Outcomes in splenic marginal zone lymphoma: analysis of 107 patients treated in British Columbia.

Splenic marginal zone lymphoma (SMZL) accounts for less than 2% of all non-Hodgkin lymphomas. We identified 107 cases diagnosed with SMZL between 1985 and 2012 from the British Columbia Cancer Agency Centre for Lymphoid Cancer and Lymphoma Pathology Databases. Patient characteristics were: median age 67 years (range 30-88), male 40%, stage IV 98%, splenomegaly 93%, bone marrow involvement 96%, peripheral blood involvement 87%. As initial treatment, 52 underwent splenectomy (10 with chemotherapy), 38 chemotherapy alone (21 chemoimmunotherapy containing rituximab, 1 rituximab alone), two antivirals for hepatitis C, and 15 were only observed. The 10-year overall survival for first-line splenectomy versus chemotherapy was 61% and 42%, respectively [Hazard Ratio (HR) 0·48, 95% confidence interval (CI) 0·26-0·88, P = 0·017]. The 10-year failure-free survival (FFS) after first-line splenectomy vs chemotherapy was 39% and 14%, respectively (HR 0·48, 95% CI 0·28-0·80, P = 0·004). Among the 38 patients who received first-line chemotherapy, FFS was similar between those receiving rituximab (n = 22) and those who did not (n = 16) (HR 0·64, 95% CI 0·31-1·34, P = 0·238). Fifteen patients transformed to aggressive lymphoma with median time to transformation of 3·5 years (range 6 months to 12 years) and the 10-year transformation rate was 18%. In conclusion, splenectomy remains a reasonable treatment for patients with SMZL.

Rituximab with high-dose methotrexate in primary central nervous system lymphoma.

The addition of rituximab (R) to chemotherapy improves outcomes in patients with systemic B-cell non-Hodgkin lymphomas, but the impact in patients with primary central nervous system lymphoma (PCNSL) receiving high-dose methotrexate (HDMTX) is unknown. Patients diagnosed with PCNSL at the British Columbia Cancer Agency (BCCA) between 2000 and 2013 were treated with ≥1 cycle of HDMTX 8 g/m(2) every 2 weeks, to best response or 10 cycles. After 2006, rituximab 375 mg/m(2) was given every 2 weeks with HDMTX for a total of 4 doses. 49 (66%) patients received HDMTX alone and 25 (34%) HDMTX+R, with a median of 5 (range 1-10) HDMTX cycles, and no difference between groups. The median follow-up was 5 years: 8.8 years (range 3.15-13.5 years) HDMTX and 1.9 years (range 0.5-7 years) HDMTX+R. The 5-year PFS was 17%, with no difference between groups (HR: 0.75, 95% CI: 0.41-1.35; P = 0.33). The 5-year OS was 38%, with no difference between the groups OS (HR: 0.73, 95% CI: 0.35-1.52; P = 0.39). In this retrospective study comparing two subgroups of patients treated in different eras, the addition of R to HDMTX did not appear to improve outcomes in PCNSL, possibly consistent with its known poor CNS penetration. It is possible that with a larger sample size, longer follow-up, or different rituximab dosing/schedule, the addition of rituximab may lead to a statistically significant improvement in outcomes. Prospective randomized trials currently in progress will more definitively estimate the impact of the addition of rituximab to HDMTX-based chemotherapy for PCNSL.

Long-term outcomes of R-CEOP show curative potential in patients with DLBCL and a contraindication to anthracyclines.

Doxorubicin plays an integral role in the treatment of patients with diffuse large B-cell lymphoma (DLBCL) but can be associated with significant toxicity. Treatment guidelines of British Columbia (BC) Cancer recommend the substitution of etoposide for doxorubicin in standard-dose R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone) (R-CEOP) for patients who have a contraindication to anthracyclines; however, it is unknown if this compromises treatment outcome. We identified all patients with newly diagnosed DLBCL who were treated in BC with curative intent with R-CEOP (n = 70) within the study period. Outcome in this population was compared with a 2:1 case-matched control group (n = 140) treated with R-CHOP and matched for age, clinical stage, and International Prognostic Index score. The 10-year time to progression and disease-specific survival were not significantly different for patients treated with R-CEOP compared with patients in the R-CHOP control group (53% vs 62% [P = .089] and 58% vs 67% [P = .251], respectively). The 10-year overall survival was lower in the R-CEOP group (30% vs 49%, P = .002), reflecting the impact of underlying comorbidities and frailty of this population. R-CEOP represents a useful treatment alternative for patients with DLBCL and an absolute contraindication to the use of anthracyclines, with curative potential.

