|Home | About | Journals | Submit | Contact Us | Français|
To report the final analysis of survival outcomes for children with newly diagnosed high-grade glioma (HGG) treated on the “Head Start” (HS) II and III protocols with chemotherapy and intent to avoid irradiation in children <6 years old.
Between 1997 and 2009, 32 eligible children were enrolled in HS II and III with anaplastic astrocytoma (AA, n = 19), glioblastoma multiforme (GBM, n = 11), or other HGG (n = 2). Central pathology review was completed on 78% of patients. Patients with predominantly brainstem tumors were excluded. Patients were to be treated with single induction chemotherapy regimen C, comprising four cycles of vincristine, carboplatin, and temozolomide. Following induction, patients underwent marrow-ablative chemotherapy and autologous hematopoietic cell rescue. Irradiation was used for patients with residual tumor after consolidation or >6 years old or at the time of tumor progression.
The 5-year event-free survival (EFS) and overall survival (OS) for all HGG patients were 25 ± 8% and 36 ± 9%, respectively. The EFS at 5 years for patients with AA and GBM were 24 ± 11% and 30 ± 16%, respectively (P = 0.65). The OS at 5 years for patients with AA and GBM was 34 ± 12% and 35 ± 16%, respectively (P = 0.83). Children <36 months old experienced improved 5-year EFS and OS of 44 ± 17% and 63 ± 17%, compared with children 36–71 months old (31 ± 13% and 38 ± 14%) and children>72 months old (0% and 13 ± 12%).
Irradiation-avoiding treatment strategies should be evaluated further in young children with HGG given similar survival rates to older children receiving standard irradiation-containing therapies.
Central nervous system (CNS) tumors are the second most common childhood malignancy and the leading cause of cancer-related death in children.1 High-grade gliomas (HGGs) are a heterogeneous group of tumors that include World Health Organization (WHO) grade III (anaplastic astrocytoma, AA) and IV (glioblastoma multiforme, GBM) neoplasms. HGGs account for approximately 10% of pediatric brain tumors and are the second most common malignant CNS tumor after medulloblastoma.2 Current therapies include surgical resection, radiotherapy, and chemotherapy, administered both concomitant with and following irradiation.2–6 Despite aggressive multi-modality therapy, historically the long-term survival has been poor, ranging from 10 to 30%.2 In addition, young children treated with cranial irradiation experience significant and irreversible neurocognitive sequelae, a phenomenon now well described in the literature.7–9 As a result, several clinical trials have attempted to delay or eliminate radiotherapy in the treatment of intracranial malignancies, particularly in very young children.10–12 Over the last two decades, marrow-ablative chemotherapy with autologous hematopoietic cell rescue has gained a role in the treatment of children with CNS tumors.13,14 Since 1991, the “Head Start” (HS) trials have focused on avoiding radiotherapy in young children and preserving neurocognitive function by incorporating high-dose, marrow-ablative chemotherapy and autologous hematopoietic progenitor cell rescue (AuHPCR) in the treatment of malignant CNS tumors. The success of this effort in the overall preservation of neuropsychological functioning and quality of life for patients treated on the HS protocols has been extensively discussed in other publications.7,8 Here, we present the analysis of survival outcomes for children with HGG treated with a uniform induction and consolidation chemotherapy strategy (regimen C) on the HS II and III protocols.
Children <10 years old with newly diagnosed HGG without prior chemotherapy or irradiation were enrolled in two sequential, prospective, nonrandomized international multiinstitution studies, HS II (1997–2003) and “Head Start” III (2003–2009). Patients with diffuse intrinsic pontine gliomas (DIPGs) were excluded. Gadolinium-enhanced magnetic resonance imaging (MRI) of the head and complete spine, as well as lumbar cerebrospinal fluid cytology, was required at the time of study enrollment. Chemotherapy was required to begin within 42 days of the most recent definitive surgical procedure. With respect to HGG, the primary aim of the HS II and III protocols was to determine the event-free survival (EFS) and overall survival (OS) from the time of study enrollment of children treated with regimen C followed by consolidation with marrow-ablative chemotherapy and AuHPCR after maximal surgical resection.
All patients were to undergo maximal possible surgical resection, consistent with preservation of neurologic function prior to enrollment in the studies. The extent of surgical resection was defined as follows: gross total resection (GTR) if no residual tumor was present, near total resection when less than GTR but more than 90% resection was achieved, subtotal resection when less than 90% but more than 50% resection was achieved, and partial resection if less than 50% of the tumor was resected. A small number of patients had biopsies only, where 10% or less of the tumor volume was resected. The degree of surgical resection was determined by central review of operative reports and of the preoperative and postoperative MRI reports; where these differed, the extent of resection was based on the imaging studies.
