of cases), blood vessels (11%), and cranial nerves (39%). In several cases larger extensions with diameter > 5 mm suspected on MR/CT were negative on MET-PET. Here MET-PET showed no infiltration of cavernous sinus (35% of cases), sagittal sinus (22%), and skull base (17%). In 5 patients with absent cavernous sinus infiltration on MET-PET the final GTV was reduced compared to the MR/CT GTV and a radiotherapy dose reduction on hippocampus (mean dose 31% (MR/CT plan) vs. 23% (MR/CT/PET plan), brain stem (24% vs. 18%), and chiasm (82% vs. 72%) was demonstrated.
MET-PET GTV are smaller compared to MR/CT GTV although we did not significantly reduce final GTV. In patients with suspected cavernous sinus infiltration on MR/CT the addition of MET-PET reduced final GTV with a dose reduction on critical brain structures.
Brain tumor progression is partly dependent on the cells’ ability to invade and grow into surrounding brain. Development of an assay to assess both tumor invasion as a measure of aggressive behavior and the inhibition of that invasion through application of chemotherapeutics could provide clinically valuable data regarding an individual patient’s response to treatment.
Using an established ex vivo invasion assay, tissue samples from 14 surgical patients radiographically diagnosed with meningioma were assessed. The actual invasion distance (μm) of tumor cells were monitored using video-microscopy over 5 days and WHO grades pathologically confirmed.
Video-microscopic surveillance revealed that 13/14 samples demonstrated no cellular invasion. Neuropathological diagnosis revealed eleven WHO grade 1 meningiomas (benign), one WHO grade 2 meningioma, and one hemangiopericytoma. The single WHO grade 3 meningioma demonstrated significant tumor invasion and inhibition by cisplatin.
These laboratory observations suggest a correlation between the histopathology (WHO grade) of a meningioma and its invasiveness (a characteristic of malignant behavior). This invasion assay may be used to assess tumor invasion and possibly response to chemotherapeutics to help individualize adjuvant treatment regimens.
There is no established standard chemotherapy for recurrent glioblastoma (GBM). In a previous trial (NOA-01), the combination of nimustine and teniposide showed efficacy in previously untreated GBM. After temozolomide has been established as the standard for primary therapy of GBM, nimustin (ACNU) and teniposide (VM-26) are now in use as salvage chemotherapy for recurrent GBM. However, there are no data on toxicity and efficacy of this regimen in recurrent GBM
In two neurooncological centers, all patients with recurrent glioblastoma who had received nimustine (90 mg/m2, day 1/42) and tenoposide (45–60 mg/m2, days 1–3/42) chemotherapy were analyzed retrospectively for progression-free survival, overall survival, and toxicity.
Thirty-five patients (median age 51 years, range 25–71 years), most of them pretreated with temozolomide were identified. The rate of progression-free patients at 6 months after initiation of therapy was 29%. The median overall survival after initiation of nimustine/teniposide was 6 months and 23% of patients survived 1 year or longer after initiation of nimustine/teniposide. Grade 4 hematotoxicity was observed in 12 of 35 patients (34%) and in 14 of 83 evaluable chemotherapy courses (17%). No high-grade non-hematological toxicity was observed.
These data support the efficacy of nimustine and teniposide combination chemotherapy in recurrent GBM. Nimustine and teniposide combination therapy is associated with a comparably high rate of high-grade hematotoxicity.
Despite new strategies the prognosis of recurrent malignant gliomas (MG) is dismal. MG are characterized by vascular proliferation, with overexpression of proangiogenic factors and receptors such as VEGF-A and VRGFR-2. Some prior studies have explored the potential role of combined treatment with irinotecan and bevacizumab. The aim of this study is to confirm efficacy and safety of this combination in non-selected consecutive patients (pts).
Data from 7 Spanish centers were collected retrospectively. All pts were >18 years, had to sign an informed consent and to present: histological documented MG; progression after radiation and temozolomide; measurable disease on MRI; have received at least 3 infusions of treatment schedule (irinotecan 125 mg/m2 and bevacizumab 10 mg/kg every 2 weeks) for a maximum of a year. Response rate (RR) was determined by MRI (performed every 6 infusions) using MacDonald criteria. Kaplan-Meier and multivariate analysis (MA) (Cox model) were performed to determine progression free survival (survival until progression or death) (PFS) and overall survival (OS).
From July 2006 to December 2007, 54 pts (35 glioblastoma, 15 anaplastic astrocitoma, 4 anaplastic oligoastrocitoma, and 1 anaplastic oligodendroglioma) were treated with this schedule. Median age: 52.3 years (25–74). Median number of prior chemotherapy: 2 (1–4). Six cases had received irinotecan previously. 33 pts (62%) with KPS≥80% and 33 pts (63%) had neurological symptoms. 33 pts received dexamethasone with median dosage: 34 mg (1–18). 22 were taking a non enzyme-inducing antiepileptic drug. The median cycles of the schedule: 9 (4–23). Toxicity grade 3–4: Asthenia in 6 pts; thrombocytopenia 1; neutropenia 1; mucositis 2; thomboembolic complications (TEC) 3; skin 1; hemorrhage out of CNS 2; and severe cognitive impairment 4. Efficacy: 31 pts had response (4 complete response); 12 pts stable disease ≥3 months and 9 progression as best response. RR was 59.6% (95% CI [45.1–73]). Mean follow-up: 5.8 months (1.9–17.9). Mean PFS was 7.4 months (6–8.6). OS was 16.17 months (8.3–24). In the univariate analysis for response no statistical difference was seen for gender, KPS, age, and dexamethasone. KPS <80% (HR 6.1 [1.3–28.2]) was significant for OS, but there was no signification for dexamethasone (HR 2.96 [0.8–10.7]). KPS<80% (HR 4.4 [1.4–13.4]) was significant for PFS, but not dexamethasone (HR 2.2 [0.9–5.3]). In the MA, only KPS was significant for OS and for PFS.
Combination of bevacizumab and irinotecan in recurrent MG improves RR, EFS, and OS when compared with previously reported data. However, this regimen is not free of severe toxicity (TEC and severe cognitive impairment) and requires a careful selection of patients.
After first published data on this treatment regimen, we conducted a prospective evaluation of this kind of chemotherapy in relapsed glioma patients. No validated standard treatment is available to substantially improve outcome.
14 adult patients with recurrent glioma were included. 12 GBM, 1 gliomatosis cerebri, and 1 brain stem glioma. 10/14 had 2 or more relapses. Mean age at first diagnosis was 43 years (range 18–72). 13 patients underwent surgery (4 total, 6 subtotal, 3 biopsies). The patients with gliomatosis and brainstem glioma had no radiation therapy. All others were treated with standard radiotherapy (60 Gy) and concomitant temozolomide, followed by various cycles of adjuvant temozolomide until 6 cycles or less, if relapse occurred. After progression second line chemotherapy contained other chemotherapy regimen. Treatment with irinotecan 125 mg/m2 and Avastin 10 mg/kg were administered every two weeks. Dose reduction was necessary in 5 patients due to toxicity; only non–enzyme inducing antiepileptic drugs were allowed; oral anticoagulants were prohibited. Response evaluation by MRI using MacDonald’s criteria and rCBV measure was performed.
Radiographic responses were noted at first evaluation after 6 weeks in 12 of the 14 patients (86%); 1 patient with gliomatosis remains stable, and 1 GBM patient progressed. At this time 3 patients progressed after 3 months; all others did not reach PFS yet. Toxicity was mild with no grade 4 hematoxicity, moderate, and rare gastrointestinal symptoms.
This combination chemotherapy has exciting response rates with acceptable toxicity. We can confirm earlier studies on this treatment. PFS rate at 6 months and TTP will be presented at the meeting.
Erlotinib and Gefitinib (EGFRinhib) are tyrosine kinase inhibitors specifically targeting epidermal growth factor receptor (EGFR). We present data of an exploratory study analyzing EGFRinhib monotherapy in patients with recurrent/progressive malignant glioma.
21 patients with recurrent/progressive malignant glioma were included in our study. EGFRinhib administration was started at a median of 1.8 years (range 0.54 to 10.95) after primary surgery. 20/21 patients had undergone radiotherapy and all patients had received at least one (range 1 to 5, median 2) line of previous systemic anti-neoplastic therapy. Patients orally received a daily dose of 100 mg or 150 mg Erlotinib or 250 mg Gefitinib.
Median age at initial diagnosis was 47.9 years (range 31.9 to 76 years). 18 patients received a total of 92.8 months (median 3.03) of Erlotinib treatment and 3 patients received a total of 16.1 months (median 6.06) of Gefitinib treatment. The best responses were partial remission in one patient under Erlotinib treatment and in two patients under Gefitinib treatment, respectively. Median time to progression was 3.05 months. Six months after start of EGFRinhib treatment, 4/21 (19%) patients had not experienced tumor progression and 6/21(29%) patients were alive. Expression of EGFRwt, EGFRvIII, PTEN, or EGFRvIII/PTEN co-expression in tumor cells did not significantly correlate with time to tumor progression or survival time. EGFRinhib administration had to be discontinued due to toxicity (grade 3 rash) in only one patient.
EGFRinhib monotherapy is associated with therapeutic efficacy in only a small percentage of malignant glioma patients. Reliable biomarkers predicting tumor response to EGFRi need to be identified.
Recurrent malignant glioma has a dismal prognosis and therapeutic options are scarce. Nimotuzumab (N) is a humanized mAb against the EGFR extracellular ligand binding domain with high affinity and specificity. After previous potentially encouraging reports on N in this setting, N with chemotherapy was applied to patients with recurrent malignant glioma in an institutional series. In a retrospective analysis, 10 patients with recurrent high-grade glioma and measurable disease (WHO III, n = 4; WHO IV, n = 6) all with EGFR amplification were treated with N (200 mg weekly × 6 as induction and twice a month for 6 months afterward) in combination with temozolomide in 3 patients (2 patients 200 mg/m2/day for 5 each 28 days and one 75 mg/m2 for 15 days out of 30) or lomustine (110 mg/m2 doses each 40 days) in 7 patients. The MacDonald criteria were employed for response assessment. The response was documented by MRI in week 8 and assessed afterward every 12 weeks. All the toxic events were seen with the combination of lomustine + N. One episode of grade 2 diarrhea and one event of seizure at the moment of the infusion were seen. Grade 3 hematological toxicity was observed in 4 of 7 and grade 2 in 1 of 7. One patient was hospitalized for febrile neutropenia. Three of 10 patients progressed at 4, 8, and 10 weeks; 2 had some evidence of improvement in neuroimaging but less than the 25% limit to assess them as partial responders (one progressed at week 19, and the other are still in response at week +10), and 5 had stable disease (2 progressed at week 12 and 30 and 3 continued in stable disease for +10, +10, and +23 weeks). Six of 7 patients who received the combination with lomustine had moderate to severe hematological toxicity. The potential of this approach should be further refined and explored.
GBM is a highly malignant tumor with a median survival of less than 15 months. Dysregulated signaling of platelet derived growth factor receptors (PDGF-Rs) is implicated in pathogenesis. Imatinib (I) plus Hydroxyurea (HU) is effective in patients (pts) with recurrent progressing GBM. In a pilot group of 30 pts with recurrent GBM the progression free survival at 6 and 24 months was 32% and 16% respectively. 37% of pts achieved disease stabilisation (SD). Despite the aggressive course of GBM, short periods of SD after primary treatment or effective treatment of relapse is common. The current Phase II study was designed to analyze the efficacy of I plus HU treatment in GBM pts with SD as sequential maintenance treatment.
