|Home | About | Journals | Submit | Contact Us | Français|
In patients with parenchymal brain masses of uncertain origin responsive to corticosteroids, primary CNS lymphoma (PCNSL) should be considered. PCNSL is a rare but aggressive brain tumor that is highly sensitive to high-dose methotrexate (HDMTX)–based chemotherapy. We report a series of six patients with brain masses without histologic confirmation suspicious for PCNSL based on clinical and radiomorphologic criteria after exclusion of some infectious conditions. All patients were treated with HDMTX. We observed two complete responses, two partial responses, and one stable disease. One patient had progressive disease and received rescue whole-brain irradiation. All patients were alive without disease progression 12–48 months after HDMTX start. No symptoms of late neurotoxicity have occurred so far. The response and survival data in this small series of patients are encouraging and suggest a benefit for patients with suspected PCNSL after initial treatment with HDMTX.
In patients with brain tumors, diagnosis is usually made by histologic evaluation. However, in some patients biopsy is precarious or not feasible due to size or location of the lesion, cerebral edema, concomitant disease, or lack of patient’s consent. Moreover, a conclusive histologic diagnosis cannot be achieved in up to 10% of biopsies (Bernays et al., 2002; Hall, 1998; Wild et al., 1990). A probable diagnosis must then be made on clinical, radiologic, and biochemical/serologic findings in this situation. Primary CNS lymphoma (PCNSL)2 should be considered in some of these patients, especially when response to corticosteroids is observed.
PCNSL comprises only 2%–5% of all primary brain tumors, but a threefold increase of the incidence has been reported since 1970 (Eby et al., 1988). PCNSL is a very aggressive tumor leading to irreversible deficits when treatment is delayed. Moreover, treatment of PCNSL differs from many other primary brain tumors, in that resection is not beneficial for survival. Survival in PCNSL is limited to 12–18 months with whole-brain irradiation (WBI) alone (DeAngelis, 1995; Nelson et al., 1992) and can be prolonged to more than 30 months with high-dose methotrexate (HDMTX)–based chemotherapy (Abrey et al., 1998; Glass et al., 1994). Patients treated with radiotherapy seem at higher risk for cognitive dysfunction than patients treated with HDMTX-based chemotherapy alone. Thus, a rapid and correct diagnosis is crucial in PCNSL.
The aim of this retrospective study was to analyze the outcome of six patients with histologically unconfirmed parenchymal brain masses suspect of PCNSL after treatment with HDMTX.
Six immunocompetent patients with brain lesions suspect of PCNSL without histologic or cytologic diagnosis confirmation were treated at our institution from 2002 to 2005.
Contrast-enhanced brain MRIs were obtained from all patients before steroid medication. Routine sequences were T2-weighted turbo spin echo sequence, fluid-attenuated inversion recovery sequence, and T1-weighted spin echo sequence with and without gadolinium contrast. The scans were evaluated for the number of lesions, location, contrast enhancement, T2 signal intensity, necrosis, edema, and proximity to the subarachnoid space, ventricles, calvarium, and meninges. Necrosis was defined as an area without enhancement inside a contrast-enhanced lesion. The extent of edema was determined on T2-weighted images and rated as moderate, intermediate, or extensive according to previously published criteria (Küker et al., 2005).
In one patient with hypothalamic lesion, 18F-fluoro-deoxyglucose (FDG) PET was also performed.
All patients had a thoracic and abdominal CT scan and underwent cytologic and histologic bone marrow examination to exclude systemic tumor manifestations.
Patients 1 and 4 had an open brain biopsy (repeated in patient 4), and patient 5 had a stereotactic procedure, and the tissue was evaluated both histologically and immunohistologically. None of these patients received steroids before biopsy. Tissue material was not sufficient for additional molecular pathology. Cerebrospinal fluid (CSF) for cytomorphologic and immunocytologic evaluation was obtained from all but one patient, who refused puncture. Protein was measured in the CSF of all patients; IgG and oligoclonal bands were measured in the CSF and serum of three patients. PCR of the third complementarity-determining region (CDR3) for detection of clonal immunoglobulin heavy-chain rearrangements was additionally performed in three patients (Gleissner et al., 2002). Herpes virus infection was excluded by consensus PCR from CSF in four patients, and enterovirus infection was excluded by PCR from CSF in two patients. Patient 1 was additionally tested for JC virus, mycobacteria, and cryptococci in CSF.