Patients were eligible to start therapy based on the institutional pathologic diagnosis. Every effort was made to obtain samples for central pathology review by leading neuropathologists in key centers in the United States and by the HS study committee neuropathologists at Children's Hospital Los Angeles (CHLA). Two of five patients from HS II underwent central pathology review, with no discordant cases. For HS III patients, 23 of 27 (85%) underwent central review, with two discordant cases. Both cases were originally classified as GBM, and later revised to AA after central review. Histology was classified according to the most recent WHO criteria. Of the 13 long-term survivors (>3 years), zero of three HS II patients had central review and nine of 10 HS III patients had central review. Two of those nine cases were discordant; both were read at the institution as GBM and later revised to AA by the central reviewer.
All patients were to be treated with regimen C, which consisted of four cycles of induction chemotherapy at 28 day intervals with vincristine (0.05 mg/kg intravenous [IV] on days 0, 7, and 14), carboplatin (560 mg/m2 or 20 mg/kg IV, whichever was less, on days 0 and 1), and temozolomide (6.5 mg/kg PO on days 0 through 4) followed by a consolidation phase with one cycle of marrow-ablative chemotherapy consisting of thiotepa (300 mg/m2 IV on days −5 through −3) and carboplatin (calculated using the Calvert formula for a desired area under the curve of 7 mg/ml/min, given IV, on days −8 through −6), with AuHPCR. Granulocyte colony-stimulating factor 5 μg/kg/day was to be administered subcutaneously commencing 24 hr after the last dose of chemotherapy in each induction cycle and continued until the postnadir white blood cell count was more than 10,000/μl. Chemotherapy was to be discontinued if disease progression or unacceptable toxicity occurred. Autologous hematopoietic progenitor cell leukopheresis was to be undertaken upon recovery from cycles 1 or 2.
Patients were to receive radiotherapy only if they were >6 years old at the time of diagnosis and/or had residual tumor at entry into consolidation chemotherapy. Radiotherapy was to be initiated following recovery from consolidation chemotherapy. Per protocol, children with localized residual tumor were to receive 54 Gy local field radiation therapy (XRT). XRT doses and volumes for rare patients with metastatic disease were left to treating investigators.
Data were collected using methodology similar to studies published from HS I and II.7,8 All subjects enrolled on the HS III study were expected to undergo a neuropsychological assessment prior to transplant and every 2 years thereafter. Unfortunately, given the small number of HGG subjects and limited follow-up data, the available neuropsychological data will be reported as descriptive.
Parents or legal guardians for each child signed an informed consent approved by the treating center's Institutional Review Board or equivalent committee. No patient was permitted to be enrolled from a participating center before the local Institutional Review Board (IRB) approval letter and consent form were further reviewed and approved by the Operating office (CHLA) IRB.
Primary endpoints for analysis were as follows: EFS—time from date of diagnosis to progression, relapse or death from any cause; OS—time from date of diagnosis to death; and irradiation-free survival—time from date of diagnosis to death or to the initiation of radiotherapy for any reason. These were estimated using product-limit estimates with Greenwood standard errors. Comparisons between groups were based on the log-rank test.
From May 2001 to December 2009, 32 children <10 years old newly diagnosed with and previously untreated for HGG were enrolled on the HS II and III protocols. Patient characteristics, tumor characteristics, and surgical outcomes are described in detail in Table 1.
Twenty of 32 children completed induction chemotherapy. Of the other 12 patients, nine had progressive disease (PD), one patient discontinued therapy per parent request, one patient died of sepsis during induction, and one patient had PD prior to starting induction and elected to start XRT right away. Table 2 details the radiographic responses to the induction chemotherapy.
Nineteen patients (59.4%) proceeded to consolidation chemotherapy. A single patient that completed induction had stable disease (SD) at the end of induction, but the family opted not to proceed to consolidation. Radiographic response was assessed on day +42 following consolidation.
Sixteen patients (50%) received XRT at some point during their management. Nine patients received XRT for recurrence after BMT, four patients had disease progression during induction therapy, two patients had residual disease after consolidation, and one patient discontinued chemotherapy on day 2 of cycle 1 due to significant progression prior to starting therapy, and proceeded directly radiotherapy. Figure 1 shows patient outcomes based on response to induction, AuHPCR, and XRT.