From Dec. 2003 up to June 2005 30 pts were included. No enzyme-inducing anticonvulsive drugs were allowed. I at the dose of 600 mg and 1000 mg of HU were given as a continuous daily treatment, follow up included blood cell count weekly and magnetic resonance imaging (MRI) every 6 weeks. The primary endpoint is 12 months progression free survival (PFS).
All pts are eligible for safety, for 6, 12, 24, and 36 months PFS and 6, 12, 24, and 36 months overall survival (OS); 25 pts were male, 5 pts female, and the median age was 44 years (32 to 71). 6 pts had primary GBM 24 pts scondary GBM following lower grade malignancy. All 30 pts had prior irradiation, 21 pts had temozolomid containing, and 9 pts nontemozolomid containing regimens. 8 pts were free from relapse, 17 pts after first, and 5 pts after second relapse. 25 pts had measurable disease in MRI scan, and 5 pts had no evidence of disease. The median follow up was 40 months. The best response was partial remission in 4 pts. SD for more than 3 months in 20 pts. 6, 12, 24, and 36 month PFS was 60% (18/30), 40% (12/30), 17% (5/30), and 17% (5/30). 6, 12, 24, and 36 month OS was 90% (27/30), 67% (20/30), 37% (11/30), and 17% (5/30). 3 of 5 pts with PFS of more than 3 years had secondary GBM (3/6). Hematotoxicity grade 2 and 3 occurred in 11 out of 30 pts (anemia grade 3: 2 pts; anemia grade 2: 4 pts; leukopenia grade 3: 2 pts; grade 2: 7 pts; thrombopenia grade 2: 4 pts) and required dose reduction of HU in 8 pts, dose reduction of I in 1 patient and G-CSF subcutaneously in 8 pts.
I (600 mg/d) and HU (1,000 mg/d) showed efficacy as maintenance treatment in pts with GBM in stage of SD with low toxicity profile. Data suggest that efficacy might be higher in pts with secondary GBM. Confirmation is necessary.
Temozolomide (TMZ) is the standard of care for patients with newly diagnosed glioblastoma (GBM) as well as recurrent anaplastic glioma (AG) and GBM. Recent data suggest enhanced efficacy of alternative schedules of TMZ at recurrence. Few data have been published on the efficacy of TMZ after the failure of TMZ using conventional dosing regimens.
A retrospective review of patients with recurrent glioma rechallenged with TMZ was conducted.
A total of 90 rechallenges were identified in 81 patients, 14 with low-grade glioma (LGG), 22 with AG, and 45 with GBM. Median age was 46 years with median Karnofsky Performance Score of 90 at rechallenge. Partial (PR) or complete responses (CR) to TMZ were observed in 32.0% of AG and 11.1% of GBM. Six-month progression-free survival (PFS) was 41.2% in LGG, 48% in AG, and 31.9% in GBM. Relevant toxicity (NCI-CTC grade 3 to 5) was observed in 32 of 90 rechallenges. Discontinuation of rechallenge TMZ was due to toxicity in 6 patients and due to progressive disease (PD) in 60 patients.
TMZ was well tolerated and generated a high response rate in patients who had previously failed TMZ. These data suggest that rechallenge with TMZ in patients who previously demonstrated disease stabilization with TMZ treatment and also patients who progressed on TMZ should be further investigated in a randomized study.
Despite advances in different therapy modalities, the treatment of recurrent glioma remains challenging. Chemotherapy is a therapeutic option and BCNU (carmustine) is one of the drugs used in this context. However, its preference to other nitrosoureas remains controversial due to a possible development of pulmonary fibrosis. Besides, little is known about its effectiveness in case of recurrence. Therefore we sought to determine safety and efficacy of BCNU treatment in recurrent glioma.
From 1995 until 2005, data of patients with recurrent glioma were retrospectively evaluated in view of possible side effects of BCNU treatment that were classified according to the NCI Common Toxicity Criteria (CTC) version 2.0. For outcome analysis, data of patients with recurrent GBM receiving either BCNU or no adjuvant therapy (control) were analyzed for known prognostic factors, overall survival (OS), and survival from relapse (SV). Univariate analysis was done by Kaplan-Meier estimates and multivariate analysis by a modified Cox regression model.
163 patients with recurrent glioma received a median total dose of 1662 mg BCNU during a median number of 5 cycles. Indications for BCNU treatment were mainly recurrent glioblastoma WHO IV (n = 84) and anaplastic glioma WHO III (n = 57). Few patients received BCNU for recurrent low-grade glioma WHO II (n = 22). 88 patients (54%) experienced drug-related side effects which were usually of mild nature (CTC I/II; 45% of all patients). Severe side effects CTC III/IV were rarely observed (9% of all patients). Only one patient (0.6%) experienced a clinically relevant pulmonary fibrosis CTC IV requiring mechanical ventilation. Outcome analysis of 134 patients with recurrent GBM revealed beneficial effects of adjuvant BCNU therapy. Both OS (17 vs. 12 months) and SV (10 vs. 4.5 months) were significantly prolonged compared to controls. In a multivariate analysis, BCNU treatment was shown to be an independent prognostic factor for prolonged survival.
In our patient sample, BCNU turned out to be a well tolerated cytostatic substance with mostly mild side effects. Contrary to the literature, the most dreaded side effect, a clinically relevant pulmonary fibrosis, was detected in one out of 163 patients only. In case of recurrent GBM, BCNU treatment proved to be an independent prognostic factor for prolonged survival. Therefore, in recurrent glioma the use of BCNU might be preferred over more aggressive chemotherapy regimens in terms of safety and patient outcome.
To evaluate the efficacy and tolerability of ifosfamide, carboplatin, and etoposide (ICE) in patients with recurrent glioblastoma.
This was an open-label, single-center phase II trial. Forty-two patients with progressive glioblastoma after surgery, standard radiotherapy, and a first-line temozolomide-based or ACNU-based chemotherapy, were enrolled. The primary endpoint was progression-free survival at 6 months (PFS-6), and secondary endpoints were response rate, toxicity, and survival. Chemotherapy consisted of Ifosfamide (700 mg/m2 on days 1, 2, and 3), carbopaltin (100 mg/m2 on day 1), etoposide (70 mg/m2 on days 1, 2, and 3), every 6 weeks. PFS-6 was 37%. The median PFS was 17 weeks. Response rate was 27%. Adverse events were generally mild (grade 1 or 2) and consisted mainly of alopecia.
This regimen is well tolerated and has activity in patients with progressive glioblastoma.
We evaluated retrospectively the efficacy and toxicity of sterotactic radiotherapy or radiosurgery retreatment in recurrent gliomas.
Between April 2005 and September 2007, 57 patients (58 lesions) with high-grade glioma, 26 (45%) WHO grade III (AA) and 32 (55%) grade IV (GBM), were treated with stereotactic radiotherapy. Minimum follow-up was 6 months. Forty patients underwent chemotherapy with temozolomide (75–150 mg/m2) or with cisplatin (100 mg/m2) and carmustine (100 mg/m2) for 5 cycles. Conformal radiotherapy after surgery or biopsy was made in all patients with a median dose of 60 Gy (2 Gy/fract). Median time between conventional and stereotactic radiotherapy was 10 months. Thirty-one lesions were treated with single fraction, while 27 with hypofractinated schedule. Median treated volume was 5.3 cc (range 0.6–33.9). Median dose was 14 Gy in radiosurgery and 20 Gy in 3 fraction in streotactic treatment; median isodose was 78%.
Median survival time after stereotactic retreatment was 16.95 months (CI 95%: 11.1–22.8 months). Twenty-one lesions have a progression disease (63.6%). Median time to progression was 10.25 months (CI 95%: 5.6–14.9 months). Toxicity: We observed 2 acute toxicities, consistent in vasogenic edema 2 months after the treatment, controlled by high doses of steroids. Late toxicity affected 2 patients: In the first case after 15 months of follow up an MRI image consistent in pseudotumoral radionecrosis was observed and subsequently confirmed by spectroscopy, PET, and Histology. A second case of a radionecrosis MRI image 11 months after SRS was evaluated as a recurrent tumor in PET imaging.
Stereotactic radiotherapy with Cyber Knife could be a therapeutic option in recurrent and multifocal high-grade gliomas. The noninvasive fixation of the skull used in Cyber Knife System allows the choose of fractionated schedules.
We have applied a form of tumor selective particle radiation known as boron neutron capture therapy (BNCT) to malignant brain tumors. We report the survival benefit of BNCT for newly diagnosed glioblastoma patients with special reference to the results of RTOG recursive partitioning analysis (RPA) classes (submitted for publication). As there has been no standard treatment for recurrent gliomas so far, it has been difficult to evaluate the survival benefit of BNCT for recurrent malignant gliomas. Here we introduce the survival benefit of BNCT for recurrent malignant glioma patients, with special reference to new RPA classification based on phase 1 and 2 trials planned for recurrent malignant gliomas advocated by New Approaches to Brain Tumor Therapy CNS Consortium (NABTT) in J Clinical Oncology (25:2601–2606, 2007).
Since 2002, we have treated 22 cases of recurrent malignant gliomas with BNCT. All cases had been treated by standard treatment including radiotherapy and chemotherapy mainly by X-ray treatments and temozolomide or nimustin prior to BNCT. After BNCT, radiographical improvement and aggravation were followed by MRI and biological activity was followed by 18F-boronophenylalanine (BPA)-PET. Cause of death was estimated by these modalities. Also, overall survival was evaluated with special reference to RPA advocated by NABTT as above.
All cases showed radiographical improvement at once. The median survival time (MST) of BNCT-treated recurrent gliomas was 11 months, while that of total NABTT cases (N = 333) was 7 months, where 2 studies have the similar distribution of the patients as to RPA. MST of NABTT RPA classes were 25.7, 17.2, 3.8, 10.4, 5.6, 6.4, and 4.9 months for class 1 to 7, respectively. The MST of BNCT-treated cases were 43, 22, 11, 8, 7, 22, and 11, for class 1 to 7, respectively. In each RPA class, BNCT showed good survival benefit for recurrent gliomas patients, although case numbers were limited. Especially BNCT could prolong the poorest RPA class from 4.9 months (NABTT) to 11 months as MST. Three cases were lost due to uncontrollable radiation necrosis, 10 were lost due to CSF dissemination, 5 were lost by local recurrence, and 2 were lost by other cause of death. Two patients are still alive at the time of analysis.
BNCT could prolong the survival of patients with recurrent gliomas not only for the good prognosis group but also for the poor prognosis group. The major cause of death was CSF dissemination after BNCT.
The present study analyzed the role of MR diffusion-weighted (DW) imaging in the detection of treatment volumes for radiotherapy. Literature data show that MR DW images may be able to detect tumor infiltration into peritumoral edema and into normal-appearing white matter (WM) allowing a better delineation of clinical target volume (CTV).
Twelve patients affected by high-grade gliomas (WHO III–IV) operated and candidate to adiuvant chemo-radiotherapy have been enrolled thus far.
Prior to radiotherapy, they underwent MRI study by MR DW imaging. The apparent diffusion coefficient (ADC) and the fractional anisotropy (FA) were analyzed. ADCs and FA were measured on volumes of interest generated from isotropic DW images represented by the enhancing tumor, hyperintense regions adjacent to enhancing tumor, and the normal-appearing WM adjacent to hyperintense regions. The same measurements were performed on symmetric locations in the contralateral hemisphere.