Molecular features of a large cohort of primary central nervous system lymphoma using tissue microarray.

The objective of this study was to evaluate the distribution and prognostic impact of a broad range of molecular attributes in a large cohort of immunocompetent patients with primary central nervous system lymphoma (PCNSL) by using tissue microarray. Patients diagnosed with PCNSL were initially identified in the BC Cancer Lymphoid Cancer clinical and pathology databases. Tissue microarrays were constructed by using archival formalin-fixed paraffin-embedded diagnostic biopsy tissue. Immunohistochemistry and fluorescent in situ hybridization studies were performed. A total of 115 patients with PCNSL with diffuse large B-cell lymphoma (DLBCL) histology were identified. The majority of cases (≥75%) had a non-germinal center B-cell phenotype according to immunohistochemistry algorithms, but cell of origin did not affect progression-free or overall survival. MYC (40%), BCL2 (75%), and programmed death-ligand 1 (29%) protein expression were common, but their corresponding gene rearrangements were rare (≤1% each), suggesting that alternate mechanisms were driving expression. There were no dual rearrangements involving MYC and BCL2. Only 22% of cases had membranous expression of major histocompatibility complex class II, suggesting a mechanism for escape from immune surveillance. Epstein-Barr virus-encoded RNA was positive in 1 immunocompetent patient. BCL6 protein expression (77%) and BCL6 rearrangements (31%) were frequent; the latter was the only factor associated with a poor prognosis in the overall cohort and in the subgroup of 52 patients treated with high-dose methotrexate-based regimens. This large population-based study shows that prominent molecular features of PCNSL are unique and different from those of systemic DLBCL. These results may better inform drug development in PCNSL.

Primary CNS lymphoma of T-cell origin: a descriptive analysis from the international primary CNS lymphoma collaborative group.

PURPOSE: To describe the demographic and tumor related characteristics and outcomes for patients with primary T-cell CNS lymphoma (TPCNSL). PATIENTS AND METHODS: A retrospective series of patients with TPCNSL was compiled from twelve cancer centers in seven countries. RESULTS: We identified 45 patients with a median age of 60 years (range, 3 to 84 years). Twenty (44%) had Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 or 1. Twenty-six (58%) had involvement of a cerebral hemisphere and sixteen (36%) had lesions of deeper sites in the brain. Serum lactate dehydrogenase was elevated in 7 (32%) of 22 patients, and CSF protein was elevated in 19 of 24 patients (79%) with available data. The median disease-specific survival (DSS) was 25 months (95% CI, 11 to 38 months). The 2- and 5-year DSS were 51% (95% CI, 35% to 66%) and 17% (95% CI, 6% to 34%), respectively. Univariate and multivariate analyses were conducted for age ( 60 years), PS (0 or 1 v 2, 3, or 4), involvement of deep structures of the CNS (no v yes), and methotrexate (MTX) use in the primary treatment (yes v no). Only PS and MTX use were significantly associated with better outcome with hazard ratios of 0.2 (95% CI, 0.1 to 0.4) and 0.4 (95% CI, 0.2 to 0.8), respectively. CONCLUSION: This is the largest series ever assembled of TPCNSL. The presentation and outcome appear similar to that of B cell PCNSL. PS 0 or 1 and administration of MTX are associated with better survival.

Impact of time from diagnosis to initiation of curative-intent chemotherapy on clinical outcomes in patients with classical Hodgkin lymphoma.

The impact of treatment delays on outcomes in Hodgkin lymphoma (HL) is currently unknown. Time from definitive histologic diagnosis to first ABVD treatment (TDT) was calculated in 810 adults with HL: 365 (45%) TDT ≤4 weeks, 369 (46%) TDT 5-8 weeks, 76 (9%) TDT >8 weeks. The 5-year overall survival (OS) was 92% TDT ≤4 weeks, 92% TDT 5-8 weeks, and 83% TDT >8 weeks (p = 0.007). The 5-year disease-specific survival (DSS) was 93% TDT ≤4 weeks, 95% TDT 5-8 weeks, and 87% TDT >8 weeks (p = 0.094). The 5-year progression-free survival (PFS) was similar between groups (p = 0.139). In the multivariate analysis, TDT >8 weeks was not associated with worse OS, DSS, or PFS. Despite the univariate association between initiation of ABVD >8 weeks and worse OS, these data do not support such cut-off to improve outcomes. Nevertheless, clinicians should make every effort possible to initiate curative-intent chemotherapy as soon as a diagnosis of HL is established.

Brief chemotherapy and involved-region irradiation for limited-stage diffuse large-cell lymphoma: an 18-year experience from the British Columbia Cancer Agency.