Toxoplasmosis, cytomegalovirus, Epstein-Barr virus, and human immunodeficiency virus infection were excluded by serologic examinations in all patients. Ophthalmologic examination was performed in all but one patient.
Response evaluation was performed according to criteria recently published for PCNSL by Abrey et al. (2005) after the third and sixth course and every three months in the first year, every four months in the second year, and every six months in the third year and the following years of follow-up. Overall survival was measured from start of treatment to death or last follow-up.
The group consisted of four women and two men with a median age of 52.5 (range, 19–75) years.
Characteristics of the patients at first presentation are summarized in Table 1. Clinical symptoms deteriorated rapidly in patients 1–3, 5, and 6 and remained stable in patient 4. Corticosteroids were given before chemotherapy to all patients, which resulted in clinical improvement in all but patient 4.
Biopsy revealed nonspecific reactive tissue in patient 1 and perivascular T-cell and B-cell lymphocytic infiltrates in patients 4 and 5, with no evidence of demyelination or immunohistologic proof of lymphoma. Histopathology of patient 4 is shown in Fig. 1. In three patients, biopsy was considered not feasible due to lesion location (patients 3 and 6) or critical condition requiring immediate corticosteroid treatment (patient 2). Cytologic examination of CSF was initially inconclusive in all patients. In patient 3, CDR3 PCR revealed monoclonal cells after treatment had already been initiated.
Five patients had only one lesion on MRI, and one patient had two lesions. Location adjacent to the CSF space was found in all patients. Contrast enhancement was rated strong in four patients and moderate in two patients. Five patients had moderate edema on T2-weighted MRI, and one patient had intermediate edema. PET performed in patient 4 showed strong FDG uptake. Individual findings on MRI are shown in Fig. 2A and andBB.
All patients were discussed interdisciplinarily. After thorough information about the lack of a definite diagnosis and discussion of alternative diagnoses and treatment options, all patients consented and were then treated with HDMTX, 4 g/m2, with dexamethasone in the first course. In patient 1, WBI was started before chemotherapy after a tentative diagnosis of glioma and was stopped after application of 14 Gy. Patient 3 additionally received ifosfamide starting with the fourth course after partial remission had been achieved with three courses of HDMTX.
After chemotherapy, two patients achieved complete remission, two partial remission, and one patient had progressive disease before he received rescue WBI. A complete loss of FDG uptake on PET was found in patient 4, but tumor size only slightly decreased on MRI. All patients were alive 12–54 months after HDMTX start without further treatment, with no evidence of active tumor and without clinical signs of late neurotoxicity (see Table 1).
Patients with symptomatic parenchymal brain masses in whom access to tissue biopsy is not feasible or histopathology is not conclusive have very limited therapeutic options, usually reduced to steroid medication and brain radiation when a malignant tumor is suspected. However, clinical and radiomorphologic features may be indicative of PCNSL in some cases. Age greater than 50, short symptom duration (median, 2.8 months) with rapid deterioration, and rapid but transient improvement with corticosteroids are typical and were essential for considering PCNSL the most probable diagnosis in our patients. Patients 1 and 4 were exceptional in that their young age was very atypical for PCNSL. A characteristic MRI morphology with location adjacent to CSF space, strong (or at least moderate) and homogeneous contrast enhancement, moderate edema, and absence of necrosis was present in all our patients (Bühring et al., 2001; Küker et al., 2005; Roman-Goldstein et al., 1992; Schwaighofer et al., 1989). These findings are, however, not specific for PCNSL and can also be observed in other brain lesions.
On the other hand, PCNSL can sometimes present with atypical morphology, which can be attributed to low-grade histology in some cases (Jahnke et al., 2005). Analysis of the CSF did not immediately provide additional information with regard to the presence of lymphoma, but it was of utmost importance to exclude an infection.