Mean ADCs values in enhancing tumor and in peritumoral hyperintense regions were increased compared with contralateral normal WM (828.16 mm2/sec vs. 588.33 mm2/sec, and 984.3 vs. 545.5 mm2/sec). No difference was seen in mean ADCs between peritumoral normal-appearing WM and the corrisponding contralateral normal-appearing WM (547.5 vs. 575.5 mm2/sec). No difference was appreciable in mean FA values between glioma regions and normal tissues. Mean FA values in peritumoral hyperintense regions were slightly decreased compared to contralateral WM (0.267 vs. 0.395).
Our data confirm the possible role of ADC in the detection of tumor infiltration in perilesional edema to be included in the CTV. ADC values do not give information on tumor infiltration into normal-appearing white matter (WM), but the limited number of observations does not allow to draw definitive conclusions. The role of FA value deserves further investigation.
The early and accurate prediction of treatment response in brain gliomas after radiation therapy using MRI perfusion and spectroscopy may provide an opportunity to switch to more beneficial to nonresponders instead of the limited delayed structural data obtained via the use of conventional MRI. Our aim was to analyze these recent modalities to evaluate the combined application on the treatment plan.
Forty-six patients with brain astrocytomas were recruited in this study, they were examined prospectively before and after radiation therapy at regular intervals over a period of two years, using conventional MRI, MRI perfusion, and spectroscopy. The findings correlated the histo-pathological grade of the tumor and different tumor response to treatment for each individual patient. For those patients with early poor response or resistant to radiation, chemotherapy was added in the form of temozolamide 175 mg/m2 for 5 days every consecutive 28 days; meanwhile for patients with good response, no further treatment was added with regular follow-up.
Initial base line radiological studies before radiotherapy were correlated with the histopathological grade. Both MRI spectroscopy and biomarkers choline/creatinine, choline/NAA were statistically significant (0.001, 0.004) respectively to differentiate between high grade (III, IV) and low grade (II), as well as MRI perfusion (P= 0.042); meanwhile, conventional MRI data were not statistically significant except for necrosis and edema (P = 0.002). For the assessment of the early response after 4 weeks of completion of radiotherapy, conventional MRI was only significant for the enhancement (P = 0.0001) Biomarkers of spectroscopy were highly significant for choline/creatinine, NAA/creatinine and choline/NAA, with P values 0.005, 0.025, 0.01, respectively. MRI perfusion was also statistically significant with P value 0.004.
Both MRI perfusion and spectroscopy are important and promising new modalities for diagnosis and early therapeutic evaluation of brain tumor response after radiation therapy for proper selection of patients in need of further chemotherapeutic treatment.
Due to the primary treatment of patients with glioblastoma multiforme (GBM) with alkylating drugs (AD) the role of a rechallenge with AD upon recurrence is subject to discussion. A new treatment option with bevacizumab, an antibody against VEGF, and irinotecan, a topoisomerase I inhibitor is currently under investigation. In contrast to an excellent radiographic response rate of 63% in the treatment of recurrent GBM as demonstrated by MRI, clinical progression and survival parameter did not match the radiological response. Since VEGF has been implicated in alteration of the blood-brain barrier (BBB) the radiological response might merely be due to this effect on BBB. We therefore evaluated the response of the regiment on recurrent GBM by early MRI and metabolic investigation by Fluroethyltyrosin (FET)-PET.
At present, 9 patients with recurrent GBM after treatment with radiotherapy and concomitant plus adjuvant temozolomide (TMZ, median 7 cycles) were included in our observation study (median observation time as yet 10 weeks) and treated with 125 mg/m2 irinotecan and 10 mg/kg bevacizumab in a six week cycle. MRI and FET-PET studies were performed prior to initiation, 3 weeks after the first and one month after the third therapy. 2 patients had FET-PET scans 3 days after the first therapy.
All 9 patients showed a decreased contrast agent enhancement on MRI scan after the first 2 therapies. 7 patients showed a homogeneous reduction of contrast enhancement (CE) by a median of 52%. 2 patients demonstrated a heterogeneous reduction of CE. Accordingly, FET-PET scans demonstrated a reduction of metabolic activity in all patients, with a focal high activity in those 2 patients with heterogeneous reduction of CE. FET-PET scans performed as early as 3 days after the first therapy demonstrated in both cases a homogenous and nearly complete reduction of FET activity. Both patients with heterogeneous reduction of CE and with a focal high activity of FET demonstrated multifocal recurrence 1 month after therapy.
The combination of bevacizumab and irinotecan seems to induce an impressive alteration of the CE and metabolic activity of recurrent GBM. Though a final conclusion can not be drawn as yet due to the short observation time, a heterogeneous reduction of CE and metabolic activity might be correlated with early progression. The observation that the reduction of FET activity was observed as early as 3 days after initial treatment indicates a direct effect of this treatment on the BBB. Further observation of the clinical outcome will further illuminate the effectiveness of this therapeutic approach.
In high-grade gliomas, assessment of response by MRI is often difficult because of the frequent persistence of residual, enhanced lesions. Other tools are needed to distinguish viable tumor from radionecrosis, and to identify new high-grade glioma (HGG) foci in a previously diagnosed low-grade glioma.
In a consecutive series of patients (pts) with glioma, different techniques, including MR spectroscopy (MRS), 2-(18F) fluoro-2-deoxy-D-glucose positron emission tomography (GPET), and dynamic contrast enhanced perfusion MRI (MRP), were performed to assess the presence of viable HGG inside uncertain T1 enhanced lesions.
24 pts were included (14 men, 10 women). In 2 pts, 2 suspected recurrences were assessed, resulting in 26 episodes. Tumor location was lobar in 18 pts, deep in 5 pts, multicentric in 1 pt. Initial surgery was only biopsy in 4 cases and resection in 22 cases, with remaining tumor in 21. Histologic diagnosis were obtained in 13 pts: glioblastoma (5), anaplastic astrocytoma (3), anaplastic oligodendroglioma (2), anaplastic ependimoma (2) grade II oligodendroglioma (1). After surgery, 23 pts were given radio-therapy and 19 pts received chemotherapy. Findings in previous MRI were new enhanced lesions in 8 cases and increased enhancement of previous lesions in 18 cases. The reason for carrying out new assessment techniques were to differentiate between radionecrosis and recurrence/remaining tumor in 21 cases, and to confirm a suspected grade increase in 5 cases. MRS was performed in 24 cases, showing a pattern of HGG in 12, radionecrosis in 11, and low-grade glioma in 1 case. GPET showed a hypermetabolic lesion in 8 cases and hypometabolic areas 18 cases. MRP demonstrated increased perfusion in 8 cases and no increase in 10 cases. Discrepancy among different techniques was observed in 8 cases. The positive predictive value (PPV) and the negative predictive value of every technique were as follows. MRI: PPV = 58.3%. MRS: PPV = 92.3%, NPV = 90%. GPET: PPV = 75%, NPV = 50%. PRM: PPV = 100%. NPV = 100%. To differentiate viable tumor from radionecrosis, RMS and PMR reached a PPV and a NPV of 100%, whereas for GPET, PPV and NPV were 66.6% and 60%, respectively.
To distinguish between viable HGG and radionecrosis, gadolinium-enhanced MRI gives a high false-positive rate, while MRS and PMR are superior to GPET to discriminate among tumor recurrence, grade increase, and radionecrosis.
The purpose of this report was to investigate the potential role of positrone emission tomography using 18F-fluoroetyltyrosine (FET-PET) in early detection of disease relapse or progression in high-grade gliomas.
FET-PET (Advance, General Electric Medical System, Milwaukee, WI) was performed in 7 patients (pts) with high-grade gliomas after a neuroradiological suspicious of relapse/progression. Ninety minutes after the intravenous injection of approximately 185 MBq of FET, late images of the brain were obtained. FET uptake in the lesion was semiquantitatively evaluated by measuring the maximal standardized uptake value (SUVmax) and compared with the results of MRI/spectroscopy or histological specimen after surgery.
In the last year (November 2006–2007) 7 pts (6 men and 1 woman; median age, 47; range: 21–69) followed at our institution underwent FET-PET after the evidence of lesion enhancement on MRI. Three had a diagnosis of glioblastoma multiforme (all radically operated) and four anaplastic astrocytoma/oligoastrocytoma (2 partial and 2 complete exeresis). Four pts received radiotherapy (60 Gy in 30 fractions) and concomitant temozolomide (75 mg/msq/daily) and three only radiotherapy. Median time from the end of post-surgical therapy and neuroradiological doubt of relapse/progression was 6.3 months. In 5 pts FET-PET demonstrated an increase of FET uptake (median SUVmax values = 2.9; range: 1.7–3.4) in enhancing lesions on MRI. One patient presented a doubtful uptake in the site of the lesion, while another one presented a “circle pattern” uptake in the site of surgery. Spectroscopy was performed in 5 of 7 pts: in 4 cases it was suggestive of relapse/progression and in 1 case of radionecrosis. One patient underwent surgery for relapse: the histological sample demonstrated an active glioblastoma (KI67 = 15%). After a median follow-up of 26.1 months, 5 patients died for disease progression and 2 patients are still alive with no clear evidence of disease relapse/progression. Concordance between FET-PET and MRI/spectroscopy or histological specimen was showed in 6 of 7 patients (85%).
FET-PET can be useful in the management of high-grade gliomas to detect relapse or tumor progression early after post-surgical treatment (chemo and/or radiotherapy). A longer follow-up and more cases are necessary to confirm this observation.
PET imaging with amino acids has the potential to image glioma invasion better than MRI, which commonly may show nonspecific T2-signal extension beyond the borders of T1-signal enhancement. Our aim is to define the potential impact of FET PET on target volume delineation as compared to MRI.
We retrospectively evaluated 10 patients affected by high-grade gliomas. They underwent both MRI and FET PET as part of their intial work up. An experienced radiation oncologist delineated T1 and T2 signal target volumes based on MRI images. FET PET target volume was defined using a SUV threshold = 1.6, contralateral SUV averaged over a 1 cm diameter region. Fusion of PET and T1/T2 signal MRI images was performed using the PMOD software and the common volume between PET and T1, between PET and T2 (intersection volumes) were computed.
Mean tumor volumes on T1-/T2-signal MRI and PET were 34±31, 21±25 and 128±84 mL, respectively. In one patient, a 1.4 threshold had to be used because of low tumor uptake. There was no significant difference in volumes derived from PET or T1-signal MRI (p = 0.14). PET volumes were were significantly smaller (p = 0.001) than when derived from the T2-signal MRI and represented only 13% of the T2-signal volumes.
Our analysis suggests that FET PET would have an impact on delineating radiation therapy target volumes in most patients and may help in radiation therapy planning. Outcome studies are needed to determine the effect of such PET image-based radiation therapy planning on progression-free and overall survival.
The noninvasive functional characterization of an intracranial lesion represents a major goal, substantially enhancing the neuromorpho-logical information obtained by conventional radiology. We set out to evaluate the contribution of technetium-99m Tetrofosmin (99mTc-TF) SPECT in the evaluation of brain tumors.
Eighty patients (37 male, 43 female, mean age 57.8 years) were prospectively enrolled in the study. The diagnosis was meningioma (27), glioblastoma (18), anaplastic astrocytoma (4), anaplastic oligodendroglioma (3), low grade astrocytoma (4), low grade oligondroglioma (1), radiation necrosis (4), intracerebral hematoma (10), hemangioblastoma (1), colloid cyst (1), swannoma (1), pituitary adenoma (1), and metastasis (5). All patients underwent SPECT imaging and within a week surgical excision was performed. From the surgical excision specimens Ki-67 antigen was assessed with the MIB-1 immunostaining method.