One patient enrolled on HS III died of toxicity due to sepsis during induction; this patient was steroid dependent and arrived at the ER in septic shock, rapidly developed Acute Respiratory Distress Syndrome (ARDS) and expired due to respiratory failure. Bronchoalveolar lavage demonstrated Candida glabrata and Candida krusei. There were no toxicity-related deaths during consolidation. Other nonfatal toxicity data are similar to those published for other patients enrolled on HS II and III on Regimens A, A2, or D, consisting primarily of cytopenias, mucositis, and anorexia, and are described in prior publications.10,11
Eleven patients experienced relapse after AuHPCR, and one patient who had SD after induction but declined AuHPCR had disease progression 2 years after stopping therapy. Of these 12 patients, eight relapsed within 12 months of AuHPCR (1, 1, 3, 3, 3, 5, 7, and 8 months after AuHPCR) and four patients relapsed at least 12 months after AuHPCR (17 months, 2, 4, and 5 years). One patient had metastatic recurrence in the paranasal sinus; the other 11 were local recurrences. Three families declined radiotherapy at recurrence; these patients died at 3, 5, and 9 months after recurrence. Nine patients received radiotherapy at the time of recurrence; four patients are still alive (3+, 42+, and 47+ months from date of recurrence) and five patients died at 11, 12, 15, 19, and 44 months after recurrence.
Among patients who were still alive, the median follow-up time was 5.2 years (range 1.1–8 years). For those patients that passed away, their time to death ranged from 22 days to 4.4 years after starting on study. The 5-year Kaplan–Meier analyses of EFS and OS for the 32 patients with HGG treated on HS II and III with regimen C were 25 ± 8% and 36 ± 9%, respectively (Fig. 2). Age-stratified EFS and OS are shown in Fig. 3. Five-year irradiation-free survival was 31 ± 10% (Supplementary Fig. S1). Overall, 11 of 32 patients survive. Nine of those 11 underwent BMT, six of whom survive without XRT. One additional surviving patient, who did not undergo BMT, had a second surgery and declined XRT, bringing the total number of patients who survive without XRT to seven.
Five patients underwent at least three serial neuropsychological evaluations that consisted of either an age-appropriate Wechsler Full Scale IQ score, Verbal IQ and Performance IQ score, and/or the parent-report Adaptive Behavior Assessment Scale, 2nd ed. Four of the five subjects were assessed prior to transplant (age range 5–49 months), with the most recent follow-up between 89 and 133 months, while the fifth subject was assessed at 79 months posttransplant. As the data available in the Supplementary Table S1 illustrates, the IQ and adaptive functioning scores between subjects varied between the high average to impaired ranges; however, it is notable that the predominance of scores within subjects were generally consistent between pretransplant and most recent follow-up, which is suggestive of similar stable findings from other HS subjects.7,8
Supplementary Table S2 shows the analysis of prognostic factors. There was no statistically significant difference in EFS or OS by pathology, age at diagnosis, or extent of resection. Patient sex was the only factor that had statistical significance, with males having better OS than females (65 ± 12% vs. 29 ± 12%, P = 0.02).
Patients with HGGs have historically experienced very poor prognosis and limited treatment options. Due to their relative rarity in children, there is a dearth of large randomized clinical trials. Many treatment options for children are therefore based on standard adult therapies with poor outcomes.19,20 Several publications have noted that pediatric and adult HGGs are molecularly different diseases, which may explain in part the lack of efficacy of adult-originated therapies in children.12,20–29
The first trial to evaluate the use of chemotherapy in the treatment of pediatric HGGs was the CCG-943 trial initiated in 1976, in which patients were randomized to receive regional field irradiation with or without adjuvant chemotherapy.30 Two important conclusions arose from this study: that extent of resection was the single most important prognostic factor, and that adjunctive chemotherapy improves outcomes in pediatric patients (EFS of 46% for children receiving radiation plus chemotherapy vs. 18% for those treated with radiotherapy alone).30,31 In contrast, adult trials using adjunct chemotherapy have had little success. In 2005, Stupp et al. demonstrated statistically significant improvement in survival for patient with GBM treated with adjunct temozolomide, reporting a 2-year OS of 27% for the irradiation plus temozolomide arm, compared to 10% for the irradiation only arm, or an absolute improvement in survival of 2.5 months.32
Since the results of CCG-943, a limited number of studies, including the seminal CCG-945 and “Baby POG” trials, have continued to demonstrate an important role for chemotherapy in the treatment of pediatric HGG.12,15–18,34 Table 3 summarizes the results of the most relevant of these trials. One commonality to all these studies is that the youngest children seem to have better outcomes than older children. Although no single explanation fully accounts for this observation, some studies have suggested that the gliomas of very young children are molecularly different from those of older children. For example, P53 gene mutations are associated with poor outcomes in HGGs.27 CCG-945 reported that among children <3 years old, 28% had p53 mutations, as compared to 70% of children 3–6 years old (P = 0.5).12 Several other molecular genetic targets have been or are being investigated for their roles in the differences between adult and pediatric HGG, including phosphatase and tensin homolog deletions, epidermal growth factor receptor gene amplification, alpha-type platelet-derived growth factor receptor, and Akt pathway activation.22–26,28 These biomarkers and others will need further research to establish their role as prognostic indicators and/or potential therapeutic targets in pediatric HGG.