PURPOSE: To evaluate clinical outcome of patients with limited-stage diffuse large-cell lymphoma (DLCL) treated with three cycles of chemotherapy followed by involved-region irradiation (IRRT). PATIENTS AND METHODS: Adults with limited-stage DLCL were treated with brief doxorubicin-containing chemotherapy regimens between 1980 and 1998. IRRT was administered 3 to 4 weeks after the third chemotherapy treatment in a dose equivalent to 30 Gy in 10 fractions. RESULTS: Three hundred and eight patients (median age, 64 years) were included, and 299 experienced complete remission. After a median follow-up of 86 months, 64 patients developed progressive disease, and 104 patients died (43 from lymphoma, three from toxicity, and 58 from other causes). Actuarial overall and progression-free survival (PFS) rates were, respectively, 80% and 81% at 5 years and 63% and 74% at 10 years. For subgroups identified using the Miller modification of the International Prognostic Index (IPI), the overall survival rates at 5 and 10 years were, respectively, 97% and 89% (no factors), 77% and 56% (one or two factors), and 58% and 48% (three or four factors), and the 5-year and 10-year PFS rates were, respectively, 94% and 89% (no factors), 79% and 73% (one or two factors), and 60% and 50% (three or four factors). Men with testicular presentation, had a definitely inferior outcome. CONCLUSION: Long-term outcome with three cycles of doxorubicin-based chemotherapy and IRRT confirms that this is a successful approach for the majority of patients with limited-stage DLCL. Subgroups with worse prognoses can be identified, and these patients should be offered alternative treatment approaches.

Unusual variants of primary central nervous system lymphoma.

Rare variants of primary central nervous system lymphoma (PCNSL) include unusual sites of presentation (eg, neurolymphomatosis and primary leptomeningeal lymphoma) and uncommon pathologic entities. Neurolymphomatosis involves peripheral nerves and nerve roots in addition to systemic and central nervous system (CNS) sites. Diagnosis requires a high index of suspicion, and treatment incorporates the principles of therapy for systemic and CNS lymphoma. Primary leptomeningeal lymphoma can present with symptoms of raised intracranial pressure or cranial or spinal polyradiculopathies. Diagnosis can be made by examining cerebrospinal fluid and incorporating immunophenotyping and molecular pathology techniques. Treatment options include irradiation and intrathecal or systemic chemotherapy. The features of PCNSL of T-cell origin and indolent B-cell PCNSL also are discussed.

Improved outcome of human immunodeficiency virus-associated plasmablastic lymphoma of the oral cavity in the era of highly active antiretroviral therapy: a report of two cases.

Plasmablastic lymphoma (PBL) is a recently described type of non-Hodgkin's lymphoma (NHL) that occurs in up to 3% of patients with HIV infection. Although the clinical-pathological features of several patients with HIV-associated plasmablastic lymphoma are documented, detailed description of clinical outcome is limited to isolated case reports. Generally, the response to lymphoma therapy is poor and survival is short. Response to highly active anti-retroviral therapy (HAART), however, has also been described. In this report, we describe the clinical course of two patients diagnosed with HIV-associated PBL in the era of HAART. One patient had a complete response to HAART, with a response-duration of 14 months, followed by relapse in the gastrointestinal tract several months after an anti-retroviral holiday. He is currently in complete remission (CR) eight months from diagnosis of relapse after receiving a full course of combination chemotherapy with modified CHOP, and 25 months from initial diagnosis. A second patient responded to brief chemotherapy in conjunction with HAART and is in clinical CR ten months from diagnosis. These cases illustrate that immunologic and virologic control with HAART may be beneficial for treating PBL and may possibly maintain continued CR. We advocate a high index of suspicion for primary PBL or its recurrence in patients with HIV infection, a history of low CD4 counts or high viral load, and oral or gastrointestinal symptoms.

Prognostic and predictive value of tumor vascular endothelial growth factor gene amplification in metastatic breast cancer treated with paclitaxel with and without bevacizumab; results from ECOG 2100 trial.