The tumor tracer uptake intensity on visual qualitative image assessment ranged from profound to very faint (in a low-grade fibrillary astrocytoma). 99mTc-TF brain SPECT managed to differentiate tumor recurrence from radiation necrosis and neoplastic from non-neoplastic intracerebral hemorrhage. Furthermore, there a was a linear clear positive correlation between the 99mTc-TF uptake and Ki-67 antigen in gliomas and in meningiomas.
Imaging by 99mTc-Tetrofosmin provides useful information on the lesion’s functional nature and could be implemented in the diagnostic workup.
201Tl-SPECT imaging is used in the assessment of supratentorial primary brain tumors (SPBT): increased uptake of isotope reflects malignant change. Commonly scans are reported using a qualitative inspection technique and described as “no increased uptake” or “increased uptake” indicating low grade and high grade tumor respectively. Previous studies have attempted to relate 201Tl uptake in SPBT to histological grade using quantitative examination of images.
Audit the use of 201Tl-SPECT in the management of SPBT and assess the benefit of routine quantitative analysis.
201SPECT scanning was performed routinely in patients with CT/MR proven SPBT. Scans were performed on a dual-headed gamma camera (Siemens E-Cam) using low energy collimation. Tomographic axial slices, oriented to brain anatomy, were analyzed. Qualitative analysis was performed by an experienced neuroradiologist and a report of “no increased uptake” or “increased uptake” relative to the rest of the cerebral cortex was reported for the tumor site. For quantitative analysis a region of interest (ROI) was drawn on the axial section at the tumor site for scintillation counting. An equivalent “mirror” ROI was also defined in the contralateral hemisphere in the same section and the skull vault. The ratio of scintillation counts in the tumor site to those of the contralateral hemisphere was calculated (T/C) and also tumor to contralateral skull ratio (T/S). Sensitivity, specificity, positive, and negative predictive values were calculated using histology. ROC analysis was performed using histology and T/C and T/S as separate test variables.
Data from 76 patients were audited. With qualitative examination of the 201Tl SPECT scan the test values for scans reported as “increased uptake” being high grade and “no increased uptake” being low grade tumors was: sensitivity 80.0%, specificity 85.4%, positive predictive value 82.4%, and negative predictive value 83.3%. ROC curve analysis of quantitative studies did not indicate significant improvement in predictive yield (area under the curve for T/C and T/S ratio versus histological grade were 0.881 and 0.917 respectively).
201Tl-SPECT imaging has high sensitivity and specificity in the diagnosis of high grade SPBT. In clinical practice quantitative analysis does not confer any benefit over qualitative analysis provided by an experienced neuroradiologist.
Gliomas demonstrate significant heterogeneity both in terms of outcome and imaging characteristics within histological grades and particularly between histopathological and genetic sub-types. Methionine uptake as measured with positron emission tomography (PET) has been shown to correlate with histological grade, but interpretation of the results is difficult unless the histological sub-type is known. We hypothesized that imaging biomarkers of other features known to reflect malignancy including tumor volume, shape, and heterogeneity could be measured from methionine PET data to support this interpretation.
Methionine PET scans performed in 39 individuals from 2003–4 were identified from the database at the Max-Planck-Institute for Neurological Research in Cologne. Scans were analyzed using the ratio to a normal uptake map as previously described. A software application was developed in R to enable standardized, highly reproducible image analysis. Tumor “masks” were defined using a 3D region-growing technique constrained by voxel connectivity. Several statistical measures were recorded for this tumor mask including the volume, characteristics of the border, and variability of the methionine uptake including texture measures. Receiver operating characteristic (ROC) analysis was performed for each of these parameters to evaluate accuracy in the discrimination of (1) grade II vs. III/IV and (2) grade III vs. IV tumors.
Ten factors were selected including peak methionine uptake, the volume of the tumor mask, measures relating to irregularity of the tumor border/shape, heterogeneity within the tumor volume, and the distribution of uptake values within adjacent brain. Each of these features produced significant results on ROC analysis (p = 0.000 to 0.010) with the volume of the region of increased methionine uptake being most effective in separating low and high-grade gliomas. Threshold values were defined from the ROC curves to identify results with high specificity for low-grade tumors. 10 of the 13 (77%) grade II tumors analyzed exhibited 3 or more of these characteristics as compared to 2 of the 26 (8%) grade III/IV tumors. A score derived from these features demonstrated a greater area under the ROC curve (AUC = 0.88) than any of the individual measures in isolation.
PET with amino-acid tracers enables multiple tumor measures to be derived from a single diagnostic investigation. Combined evaluation of multiple parameters in an individual tumor may be used to improve accuracy in the noninvasive characterisation of primary brain tumors.
Differentiation of tumor recurrence/progression from necrosis after radiation and chemotherapy remains a challenging problem. Conventional imaging by MRI is not always reliable, because it can not always differentiate tumor progression from necrosis. The aim of the study was to evaluate the potential of F-DOPA PET in this setting.
Twenty-five patients with recurrent brain gliomas or radio-/chemonecrosis, suspected on the basis of MRI, were studied with F-DOPA PET. For semi-quantitative analysis the portion of the tumor with the highest F-DOPA accumulation was selected and tumor to contralateral normal cortex ratios (T/N) were determined by region-of-interest analysis. Results were correlated with histopathological findings or, in cases without surgery or biopsy, with subsequent clinical course and MRI findings.
Histological examinations showed viable tumor in 11 and necrosis in 4 cases. Seven other cases were considered to have tumor and 3 to have necrosis based on neurological symptoms and findings on follow-up MRI. Using the mean T/N a statistical significant difference (p = 0.004) was found between recurrence of high-grade glioma (n = 17; mean 2.4, 95% CI 2.0–2.8) and necrosis (n = 7; mean 1.5, 95% CI 1.2–1.8). There was no significant difference between high-grade glioma and low-grade glioma (n = 3; mean 1.9, 95% CI 1.4–2.3) or between low-grade glioma and necrosis.
F-DOPA PET demonstrated excellent visualization of high-grade and low-grade glioma. Our preliminary data suggest that F-DOPA PET seems to be a complementary imaging tool in differentiating tumor recurrence/progression from necrosis in contrast enhancing lesions on MRI in patients with high-grade glioma.
Gadolinium-enhanced MR imaging (Gd-MRI), together with the evaluation of the neurological condition of the patient, makes part of the clinical assessment in glioma. Interpretation of contrast-enhancement remains however challenging. In case of clinical worsening without corroborating changes on Gd-MRI, functional imaging could play a role in the assessment of progression. Several amino acids accumulate intensively in malignant cells, as a result of an overexpression of L-type amino acid transport system. The low uptake of these compounds in normal brain makes them valid alternatives for FDG in brain gliomas. We evaluated 2-18F-Fluoromethyl-L-Phenylalanine (FMP), a radio-labeled amino acid, for positron emission tomography (PET) imaging in a prospective series of 23 treated glioma patients (13 males and 10 females, average age 48.6 y [range 25–77]); WHO differentiation grade was I in 1 patient, grade II in 7 patients, grade III in 5 patients, and grade IV in 10 patients. Clinical worsening was observed in 20 patients, consisting on new focal signs, cognitive deterioration and/or increased seizures. Stable clinical condition was noted in 3 patients. Gd-MRI was evaluated as disease-free in 2 patients, progression-free in 8, and progressive disease in 15, respectively. Dynamic PET studies were acquired after injection of 156 ± 57 MBq FMP. Both dynamic and summed data were reconstructed using iterative algorithms, and corrections for attenuation and scatters were applied. Volumes of interest over the tumor and normal brain tissue were manually placed. Accumulation of FMP in nonaffected brain tissue was low, rendering high target-to-background ratios. Intense FMP accumulation at the tumor site was observed in 20 patients, limited focal uptake in 1 patient, and absence of tracer accumulation in 2 patients. Patients with clinical worsening showed intense FMP uptake, except one who presented also disease-free MRI. Second line therapy was considered, except in this latter case, consisting on radiotherapy or chemotherapy in 6 patients and study treatment in 12; 1 patient refused treatment. During a 4 months follow-up period, rapid tumor progression was observed in 10 of them; 4 patients improved and 4 remained stable. In patients without clinical deterioration, intense FMP uptake was noted in 1 case, and slight focal or no uptake in 2 cases. These 3 patients, together with the patient mentioned above, were followed up conservatively, neither showing further clinical worsening nor MRI progression. FMP-PET offered useful information in the assessment of recurrent glioma, significantly correlating with the evolution of the patients (Pearson chi-square 0.011). We therefore believe that FMP-PET could be used in patients showing clinical progression but non-corroborating MRI findings. A further evaluation in a large series is planned.
Helical tomotherapy system (HT), a new intensity-modulation method, combines an intensity-modulated fan beam with the helical motion of the gantry relative to the patient. The higher specificity and sensitivity of 11C-methionine positron emission tomography (MET-PET) in brain tumors has been demonstrated in previous studies and may be useful for delineation of tumor volume. We developed a technique that incorporates routine integration of MET-PET in hypo-fractionated high-dose irradiation using simultaneous integrated boost technique (SIB) by HT. In the cases described below, we examined the combined use of MET-PET and MRI to quantify the effect of MET-PET in defining tumor volume and in monitoring the tumor’s response to radiation for SIB by HT.
We performed SIB by HT to 15 patients of GBM with gross total or subtotal resection. The gross tumor volume (GTV)-1 was defined as the area of intensive MET uptake, including the surgical resection surface. GTV-2 was defined as the area of moderate MET uptake around the resection cavity. The planning target volume (PTV)-1 encompassed GTV-1 plus a 5 mm margin, and PTV-2 encompassed GTV-2 plus a 2 mm margin. SIB by HT was performed in 8 fractions, planning the dose for GTV-1 at 68 Gy (biologically effective dose: BED = 126 Gy), PTV-1 at 56 Gy (BED = 95 Gy), and PTV-2 at 40 Gy (BED = 60 Gy). Concomitant chemotherapy consisting of temozolomide was given to all patients. For the follow-up study, MET-PET was carried out 3 months following SIB in 10 of 15 patients.
Our target planning demonstrated that the conformity index was 95% and the homogeneity index was 107%. No severe acute toxicity was observed in all patients. All patients were able to complete their treatment. In acute phase, radiation necrosis didn’t develop in the area of PTVs. In 9 of 10 patients, the uptake value on MET-PET was considerably decreased or unchanged 3 months following SIB. In just one patient, the uptake value on MET-PET was increased 3 months following SIB. In this patient, local tumor recurrence within 6 months had been demonstrated after all. In our series, actuarial local tumor control rates at 6 months and 1 year were 88% and 67%, respectively. Actuarial overall survival rate at 1 year was 77%.
Our series demonstrate that SIB with HT planning using MET-PET offers excellent target coverage and uniformity, and that MET-PET has great efficacy for monitoring treatment response after SIB. Our regimen of SIB contributed to the control the regional tumors, resulting in better survival of patients.
We have previously shown that 18F-TYR PET imaging accurately delineates irradiated skull base meningiomas with tumor-to-cortex activity ratios ~2.5. The aim of the present study was to evaluate whether 18F-TYR PET may be a good indicator of the response of meningiomas to radiotherapy (RT), which is difficult to evaluate using only MRI.