The survival outcomes of the current study, although not statistically significant, show a trend that supports the theory of a biological difference between the HGG of younger versus older children. Our study included nine patients under the age of 3 years, with three of those patients being 6 months or younger, and seven of them <24 months of age at the time of diagnosis. Three patients died (two developed PD during induction therapy, one had relapse 3 months post-AuHPCR), one patient left the study and was lost to follow up after 2 years, and five patients survive 28+ to 72+ months after diagnosis. Four of these patients underwent AuHPCR and one family refused. One of these patients relapsed 6 months post-AuHPCR, but was successfully treated with radiotherapy. Overall, for our patient below 3 years of age, the 5-year EFS and OS were 44 ± 17% and 63 ± 17%, respectively, compared to children 36–71 months old (31 ± 13% and 38 ± 13%) and children >72 months old (0% and 13 ± 12%), P = 0.36 for EFS and P = 0.32 for OS. The small number of patients likely contributed to the lack of statistical significance, but the overall trend of this and other reported studies suggests that these youngest of patients have relatively more favorable survival rates with irradiation-avoiding strategies than their older counterparts with irradiation and chemotherapy.
The psychological and neurological burden of cranial radiation in children has been well documented.7–9 The HS protocols were designed to eliminate or delay irradiation in the youngest children with the goal of preserving neuropsychological function and quality of life. The positive impact of the HS protocols on both of these metrics has been published previously.7,8,10 Unfortunately, we have an insufficient number of subjects and follow-up testing data to report on the neurocognitive outcomes for these particular 32 patients. However, the limited outcomes for the HGG patients treated with the HS II and III protocols display a relative consistency of neurocognitive and adaptive functioning overtime.
In terms of survival, the results of HS protocols have been mixed depending on the type of CNS tumor. For medulloblastomas, the survival compares favorably to other treatment protocols,10 while treating patients with DIPGs showed no survival benefit.28,33 Few data have been previously available for patients with HGG, and this is largest analysis of patients with HGG treated on a HS protocol, showing response rate, EFS, and OS that are comparable to other irradiation-avoiding protocols.
There are several limitations to the current study. The lack of neurocognitive data does not allow us to make direct comments about the potential benefits of regimen C in HGG with regard to preservation of neurocognitive function. There was no collection or analysis of biological markers to further differentiate tumors on a molecular basis, such as in CCG-945. The relatively small number of patients makes it difficult to make generalizations about our data. Our analysis of prognostic factors such as diagnosis, age, and extent of resection failed to find any statistically significant differences where other studies have done so. The difference in OS between males and females may be an idiosyncrasy of our small patient sample and likely does not reflect a biologic difference between the sexes. Other studies have shown no difference in survival, and one study showed improved OS for females.15
The survival for young children treated with this irradiation-avoiding strategy is similar to that of older children with standard irradiation-containing therapies. This outcome may reflect biological differences between HGG in young children versus older children. These relatively favorable EFS and OS are likely achieved with relative preservation of quality of life and intellectual functioning as compared to regimens using upfront irradiation; while our surviving patient numbers are too few to document this within this specific patient population, it is quite reasonable to extrapolate from other well-published reports on long-term outcomes on larger numbers of young children treated with the “Head Start” strategy for medulloblastoma and other malignant CNS embryonal tumors. These data suggest that, while improved outcomes with BMT remain unproven, irradiation-avoiding treatment strategies such as HS deserve further development and evaluation in young children with HGG and would best be tested in a multicenter trial comparing conventional chemotherapy with marrow-ablative chemotherapy.
Grant sponsor: Michael Hoefflin Foundation; Soccer for Hope Foundation; Maddi's Closet; Alex's Lemonade Stand; Pediatric Cancer Research Foundation of Southern California; Children's Hospital of Los Angeles Pediatric Residency Program.
Supporting Information: Additional Supporting Information may be found online in the supporting information tab for this article.
Conflict of Interest: Dr. Espinoza is a clinical advisor to Birich Technologies, a small biotechnology firm with no products related to the diagnosis or treatment of CNS malignancies, or related research applications.