A variety of differential diagnoses cannot be ruled out in our patients with absolute certainty. CNS sarcoidosis has especially been considered in patient 4 due to hypothalamic lesion location (Bullmann et al., 2000). However, biopsy did not show characteristic noncaseating granulomas in any patient. Moreover, isolated CNS sarcoidosis is very rare, and systemic manifestations were excluded by CT in all patients and additionally FDG PET in patient 4. Tumefactive multiple sclerosis (MS) with its variant Marburg-type MS typically shows a heterogeneous and annular contrast enhancement that differs from the radiologic appearance of lesions in our patients (Capello and Mancardi, 2004; Dagher and Smirniotopoulos, 1996; Kurihara et al., 1996; Mendez and Pogacar, 1998; Silva et al., 1999; Tan et al., 2004). Additionally, no signs of demyelination were detected in biopsies of our patients. Acute disseminated encephalomyelitis, a monophasic inflammatory demyelinating disease with long-term response to corticosteroids, usually affects children and young adults, in contrast to the older age of PCNSL patients (Marchioni et al., 2005; Menge et al., 2005; Schwarz et al., 2001). Acute disseminated encephalomyelitis should especially be considered in our two younger patients (patients 1 and 4). Autoimmune vasculitis can secondarily present with CNS manifestations (Siva, 2001); however, primary angiitis of the CNS, affecting mostly young adults (Volcy et al., 2004), is extremely rare (MacLaren et al., 2005; Rehman, 2000). Inflammatory pseudotumors, nonneoplastic lesions of unknown etiology, typically arise from meningeal structures in children and young adults and are moderately responsive to steroids (Mombaerts et al., 1996).
PCNSL seems a common diagnosis in patients with initially “unclear” brain masses responding to steroids. Of the four patients reported by Alderson et al. (1996), histology was nondiagnostic in all, with demyelination in two patients and nonspecific inflammation in one patient; however, corticosteroids were given in three patients before biopsy. All patients relapsed 7–11 months after initial response to corticosteroids with symptomatic new lesions, biopsy then revealing B-cell PCNSL. In a series of 12 patients analyzed by Bromberg et al. (2002), five were eventually diagnosed with PCNSL, with recurrence and death within 4–11 months after first presentation in four. Of the remaining patients, MS was diagnosed in two, thalamic infarct and renal carcinoma metastasis in one patient each, and no definite diagnosis could be made in three patients after follow-up ranging from three to eight years. A literature search of case reports performed by the authors revealed PCNSL in about 50% of 78 patients with similarly vanishing brain tumors. Patients with PCNSL tended to be older, and recurrence is most likely within 18 months, but clinical or radiologic features could not definitely distinguish PCNSL from non-PCNSL patients.
Newer and more sophisticated methods for the detection of PCNSL might contribute to diagnosis establishment in the future. The detection of monoclonal B-cells can be facilitated by IgH PCR as shown in one of our patients, but this method is not so easily available or fast enough and is positive only in a small number of PCNSL patients (Gleissner et al., 2002). PET imaging, used in one of our patients, has been shown to be beneficial in distinguishing between highly proliferative brain tumors like PCNSL, infectious lesions, and necrosis (Hoffman et al., 1993; Roelcke and Leenders, 1999). Advances in MRI technology such as diffusion-weighted protocols and MR spectroscopy imaging have recently been studied in PCNSL and may be routinely implemented (Harting et al., 2003; Hartmann et al., 2003; Raizer et al., 2005; Reiche et al., 2007). Diffusion-weighed MRI may be helpful in separating PCNSL from diseases associated with cytotoxic (ischemic) edema (Moritani et al., 2004). Perfusion-weighed MRI has recently been reported to produce characteristic findings in PCNSL as compared to high-grade gliomas (Hartmann et al., 2003). Limited data on MR spectroscopy suggest characteristic spectra in PCNSL (Harting et al., 2003).
When histology shows mixed lymphocytic infiltrates, clonality analysis using IgH PCR should always be attempted. However, feasibility is often limited by an insufficient amount of material for PCR analysis (Hoeve et al., 2000).
Neither response to corticosteroids nor response to HDMTX is evidentiary for PCNSL. Both agents are successfully used to treat many conditions considered differential diagnoses, although typically at lower doses. PCNSL seemed the most ominous differential diagnosis when not rapidly and adequately treated. Thus, the decision to treat with HDMTX appeared reasonable also in patients 1 and 4, in whom the diagnosis of PCNSL seemed somewhat less likely. HDMTX monotherapy is relatively safe when all precautions are observed. Therefore, it is mandatory that patients are treated only at experienced institutions with sufficient expertise.
Based on our findings, patients with clinical and radiomorphologic features suggestive of PCNSL should be treated initially with a chemotherapeutic regimen including HDMTX, with WBI being postponed until progress or relapse.
2Abbreviations used are as follows: CDR3, third complementarity-determining region; CSF, cerebrospinal fluid; FDG, 18F-fluorodeoxyglucose; HDMTX, high-dose methotrexate; MS, multiple sclerosis; PCNSL, primary CNS lymphoma; WBI, whole-brain irradiation.