We here report preliminary results on 4 patients who underwent one PET/CT before RT and another one on the mean 16 months (range 14–23 months) after RT. F-TYR PET/CT studies were acquired 15 minutes after IV injection of 300 MBq F-TYR, using a Gemini Dual system. The metabolic activity was assessed by calculating the tumor to cortex activity ratios.
In 3 patients, the uptake ratio decreased by 25 to 45%. In 1 patient, this parameter slightly increased. The volumes of high uptake remained stable. The patient whose ratio increased received the lowest dose (50 Gy versus 54, 58 and 60 Gy for the others). MRI images showed stable volumes for the three first patients and a reduction for the last one.
Although preliminary, these results suggest that 18F-TYR PET may give indications about post-RT evolution of meningiomas. A longer follow-up and a larger population of treated patients are needed to fully assess the predictive value of this technique. Its ability to be an early predictor of the response to treatment is an intriguing possibility.
Complications due to radiosurgery (SRS) of intracranial meningiomas are reported in particular for parasagittal locations. Post-radiosurgery peritumorous edema can give rise to increased intracranial pressure, seizures, or neurological deficits. Edema formation may be due to vasoactive factors secreted by the tumor and/or reactions related to vascular anatomy like disturbance of circulation. The latter may be caused by occlusion of superior sagittal sinus and bridging veins. Using multimodality imaging registered in BRAIN LAB Iplan/Brainscan complications of radio-surgery vascular occlusion may be anticipated.
A 26 year old female patient was considered for radiosurgery. She had a history of three Simpson 3 resections for a paras-agittal meningioma (WHO grade I) within a time frame of 3 years. Two years after the last operation with excision of the tumor and the obliterated central part of superior sagittal sinus an asymptomatic small recurrence was suspected on MRI at the caudal part of sagittal sinus. Clinically the patient suffered from a persistent grade 3 paresis of the extensors of the right lower limb after the second craniotomy due to a venous infarction of left parietal cortex after tumor resection. Multimodality imaging was performed using contrast enhanced CT, GAD enhanced T1 3D MRI, DTI, MR venography, and 11C methionine PET (ECAT HR+). Studies were registered using IPLAN Image.
With 11C methionine PET meningioma was visualized and discriminated from the patent superior sagittal sinus identified by MR venography and 3D MRI. Using DTI with fiber tracking the corticospinal tract with the left motor strip was identified. With 3D MRI and MR venography draining veins from the left motor cortex were identified. SRS treatment planning showed high dose coverage of the entrance region of the cortical draining veins into the patent part of the superior sagittal sinus. We considered this an unacceptable risk factor for venous occlusion and additonal cerebral infarction and consequently an increase of paresis. Consequently the patient was not further evaluated for SRS, and she was offered fractionated stereotactic radiotherapy.
We demonstrated that multimodality imaging registered in SRS planning could predict worsening of clinical symptoms due to anticipated SRS induced venous occlusion of draining veins into superior sagittal sinus. Multimodality imaging in stereotactic planning can provide important additional information, ultimately leading to alteration of treatment strategy.
Meningiomas add up to 30% of CNS tumors. Atypical meningiomas show a high index of recurrence 5 years after complete resection. Sometimes, meningiomas with histological diagnosis of benign meningioma show clinical characteristics of atypical meningioma. Additional criteria for better classification of meningiomas will improve clinical decisions like resection extension, additional radiotherapy needs, treatment selection, and patient follow up strategy after surgery. In this communication, we show differences in molecular profiles based on High Resolution Magic Angle Spectroscopy (HRMAS) spectra of 30 meningioma biopsies for the distinction between benign and atypical meningiomas.
HRMAS spectra and consequent biochemical profile determination were obtained for 30 samples of human meningioma tissue (20–40 mg), of which 23 were benign and 7 were atypical. The whole HRMAS study was performed at 4ºC. All samples were analyzed by post-HRMAS histopathology to assess the tissue integrity and double validate histological diagnosis. Statistical analysis was performed using in-house MATLAB scripts and the LIBRA statistical multivariate analysis library. One orthogonal component removal was calculated in ramdonly chosen 50% of cases and then applied to the remaining 50%. Principal component analysis (PCA) was applied to the set of spectral vectors. Principal components chosen explained at least 90% of the variance.
NMR spectra showed narrow line widths and adequate signal-to-noise ratios with well resolved spin-spin multiplicities. The comparison among spectra shows clear differences between benign and atypical meningiomas. The phospholipid pattern and other well known signals like alanine and lactate seem to have some correlation with the meningioma grade. Additionally, some signals belonging to unsaturated fatty acids seem to also discriminate the two groups of tumors. PCA analysis show major difference between benign and atypical meningioma for PC1 which major contributions are lactate, glutamate, unsaturated fatty acids, and the peaks in the choline region.
HRMAS allows discriminating between molecular profiles of benign and atypical meningiomas. The use of orthogonal signal correction allows to remove spectral component which variations are orthogonal to the diagnosis variable (sample degradation, sample contamination, etc). Metabolic discrimination between benign and atypical meningioma include the levels of phospholipids, glutamate, and fatty acids, which have been associated in previous studies to proliferation, recurrence, and partial necrosis, respectively. These metabolites may be the basis for a non-invasive classification of benign and atypical meningioma. Addition of new cases would probably improve the outcome of the classification method.
The study of pediatric brain tumor metabolic profiles may help in the understanding of the disease. Valuable information about the metabolites enrolled in the disease can enhance our knowledge in this respect. High-resolution magic angle spectroscopy (HRMAS) is nowadays being broadly applied in metabolomics of intact tissues and cells, and has been successfully applied to the classification of different types of cancers. Subtle biochemical changes can be detected in intact tissues by HRMAS revealing the status of tumor microheterogeneity, unveiling tumor metabolic alterations before they are morphologically detectable. Those changes can also be correlated to histopathological features and diagnosis. In this work, we present metabolic profiles of pediatric brain tumors by HRMAS and possible correlations with the different tumor types and grades.
HRMAS spectra were measured at 4ºC on 22 samples of tumor tissue (17–33 mg) from pediatric patients (1–15 years old). The study included 8 medulloblastomas, 6 pilocytic astrocytomas, 3 low-grade astrocytomas, and 5 ependymomas. Diagnosis was confirmed histopathologically in all samples using standard protocols. Besides the comparisons between tumor groups, correlation with tumor grade was explored.
HRMAS spectra showed increased resolution and SNR compared to in-vivo MR spectroscopy (MRS). Choline compounds, which are highly overlapped in in-vivo spectra, were resolved by ex-vivo HRMAS. All aminoacids were identified, including relevant glutamine and glutamate. Good agreement between metabolites detected ex-vivo and in-vivo was observed. Some metabolite resonances showed significant differences between tumor groups.
Agreement between in-vivo and ex-vivo MRS suggests that HRMAS can improve resolution and provide a link between in-vivo spectroscopy and neuropathologycal analysis. Particular attention should be placed in Choline compounds, Glutamate, Glutamine and GABA which are often unresolved in common clinical scanners. Some metabolites not previously detected by ‘in vivo’ spectroscopy showed also some potential as biomarkers of tumor type and grade but their metabolic role in the disease is not well understood yet. Overall, we achieved metabolic discrimination between pediatric tumors according to the PCA. Our work suggests that HRMAS spectroscopy is a helpful technique to find metabolic patterns in pediatric brain tumors, to discriminate among tumor groups, and to correlate them with pathological pathways. The potential of the method as support to the histopathological analysis is also suggested.
Brain abscesses are rare and often unrecognized. Gold standard in diagnosis is MR-imaging but the differentiation between gliomas and cerebral metastasis of other tumors is still a not well solved problem. Diffusion-weighted imaging in combination with the determination of apparent diffusion coefficient seems to be a useful help. We subsequently present our experiences with this type of imaging.
We examined a group of 83 patients with 7 brain abscesses, 33 gliomas, and 43 intracerebral metastases.
In the abscess group was one false-positive and one false-negative result. In the other groups all results were right positive. One patient with a cavernoma showed a false-positive result with a hyperintensity on DWI and a reduced apparent diffusion coefficient. The false-negative result was seen in a patient with a frontal sinus fracture and a posttraumatic cerebritis.
Diffusion weighted MR-imaging has a great value for the identification of brain abscesses. Problematic is the registration of small, fast regressive lesions and early states of cerebritis.
A 60 year old patient was diagnosed and treated for glioblastoma multiforme in 2005. In August 2005 the only presenting symptoms were migraine like visual complaints. In September 2005 occipital debulking was performed and followed by radiotherapy combined with chemotherapy and chemotherapy with temozolamide according to the Stupp protocol. She remained stable until May 2007. Control MRI of the brain showed a contrast enhancing lesion at the site of the previous tumor. At first there were no physical complaints, but because of progressive hemianopsia and partial visual seizures, treatment with temozolamide was started based on the diagnosis “recurrent Malignant Glioblastoma.” After the first course of temozolomide, chemotherapy was discontinued due to side effects and based on second thoughts about the correct diagnosis. A differential diagnosis of radionecrosis, mimicking the recurrence of tumor, was considered. All clinical symptoms and the findings on MRI also could be attributed to a diagnosis of radionecrosis. It left us with a clinical dilemma, since recurrent glioblastoma warrants a totally different aproach regarding therapy, opposed to radionecrosis. Several ancillary investgations were made: MR spectroscopy, PET scanning, and expert opininion (based on symptoms and MRI findings) suggested a diagnosis of radionecrosis. Based on these findings a wait and see policy was initiated, with frequent control MRI’s. At the end of 2007 the size of the enhancing lesion had grown, and the patient was increasingly steroid dependent. Then a descision was made to perform another debulking operation with two objectives: To decompress and to obtain PA material. Craniotomy followed in February 2008, resulting in clinical improvement in providing the PA diagnosis of recurrent glioblastoma. In summary this case demonstrates the difficulties in distinguishing radionecrosis from the recurrence of glioblastoma multiforme. MR spectroscopy and PET scanning were not helpful, maybe even harmful in making the correct diagnosis. The patient will resume chemotherapy after recovering from her operation and steroid myopathy. The role of MR spectroscopy and PET scanning for distinguishing recurrent gliobastoma from radionecrosis will be discussed.
Prospective radiological grading of primary gliomas has inherent limitations but can provide relevant informations for the clinical management of the patients. Sensitivity of conventional MRI in defining the grade of gliomas ranges from 55.1 to 83% (Bulakbasi N. 2004, Zonari P. 2007, Law M 2003). To assess the accuracy of perfusion (PWI), diffusion (DWI), and proton-spectroscopy, we collected all patients with MRI suggesting glioma in absence of necrosis, relevant edema, and enhancing volume < 30% total tumor volume, from December 2006 to December 2007.
Soon before surgery, 39 patients (11 m,. 5 f,. age: 26–77) were investigated by 1.5 T MRI with a protocol including DWI, PWI, and proton spectroscopy. ROI analyses were performed on ADC and rCBV maps on the different T2 imaging portions of the tumor and on the contralateral normal region. After surgery histopathological analysis confirmed the diagnosis of low-grade glioma in 26 patients (4 fibrillary astrocytoma, 5 oligodendroglioma, 15 oligoastrocytomas, 1 pilocytic astrocytoma, 1 ganglioglioma). Other patients had a grade III (7 anaplastic oligoastrocytomas, 3 anaplastic astrocytomas) and grade IV (3 pts.) gliomas.