PURPOSE: Clinically validated biomarkers for anti-angiogenesis agents are not available. We have previously reported associations between candidate VEGFA single-nucleotide polymorphisms (SNP) and overall survival (OS) in E2100. The associations between tumor VEGFA amplification and outcome are evaluated here. EXPERIMENTAL DESIGN: E2100 was a phase III trial comparing paclitaxel with or without bevacizumab for patients with metastatic breast cancer. FISH to assess gene amplification status for VEGFA was conducted on paraffin-embedded tumors from 363 patients in E2100. Evaluation for association between amplification status and outcomes was conducted. RESULTS: Estrogen receptor (ER)+ or progesterone receptor (PR)+ tumors were less likely to have VEGFA amplification than ER/PR- tumors (P = 0.020). VEGFA amplification was associated with worse OS (20.2 vs. 25.3 months; P = 0.013) in univariate analysis with a trend for worse OS in multivariate analysis (P = 0.08). There was a significant interaction between VEGFA amplification, hormone receptor status, and study arm. Patients with VEGFA amplification and triple-negative breast cancers (TNBC) or HER2 amplification had inferior OS (P = 0.047); amplification did not affect OS for those who were ER+ or PR+ and HER2-. Those who received bevacizumab with VEGFA amplification had inferior progression-free survival (PFS; P = 0.010) and OS (P = 0.042); no association was seen in the control arm. Test for interaction between study arm and VEGFA amplification with OS was not significant. CONCLUSION: VEGFA amplification in univariate analysis was associated with poor outcomes; this was particularly prominent in HER2+ or TNBCs. Additional studies are necessary to confirm the trend for poor OS seen on multivariate analysis for patients treated with bevacizumab.

International Prognostic Score in advanced-stage Hodgkin's lymphoma: altered utility in the modern era.

PURPOSE: The International Prognostic Score (IPS) is the most widely used risk stratification index for Hodgkin's lymphoma (HL). It is based on patients treated before 1992 and predicts 5-year freedom from progression (FFP) and overall survival (OS) ranging from 42% to 84% and 56% to 89%, respectively. The IPS has not been validated in a recently treated population in which outcomes have improved compared with historic results. PATIENTS AND METHODS: By using the British Columbia Cancer Agency Lymphoid Cancer Database, we identified all patients age ≥ 16 years newly diagnosed with advanced-stage HL (stage III to IV, or stage I to II with "B" symptoms or bulky disease ≥ 10 cm) from 1980 to 2010, treated with curative intent with doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) or an ABVD-equivalent regimen with complete clinical information. RESULTS: In all, 740 patients were identified. Five-year FFP and OS were 78% and 90%, respectively. The IPS was prognostic for both FFP (P < .001) and OS (P < .001), with 5-year FFP ranging from 62% to 88% and 5-year OS ranging from 67% to 98%. Analysis limited to patients age 16 to 65 years (n = 686) demonstrated a narrower range of outcomes, with 5-year FFP ranging from 70% to 88% and 5-year OS ranging from 73% to 98%. CONCLUSION: The IPS remains prognostic for advanced-stage HL, but the range of outcomes has narrowed considerably. This improvement in outcome with ABVD should be acknowledged before consideration of alternate initial therapies and when comparing results from current trials with those of historic controls.

A phase II trial of a neoadjuvant platinum regimen for locally advanced breast cancer: pathologic response, long-term follow-up, and correlation with biomarkers.

PURPOSE: The purpose of this study is to determine the response, tolerability, and long-term outcome of a neoadjuvant platinum-containing regimen for locally advanced breast cancer (LABC) and to search for a correlation between pathologic complete response (pCR) and predefined biomarkers in this cohort. PATIENTS AND METHODS: Patients with LABC received 8 cycles of either sequence A or B. Sequence A was doxorubicin 60 mg/m(2) and paclitaxel 175 mg/m(2) (AT) every 3 weeks x 4 followed by cisplatin (C) 60 mg/m(2) and paclitaxel 90 mg/m(2) (CT) every 2 weeks x 4. Sequence B was CT x 4 (with paclitaxel dose escalation) followed by AT x 4. In addition to estrogen receptor (ER) and HER2, immunohistochemistry for MDR-1, MRP-1, topoisomerase IIalpha (topo IIalpha), and p53 was performed. RESULTS: A total of 88 patients were evaluable for response and toxicity. Median follow-up was 97 months. The overall pCR rate was 21.5%. For subgroups ER+/HER2-, HER2+ and double negative (ER-/HER2-) disease, the pCR rates were 5.9%, 23.3%, and 35%, respectively (P = .006). Five-year overall survival for the entire cohort was 71.1%. Five-year overall survival was 88.1% (95% CI, 77.1%-99.1%) for the ER+/HER2- group compared with 68.5% (95% CI, 51.3%-85.7%) and 49.5% (95% CI, 27.4%-71.6%) in the HER2+ and "double-negative" group, respectively (P = .0077). Overexpression of topo IIalpha was correlated with pCR (P < .001). There were no toxic deaths. CONCLUSION: A platinum-containing neoadjuvant regimen was well tolerated and achieved a pCR comparable to other recent studies of multiagent chemotherapy. Further studies tailored for specific breast cancer subtypes are required.