Correct identification of tumor grade was performed in 80% (31/39) on the basis of preoperative conventional MRI. Association of advanced MRI techniques allowed to increase this fraction to 87% of the cases (34/39). Fibrillary astrocytomas have high ADC value (1.3–1.5) and low rCBV values (0.4–1.1). Oligoastrocytomas grade II had ADC values ranging between 1.8–3 and rCBV values ranging between 0.3–1.4 with respect to contralateral normal values. Cho/Cr ratio was < 1 in all low-grade tumors except for oligodendroglioma. Oligodendroglioma is characterized by high value of rCBV (0.9–3.7, mean 1.9) and high Cho/Cr ratio. GBM had high rCBV (0.8–1.6, mean 1.5) and ADC (0.5–1.6, mean 1.3) These preliminary results confirm the diagnostic role of advanced MRI in association with standard imaging. The limitation is differentiation between grade II and III tumors. Threshold values and ROI positioning into the lesion and in peritumoral tissue is of critical relevance.
The objective of this study was to investigate flow-cytometric DNA values of intracranial tumors, and to establish DNA analysis as a potential prognostic parameter.
We prospectively evaluated twenty-one patients (9 males, 12 females, mean age 60.4 years) with brain tumors. All patients had been operated on. The diagnosis was glioblastoma multiforme (7), anaplastic astrocytoma (1), typical meningioma (8), anaplastic meningioma (2), metastasis (2), and hemangioblastoma (1). Flow cytometry analysis was performed in the fresh tumor tissue specimens. The presence of DNA ploidy and the percentage of cells in G0/G1, S, and G2/M phases of the cell cycle were measured for each patient.
DNA distribution profiles in the tumors revealed 8 cases of diploidy and 13 cases of aneuploidy. Malignant tumors had a higher incidence of aneuploidy and higher S phase. Specially for meningiomas there was a positive correlation between meningioma’s malignancy and level of aneuploidy (r = 0.854, p<0.001).
DNA ploidy and S phase as determined by flow cytometry are useful indicators of different biological behaviors in brain tumors.
In glioblastoma patients, tumor progression is related to angiogenesis, which is mediated in part by the vascular endothelial growth factor (VEGF-A) and its receptor VEGFR-1. In this study, VEGF-A and VEGFR-1 tumor levels, and their relationship with survival are analyzed.
Levels of VEGF-A and VEGFR-1 were quantified in tumor tissue extracts of 25 glioblastoma patients. Quantification was performed using ELISA techniques (Human VEGF, Biosource and Quantikine Human sVEGFR1, R&D Systems). Survival data were analyzed using the Kaplan-Meier method. Patients were classified in two groups: Group A (survival <4 months), and Group B (survival ≥4 months). The quartile 75 of survival was taken as a reference.
Median survival time was 7.83 months (CI 95%, 2.81–12.86). Patients who died from non-tumor-related causes were excluded from the analysis. In Group A (n = 6), a positive trend to higher levels of VEGFR-1 was found, compared with Group B (n = 13). (76.7±90.6 vs. 28.4±14.0 pg/mg protein; p = 0.07). With regard to VEGF-A levels, no differences between Group A and Group B were observed (3.9 ±3.1 vs. 5.4±6.9 ng/mg protein).
In glioblastoma patients, increased tissular levels of VEGFR-1 could be associated with reduced survival. Therefore, a greater number of cases should be studied in order to confirm these results.
Supported in part by a Grant SEOM 2004.
Some inflammatory markers, such as interleukin-6 (IL-6), are important mediators of tumor growth. Glioblastomas are tumors with significant vascular proliferation, inflammation, and necrosis; therefore these mechanisms must play a crucial role in the development of the disease. This study assesses pre-surgery circulating levels of inflammatory and angiogenic markers in glioblastoma patients.
The study included 25 patients with histologic diagnosis of glioblastoma. A group of 40 healthy volunteers with a similar distribution of age and sex was assessed as a control group. Samples of peripheral blood were collected from patients within one week before surgery. As inflammation markers, circulating levels of IL- 6 (ELISA), tumor necrosis factor alpha (TNFα)(ELISA), fibrinogen (Fg) (turbidity assay), sialic acid (SA) (colorimetric enzymatic method) and C-reactive protein (CRP) (immunological agglutination technique) were quantified. As angiogenesis markers, serum levels of vascular endothelial growth factor (VEGF), and thrombospondin-1 (TSP-1) were determined by ELISA. Results are expressed as mean ±SD. Significance of the differences between means was evaluated using the Student’s t-test for independent data.
We found a statistically significant increase in levels of all inflammatory markers in glioblastoma patients as compared with healthy controls (Fg : 333±178 vs. 238±44 mg/dl, p<0.001; SA: 77±25 vs. 53±9 mg/dl, p<0.0001; CRP: 16.6 ±27.8 vs. 1.6±1.7 mg/l, p<0.001; IL-6: 4.4±9.9 vs. 1.0±0.9 pg/ml, p<0.05).
Levels of VEGF were significantly increased (VEGF: 226±218 vs. 146±101 pg/ml, p<0.05). However, TSP-1 levels in glioblastoma patients were not different with respect to the control group. The Pearson test was used for correlation analysis between inflammatory and angiogenic markers. A significant correlation between VEGF and Fg (R = 0.65, p<0.01), and between VEGF and SA (R = 0.51, p<0.05) was observed.
Glioblastoma patients present significantly elevated levels of circulating inflammatory and proangiogenic markers.
Supported in part by a Grant SEOM 2004.
Astrocytomas represent the most frequent primary tumors of the central nervous system. There are many reports on the clinical use of MIB-1 labeling index (LI) which linked to proliferative activity in astrocytomas, and its significance varies from one study to another. There are no molecules known which are directly linked to the MIB-1 LI in astrocytomas. The aim of this study was to evaluate the clinical value of MIB-1 LI in our human astrocytic tumor cases of different histopathological grades (WHO) and explore the molecules that could influence the MIB-1 LI.
An immunohistochemical study of the MIB-1 proteins using the avidin-biotin-peroxidase method was performed and MIB-1 LI was determined in 70 astrocytomas (18 grade II, 10 grade III, and 42 grade IV). Clinical variables considered included gender, age, extent of surgical resection, tumor location and survival. In the same cases, EGFR, EGFRvIII, PTEN, PDGFR, and S1P1 (Sphingosine-1-phosphate 1 receptor) were detected and quantified by QRT-PCR, Western blotting, and immunohistochemistry.
Our data showed increasing values of MIB-1 LI with increasing grade of malignancy in astrocytomas (2.2% grade II, 5.4% grade III, and 26.7% grade IV). Kaplan-Meier survival curves for 40 patients with glioblastoma showed that high MIB-1 LI correlates with short survival (p<0.005). On the other hand, univariate analysis did not show any correlation between survival and gender, extension of surgical resection, tumor location, or patient’s age at surgery. Among molecules tested, only low expression levels of S1P1 were significantly correlated with high MIB-1 LI in glioblastomas (p<0.05).
This study establishes MIB-1 LI as an important predictor for the grade of astrocytic tumors and prognostic factor in human glioblastomas. S1P1 may play an important role in proliferative activity in glioblastomas.
YKL-40 is a glycoprotein belonging to the chitinase family secreted by chondrocytes, synovial cells, neutrophils, and osteosarcoma cell lines. YKL-40 acts as growth factors for vascular endothelial cells and fibroblasts, and as adhesion and migration factor for vascular smooth muscle and endothelial cells. It also promotes degradation of ECM and angiogenesis and has a potential role in matrix remodeling like other molecules such as MMPs, laminin, collagen, and plasminogen. The function of YKL-40 protein in human glioma biology is unknown. It seems to be involved in tumor invasion and angiogenesis; it is expressed in infiltrating macrophages and in proliferating endothelial cells, and its expression shows correlation with survival time in patients with advanced cancer.
Recent studies identified different expression of YKL-40 in the cytoplasm of tumor cells, blood vessels, extracellular matrix, and in reactive astrocytes in high-grade gliomas, whereas it was absent in low-grade gliomas and normal brain. On the hypothesis that this molecule is involved in the biology of a subset of glioblastoma and is a predictive marker of aggressive behavior, we performed an immunohistochemical and Western blot investigation on the expression and role of YKL-40 in 70 patients with glioma, to evaluate relationship with clinical outcome (time to tumor progression [TTP] and overall survival). The patients were 50 glioblastomas, 5 oligoastrocytomas, 5 anaplastic oligoastrocytomas, 3 astrocytomas, 3 anaplastic astrocytomas, 2 oligodendrogliomas, and 2 anaplastic oligodendrogliomas. We also investigated the co-expression of YKL-40 with EGFR, PTEN, p53, and gene metylation status of MGMT. Our results showed variable immunoreactivity for YKL-40 in glioblastoma tumor cells, and all low-grade gliomas were negatives. In some cases, positivity in proliferating endothelial cells and in reactive astrocytes was also detected. Co-expression with EGFR was observed in 25% and with MGMT in about 50% of glioblastoma. Western blot analysis was performed in the aim to evaluate the specifity of YKL-40 in 15 patients; YKL-40 showed a clear band at 40 kD with variable intensity. Clinical data now available for glioblastoma patients, suggest a correlation of YKL-40 expression with shorter time to tumor progression (median TTP 5.2 months), while tissues weakly positive or negative showed a longer median TTP (8.2 months). These findings suggest that YKL-40 is associated with higher grade astrocytic tumors and to shorter time to tumor progression in glioblastomas.
In gliomatosis cerebri (GC), defined as a diffuse neoplastic glial cell infiltration of the brain, up-front chemotherapy is often proposed as an alternative to the radiotherapy. GC invades in various proportion both white matter (WM) and grey matter (GM), as reflected by the GM/WM ratio. The GM/WM ratio was estimated in 70 patients with GC (42 men and 28 women; median age 47 years) treated with up-front chemotherapy (13 PCV, 57 temozolomide). Median GM/WM was 30%. Compared to group B (33 patients with GM/WM>30%), the 37 patients from group A (defined as GM/WM≤30%) had better performance status (p = 0.04), higher response rate to chemotherapy (27/37 vs. only 9/33; p = 0.0005), longer progression free survival (19.4 vs. 11.7 months, p = 0.02), and longer overall survival (56.1 vs. 26.4 months; p = 0.003). There was no significant correlation with histological subtype (oligodendroglial vs. astrocytic or mixed GC), grading, tumor localization, or laterality. The deletion of chromosomes 1p and 19q tended to be more frequent in group A (8/17 [47%] vs. 1/9 [11%] in group B [p = 0.07]). These data suggest that GM/WM ratio is a prognostic and predictive marker in GC.
The human telomerase reverse transcriptase (hTERT) gene is upregulated in a majority of malignant tumors. A variable tandem repeat, MNS16A, has been reported to be of functional significance for hTERT expression. Published data on the clinical relevance of MNS16A variants in brain tumors have been contradictory. The present population based study in Nordic countries and the UK evaluated brain tumor risk and survival in relation to MNS16A minisatellite variants in 648 glioma, 473 meningioma, and 1,359 age, sex, and geographically matched controls. By PCR based genotyping all study subjects with fragments of 240 or 271 bp were judged as having short (S) alleles and subjects with 299 or 331 bp fragments as having long (L) alleles. Risk of glioma or meningioma was estimated with logistic regression adjusting for age, sex, and country. Overall survival was analyzed using Kaplan-Meier estimates and equality of survival distributions using the log-rank test, and by Cox proportional hazard ratios. The MNS16A genotype was not associated with increased risk of occurrence of glioma or meningioma. For glioblastoma (GBM), overall survival differed significantly with MNS16A genotype, with median survivals of 15.3, 11.0, and 10.7 months for the LL, LS, and SS genotypes, respectively. The hazard ratio for death having the SS or LS genotype compared with the LL genotype was significantly increased, 2.05 95% CI (1.35–3.12). In contrast to a previous publication showing better survival for the MNS16A SS genotype, our data indicated a better outcome for the LL genotype. Collected data from our and previous studies indicate that the MNS16A genotype results are contradictory and that the genotype is likely to be of minor relevance for glioblastoma risk and outcome.
Over 2,000 Canadians are diagnosed yearly with glial neoplasms; treatment for many such tumors involves surgical resection then chemotherapy and radiotherapy. Recently, Hegi et al. demonstrated that patients with gliomas unable to repair lesions induced by the DNA alkylating agent temozolomide (TMZ), have a more favorable prognosis than those that maintain alkyl-repair capacity. Repair capacity was gauged by promoter methylation of the gene encoding MGMT, an alkyl-DNA repair protein: gene silencing by promoter methylation increases TMZ cytotoxicity. As such, expression of MGMT is a useful predictor of patient response to TMZ. However, not all responders in the Hegi study had methlyated tumors, so other markers of benefit from TMZ may exist. In the mouse, absence of the tumor suppressor p53 increases the sensitivity of astrocytes to DNA alklyating agents. Because altered p53 function is also common in human gliomas, we reasoned that p53 might be involved in determining glioma chemotherapeutic response: p53 regulates apoptosis and cell cycle arrest, so may affect sensitivity to DNA damage. We examined the effects of altering expression of p53 on cellular response to treatment with TMZ. RNAi for p53, Pifithrin, a p53 inhibitor, and the p53-targeting viral protein E6 were used to inhibit p53 function in glioma cell lines. Following exposure to TMZ cell viability was monitored and in all cases demonstrated that absence of p53 enhances the cytotoxic effects of TMZ. Based on this data, we predict that p53 expression status may be an independent predictor of response of glial cells to treatment with TMZ and, in addition to MGMT methylation status, might be a marker of benefit from TMZ therapy.
The integrin alpha v beta 3 (αvβ3) has been appointed a key player in the angiogenesis of malignant gliomas. However, there is still a controversy about the expression and distribution of αvβ3 integrin due to malignancy. The aim of our study was to assess the extent and pattern of αvβ3 integrin expression within primary glioblastomas and low-grade gliomas.
Immunohistochemical stainings from tumor samples (GBM: n = 12; LGG: n = 4) for detection of αvβ3 integrin (clone LM609) were performed simultaneously and automatically to reach a maximum of accuracy. Staining was quantified by a special imaging software, which was calibrated for this purpose to xenotransplanted human melanoma cells (M21) expressing αvβ3. Within a selected area, it measured (1) the mean intensity of the desired immunohistochemical staining and (2) the immunohistochemically positive fraction at a desired immunohistochemical staining intensity. To further analyze the distribution of the integrin subunits αv and β3, Western blot analysis from corresponding histological sections (GBM: n = 20; LGG: n = 5) was performed. The microvascular distribution to αvβ3 expression in these tumors was determined by costaining with endothelial cell specific markers (CD31).
The expression of αvβ3 was found to be significantly higher in glioblastomas than in low-grade gliomas, whereby focal strong reactivity was restricted to glioblastomas only. Different glioblastomas show a very heterogeneous expression of αvβ3 integrin which ranges from slightly to very strong integrin expression similar to high expressing non-CNS tumors. Subsequent analysis revealed that not only endothelial cells contribute to the overall amount of αvβ3 integrin in the tumors. In malignant gliomas, more than three quarters of the overall integrin expression (about 85 %) derived from glial tumor cells. Moreover, Western blot analysis demonstrated a significant difference in the expression of the β3-integrin subunit between glioblastomas and low-grade gliomas.
The presented data lead to new insights in the pattern of αvβ3 integrin in gliomas. Since antiangiogenic therapy with integrin αvβ3-antagonists has reached clinical trials for patients with malignant gliomas minute detection and quantification of αvβ3 integrin expression could be a prerequisite for selection of patients suitable for this kind of additional therapy.
A serious outcome of the patients with brain tumors is developed by peritumoral brain edema (PTBE). The development of PTBE is influenced by many factors including tight junction proteins, such as occludin. We studied the correlation with PTBE volume and survival time relevant to occludin expression in various pathology of brain tumors. Fresh-frozen specimens from sixty patients with brain tumor were obtained during surgery and confirmed pathologically. Occludin expression was investigated by Western blot analysis. PTBE volume was measured by using preoperative magnetic resonance image (MRI) and survival time in each patient was also estimated retrospectively. Occludin was detected in 41 (68.3%) brain tumors and the other 19 (31.7%) was not. Although it was revealed the lowest expression in high-grade gliomas, its expression was variable according to pathology of brain tumors (p > 0.05). The difference of PTBE volume between occludin-positive and negative brain tumors was significant (p = 0.002). The mean survival time was prolonged in occludin-positive tumors comparable to negative one (p = 0.001). This study suggests that occludin expression is well relevant to the PTBE development in brain tumors and may affect the prognosis for survival of patients.
Meningiomas account for about 24–30% of primary intracranial tumors. Most of these tumors are benign (WHO grades I and II) and can be cured by surgical resection. Recurrence, however, represent the major factor influencing patient’s outcome and occur between 10% and 25% of the cases undergoing complete resection of the tumor. The molecular mechanisms associated with recurrence are still unclear. In order to identify predictive candidate target genes and/or molecular markers of recurrence we performed gene expression profiles in recurrent and primary tumors. A total of 45 patient samples were analyzed by using oligonucleotide microarrays, including 4 samples of nonneoplastic postmortem meninges, 17 WHO grade I, 17 WHO grade II, and 7 recurrent (WHO grades I and II) tumors. Expression profiles of primary tumors was compared to that of recurrences. Statistical analyses were performed by using GEPAS (CIPF, Valencia, Spain) and GSEA (Broad Institute, Cambridge, Massachusetts). Using a discriminant approach analysis we identified genes with differential expression between recurrent and primary tumors (FDR<0.15). These genes were related to the WNT, MAPK, TGF-β signaling pathways, regulation of actin cytoskeleton, and focal adhesion. Our results revealed candidate markers associated with recurrence that were tested in tissue arrays. Additional studies of these genes are guaranteed to provide useful diagnostic and prognostic markers.
Several clinical reports have described a correlation among different kinds of cancer and increased serum levels of ferritin which is supposed to represent a non-specific tumor marker. With respect to brain tumors, this also holds true for neuroblastomas, whereas detailed information regarding the release of ferritin from glial tumors is still missing and the role of ferritin in brain tumor development is still poorly defined. Besides its main function as a cellular iron storage protein, certain secreted ferritin isoforms—i.e., heavy(H)-chain rich acidic isoferritins—have been described which reveal immunosuppressive properties and are able to stimulate apoptosis. Notably, the generation of reactive oxygen species appears of high relevance for ferritin mediated apoptosis which also involves the upregulation of p53 and stimulation of the Fas (CD95) pathway. Here we show that culture supernatants (CM) collected from certain cell lines derived from human glial tumors as well as rat glioma C6-CM are able to stimulate apoptosis. Since this effect is strongly inhibited by neutralizing antibodies raised against the ferritin H-chain, this effect is supposed to be mediated by H-chain rich isoferritins released from the cells into the culture medium. Intriguingly, not all examined tumor cell line culture supernatants as well as those obtained from primary, non-transformed mouse astrocytes show an apoptosis inducing potential. Notably, addition of the anti H-chain antibody to these CM uncovered a proliferation stimulating activity which has to be characterized in detail.
To detect molecular determinants associated with the development of glioma resistance to anti-angiogenesis and its therapeutic and clinical significance. Angiogenesis inhibitors inhibit the growth of experimental malignant gliomas in vivo and induce dynamic changes in tumor vasculature. Vascular regression is followed by vascular normalization afterwhich the tumor vasculature starts to actively grow.
By proteomics, we initially investigated the pattern of protein expression associated with the administration of one inhibitor, PF-4/DLR, that acts simultaneously on angiogenesis and invasion. The phase of normalization was associated with the increased expression of integrin linked kinase (ILK) and of its respective downstream molecules, expression of which was instead reduced during the phase of vascular regression. ILK expression was associated with an increased expression of molecules mediating signal transmission from integrin and growth factor receptors and implicated in cell growth/survival, cell cycle progression, tumor invasion, and angiogenesis. These data suggested ILK as one of the proteins associated with the development of resistance to anti-angiogenesis. We then silenced the activity of ILK to enhance the response of the tumor vasculature to anti-angiogenesis. By the use of siRNA, we started the silencing at the beginning of the phase of vessel normalization to prolong the period of normalization. PF-4/DLR was administered in combination with ILK siRNA, resulting in a significant increase in the duration of the phase of vessel normalization, and in a longer glioma growth control. This was associated with a decrease expression and activation of the downstream molecules associated with ILK activity. Finally, to determine the clinical relevance of these findings, we investigated the expression of ILK in samples of human gliomas of different histological grade and subtypes, from primary tumors or after treatment with chemo or radiotherapy. ILK expression correlated with tumor grade and subtypes. Its expression was higher in recurrent than in primary tumors, and particularly in those which were resistant to radio or chemotherapy. In high-grade gliomas, ILK expression correlated with time to recurrence.
ILK is a prognostic factor implicated in glioma resistance to therapy.
Sphingosine-1-phosphate (S1P) is a bioactive lipid that signals through a family of five G protein-coupled receptors, termed S1P1–5, and regulates cellular proliferation, migration, survival, and angiogenesis. We investigated the expression and the role of S1P receptors in human astrocytomas. Astrocytomas of various histologic grades expressed four types of S1P receptors, S1P1, S1P2, S1P3, and S1P5 by quantitative RT-PCR analysis. Expression of S1P1 receptor was significantly lower in glioblastomas than those in normal brains and low-grade astrocytomas. Immunoblot using specific antibody against S1P1 receptor showed that normal brain, low-grade astrocytomas and anaplastic astrocytomas expressed more S1P1 receptor protein than glioblastomas. Immunohistochemistry showed that S1P1 receptor was immunolocalized predominantly to the majority of astrocytes in the normal brain, but no staining was observed in neoplastic astrocytes in glioblastoma specimens. Patients with glioblastomas whose tumors showed high levels of S1P1 receptor expression had 76% survival rate at 1 year, whereas those with low levels of S1P1 had 37% survival rate at 1 year. Furthermore, we examined the mechanism of S1P1 receptor pertaining to glioma cell proliferation and migration by manipulating S1P1 gene in glioma cell lines. Forced expression of S1P1 receptor in low expressor cell lines (U87, U251) resulted in decreased cell growth concomitant with the activation of S1P1 receptor. Cells transfected with small interfering RNA of S1P1 receptor in high expressor cell lines (T98G, G112) promoted cell proliferation. In the migration assays, no significant change was observed by either overexpressing or knocking down S1P1 receptor in glioma cell lines. This is the first demonstration that S1P1 receptor signaling negatively controls cell proliferation in glioblastomas. Dysregulation of S1P1 receptor expression or function may underlie glioma proliferation but not migration.
Cancer stem cells (CSCs) are defined as a subpopulation of cells within a tumor that give rise and maintain the entire tumor. CSCs have been discovered in leukaemia, breast, colon, and prostate cancer and recently also in various brain tumors including glioblastoma multiforme (GBM). GBM, the most common and most aggressive primary brain tumor in adults, is a highly anaplastic tumor composed of more or less poorly differentiated cells with extremely variable histological appearances. We investigated cancer stem cell properties in two established rat glioma models, N29 and N32, induced by transplacental ethyl-nitroso-urea injection. This mode of tumor induction has previously been proposed to mimic the development of human glioblastomas from subventricular zone stem or progenitor cells. Both N29 and N32 cells readily form free-floating spheres with self-renewing capacity when cultured in specialized stem cell medium. N29 and N32 express the putative cancer stem cell marker CD133 and the stem/progenitor markers musashi and nestin in vitro concomitantly with the expression of the neural lineage markers glial fibrillary acidic protein, doublecortin, ßIII-tubulin, and CNPase. When subjected to differentiating conditions, neither N29 nor N32 changed their phenotype. Importantly, both models show a very high clonogenicity in vitro and a robust tumorigenicity in vivo. Unsorted single N29 and N32 tumor cells formed clones with a very high frequency in vitro, and intracerebral inoculation of as few as 10 N29 and N32 tumor cells, respectively, gave rise to a tumor. This study shows that both N29 and N32 glioma cells display properties of CSCs. In contrast to other reports of cancer stem cell-like cells in similar experimental gliomas, we demonstrate that sphere formation does not necessarily alter the CSC features of the tumor models. We use these two tumor models to develop and improve immunotherapy of glioblastoma in ongoing clinical trials. Thus, we conclude that our experimental therapies are directed against CSC-containing brain tumors which have a close resemblance to glioblastoma.
Adult human mesenchymal stem cells have been successfully used in animal experiments as a source for cell based strategies. Their differentiation potential, their expandability as well as their selective recruitment by malignant glioma cells makes them ideal candidates for cell-based treatment strategies. Aim of our study was to evaluate whether in vitro expansion of stem cells or interaction of stem cells with malignant cells bears an immanent risk of tumor formation.
Adult human mesenchymal stem cells (hMSCs) were obtained from bone marrow donations of healthy volunteers. Cells have been characterized by FACS analysis and differentiation assays. Kept in standard culture conditions, cell morphology, proliferation rates, and potential anchorage independent growth was analyzed. Cells were implanted into immune-deficient rats subcutaneously as well as intraveneously. Expression level of telomerase as an indicator for immortalization was detected using RT-PCR.
Isolated cells from healthy donors were characterized as mesenchymal stem cells by typical marker expression in FACS analysis and by differentiation into bone or fat producing lineages. In vitro, an initial low proliferation rate was followed by a replicative phase of senescence with no detectable proliferation. Cells of samples, which overcame the state of senescence, showed a marked increase in proliferation rate, a changed morphology and anchorage independent growth potential as an indicator for an acquired malignant potential. In animal experiments, these transformed cells showed local tumor formation. RT-PCR showed signs of hTERT-overexpression in malignant transformed cells.
Stem cell isolation from glioma patients for autologous reimplantation after transfection as a cellular vector bears a potential risk of stem cell tumor induction. Further studies are necessary in order to minimize this risk, inherent in expansion of stem cells in vitro.
While neurosphere- as well as xenograft tumor-initiating cells have been identified in gliomas, the resemblance between glioma cells and neural stem/progenitor cells as well as the prognostic value of stem/progenitor cell marker expression in glioma are poorly clarified.
Viable glioma cells were characterized for surface marker expression along the glial genesis hierarchy. Six low-grade and 17 high-grade glioma specimens were flow-cytometrically analyzed for markers characteristics of stem cells (CD133); glial progenitors (PDGFRα, A2B5, O4, and CD44); and late oligodendrocyte progenitors (O1). In parallel, the expression of glial fibrillary acidic protein (GFAP), synaptophysin, and neuron-specific enolase (NSE) was immunohistochemically analyzed in fixed tissue specimens. Irrespective of the grade and morphological diagnosis of gliomas, glioma cells concomitantly expressed PDGFRα, A2B5, O4, CD44, and GFAP. In contrast, O1 was weakly expressed in all low-grade and the majority of high-grade glioma specimens analyzed. Co-expression of neuronal markers was observed in all high-grade, but not low-grade, glioma specimens analyzed. The rare CD133 expressing cells in low-grade glioma specimens typically co-expressed vessel endothelial marker CD31. In contrast, distinct CD133 expression profiles in up to 90% of CD45-negative glioma cells were observed in 12 of the 17 high-grade glioma specimens and the majority of these CD133 expressing cells were CD31 negative. The CD133 expression correlates inversely with length of patient survival. Surprisingly, cytogenetic analysis showed that gliomas contained normal and abnormal cell karyotypes with hitherto indistinguishable phenotype.
This study constitutes an important step toward clarification of lineage commitment and differentiation blockage of glioma cells. Our data suggest that glioma cells may resemble expansion of glial lineage progenitor cells with compromised differentiation capacity downstream of A2B5 and O4 expression. The concurrent expression of neuronal markers demonstrates that high-grade glioma cells are endowed with multi-lineage differentiation potential in vivo. Importantly, enhanced CD133 expression marks a poor prognosis in gliomas.
Mesenchymal stem cells (MSC) have received much attention as therapeutic vehicle because they have migratory capacity toward the malignant glioma cells. In the present study, we evaluated new therapeutic strategy by using combination of intratumoral inoculation of MSC expressing suicide gene (cytosine deaminase) and 5-FC administration on malignant glioma cells.
At first, the cell migration assay was performed using double chamber dishes. MSCs were placed in the upper chamber with 8-μm pores, and 9L cells were placed in the lower well. We could confirm that MSCs possessed significant migratory capacity for glioma cells in vitro. Next, to evaluate the antitumor effect of 5FC/MSC-CD system in vitro, 9L cells were cocultured with MSC-CDs in the medium containing 5FC. We could find that this system had antitumor effect on the glioma cells. In addition, to confirm bystander effect, which did not require cell-to-cell contact, chemosensitivity assay was performed using double chamber dish. MSC-CDs were placed in the upper chamber with 0.4-μm pores, and 9L cells were placed in the lower wells in the medium containing 5FC. We could investigate strong bystander effects of this system. Following these in vitro studies, we investigated the migratory capacity of MSCs in vivo. 9L cells were implanted into the right basal ganglia, and then MSCs were injected directly into the implanted glioma. We comfirmed that MSCs could migrate and proliferate in the established tumor by immunohistochemical analysis. To evaluate antitumor effect of 5FC/MSC-CD system in vivo, MSC-CDs were inoculated into the center of already implanted 9L tumor through the same injection tract and treated by systemic administration of 5FC. The survival term of tumor implanted rats in the treatment group was significantly longer than that in the control group.
We demonstrated the migratory capacity of MSCs and the antitumor effect of 5FC/MSC-CD in vitro and in vivo. Molecular therapy employing MSCs as a targeting vehicle would be promising as a new therapeutic approach for highly invasive malignant glioma.
Glioblastoma multiforme (GBM) is a brain tumor with a very poor prognosis due to inevitable recurrence. During the last few years, a contingent of cells within the tumor, so-called stem-like tumor cells (STC), has been characterized in GBM. These cells have similar properties to stem cells and can, with a limited number of cells injected in vivo in animals, recapitulate an entire initial tumor. STC are also resistant to current radio- and chemo-therapeutic treatments in vitro. Therefore, STC are considered to play a major role in the recurrence observed in GBM patients. These cells may be appealing targets for new therapeutic approaches such as immunotherapy. In order to expand STC, tumor samples from GBM patients are dissociated mechanically and cultured in a serum-free medium previously described (Reynolds & Weiss, 1992) in presence of EGF and bFGF. After emergence of the first neurospheres (between 4 to 21 days), cells are dissociated with trypsin. Neurospheres cultures are then passaged every 21 days after trypsin dissociation. Up to now, we have tried to expand STC from 21 GBM samples and 4 neurosphere cultures have been amplified up to 5, 7, 7, and 10 passages. Phenotypic analyses show that cells derived from all the neurosphere cultures strongly express nestin and A2B5 markers. Three out of four express CD133 but only one at high expression level. Concerning the immunological characterization, cells derived from all neurosphere cultures express HLA class I molecules, essential for specific recognition of tumor cells by cytotoxic T lymphocytes. In addition, a tumor antigen, EGFRvIII, previously described in GBM, is found in one neurosphere culture. In conclusion, these preliminary results show that STC are potentially interesting targets for immunotherapy.
It is important to expand human neural stem/progenitor cells (hNS/PCs) in vitro to obtain large amounts for clinical applications. However, the long-term maintenance of hNS/PCs in vitro may cause DNA damage due to repeated passages, which might result in malignant transformation. Furthermore, hNS/PCs and their immediate progeny are considered to carry a credible risk for malignant transformation, because many of the molecular determinants that regulate normal neurogenesis are also apparently involved in tumorigenesis. However, the in vivo tumorigenicity of hNS/PCs, which are maintained for more than 250 days in vitro (DIV), has never been described and is still uncertain. Therefore, we examined the in vitro properties of hNS/PCs cultured for approximately 250 DIV (hNS/PCs-250) and approximately 500 DIV (hNS/PCs-500), and transplanted them into the striatum of immunodeficient mice to evaluate their tumorigenicity by continuous monitoring with bioluminescence imaging (BLI) for approximately 6 months in combination with conventional histology. In vitro, the hNS/PCs-250 exhibited a higher growth rate and greater neurogenic potential than hNS/PCs-500, which showed greater gliogenic potential. In vivo, both hNS/PCs-250 and -500 differentiated into neurons and astrocytes 4 weeks after being transplanted into immunodeficient mice, and hNS/PCs-250 exhibited better survival than hNS/PCs-500 at this time point. We also monitored hNS/PCs-250 grafted mice for long-term with BLI, and found that the grafted cells had survived stably and differentiated properly into neurons and astrocytes, even 6 months after the surgery. Consistent with the result of BLI, we did not detect any evidence of tumorigenicity by grafted hNS/PCs contrast to the result of grafted U87 glioblastoma cell line during the 6-month observation period. Furthermore, neither PCNA-/Ki67-positive proliferating cells nor significant malignant invasive features were detected histologically, suggesting that hNS/PCs were not tumorigenic. These findings support the idea that hNS/PCs may represent a safe and appropriate cell source for regenerative therapies for neurological disorders.
Medulloblastomas (MBs) are the most common brain tumors in childhood. A deeper knowledge of MB biology, including the mechanisms of tumor initiation and the identification of the cells of origin, raises the possibility to improve the treatment of this malignancy by rendering therapeutic approaches more selective and reducing side-effects of therapy. In our study we focused on the characterization of cancer stem cells (CSCs) in murine MBs. CSCs have been identified in different brain tumors, including MB, as tumor founding cells, able to sustain growth and progression of the tumor and to give rise to relapse. In our model, mice carrying heterozygous patched mutations develop spontaneously MBs, which mirror the human pathology. We isolated CSC lines from 2 different murine MBs (MB CSCs), culturing the specimens under the standard conditions (i.e., in the presence of EGF and FGF2) employed to isolate normal neural stem cells (NSCs). These MB-derived lines display several features that qualify them as CSCs, such as extensive proliferation, self-renewal, multipotency, as well as aberrant differentiation. Moreover, MB CSCs are endowed with tumor initiation ability, which was assessed through orthotopic and subcutaneous injection in mice. MB is thought to arise from cerebellar neural progenitors that undergo malignant transformation; in particular, the desmoplastic variant seems to derive from the transformed precursors