Population-based analysis of incidence and outcome of transformed non-Hodgkin's lymphoma.

PURPOSE: To assess the incidence and predictive factors for development of transformed lymphoma in a population-based series of patients with follicular lymphoma (FL). PATIENTS AND METHODS: The Lymphoid Cancer Database was used to identify patients with FL diagnosed and treated in the province of British Columbia, Canada. Transformed lymphoma was defined as the development of aggressive non-Hodgkin's lymphoma (NHL) in patients with FL. Factors present at the time of initial diagnosis of indolent NHL and at transformation were analyzed for their impact on risk of transformation and subsequent outcome. RESULTS: Between 1986 and 2001, 600 patients with newly diagnosed FL met the inclusion criteria. With a median follow-up of 109 months (range, 10 to 244), 170 (28%) developed transformation, 107 (63%) based on biopsy confirmation. The annual risk of transformation was 3% continuously through 15 years. A multivariate analysis of clinical factors at diagnosis identified advanced stage as the only predictor of future transformation. The median post-transformation survival was 1.7 years. The 5-year survival was superior for patients with limited extent transformation compared with those with advanced cases (66% v 19%, P < .0001). Patients with transformation based on clinical versus histological criteria had an identical median survival of 1.8 years (P = .2). CONCLUSION: The annual risk of transformation of FL is 3% continuing without plateau beyond 15 years. Advanced stage at diagnosis is predictive of future transformation. Clinically diagnosed transformation has an equal impact on outcome as biopsy proven transformation.

Brain parenchyma involvement as isolated central nervous system relapse of systemic non-Hodgkin lymphoma: an International Primary CNS Lymphoma Collaborative Group report.

Isolated central nervous system (CNS) relapse involving the brain parenchyma is a rare complication of systemic non-Hodgkin lymphoma. We retrospectively analyzed patient characteristics, management, and outcomes of this complication. After complete response to initial non-Hodgkin lymphoma treatment, patients with isolated CNS relapse with the brain parenchyma as initial relapse site were eligible. Patients with isolated CNS relapse involving only the cerebrospinal fluid were not eligible. Information on 113 patients was assembled from 13 investigators; 94 (83%) had diffuse large B-cell lymphoma. Median time to brain relapse was 1.8 years (range, 0.25-15.9 years). Brain relapse was identified by neuroimaging in all patients; in 54 (48%), diagnostic brain tumor specimen was obtained. Median overall survival from date of brain relapse was 1.6 years (95% confidence interval, 0.9-2.6 years); 26 (23%) have survived 3 years or more. Median time to progression was 1.0 year (95% confidence interval, 0.7-1.7 years). Age less than 60 years (P = .006) at relapse and methotrexate use (P = .008) as front-line treatment for brain relapse were significantly associated with longer survival in a multivariate model. Our results suggest systemic methotrexate is the optimal treatment for isolated CNS relapse involving the brain parenchyma. Long-term survival is possible in some patients.

Effect of letrozole versus placebo on bone mineral density in women with primary breast cancer completing 5 or more years of adjuvant tamoxifen: a companion study to NCIC CTG MA.17.

PURPOSE: Aromatase inhibition depletes estrogen levels and may be associated with accelerated bone resorption. The National Cancer Institute of Canada Clinical Trials Group (NCIC CTG) study MA.17B evaluated bone turnover markers and bone mineral density (BMD) in postmenopausal women randomly assigned to MA.17, a placebo-controlled trial of letrozole after standard adjuvant tamoxifen. PATIENTS AND METHODS: Eligible women had a baseline BMD T score of at least 2.0 in either the hip or L2-4 spine; all received calcium 500 mg and vitamin D 400 U daily. Percentage change in BMD (L2-L4 spine and hip) at 12 and 24 months, rate of osteoporosis, and change in markers of bone formation (serum bone alkaline phosphatase) and resorption (serum C-telopeptide and urine N-telopeptide) at 6, 12, and 24 months were compared. RESULTS: Two hundred twenty-six patients (122 letrozole, 104 placebo) were enrolled. Baseline characteristics were similar in the two groups, including BMD, median age of 60.7 years (81% < 70 years), and median follow-up of 1.6 years. At 24 months, patients receiving letrozole had a significant decrease in total hip BMD (-3.6% v -0.71%; P = .044) and lumbar spine BMD (-5.35% v -0.70%; P = .008). Letrozole increased urine N-telopeptide at 6, 12, and 24 months (P = .054, < .001, and .016, respectively). No patient went below the threshold for osteoporosis in total hip BMD, whereas at the L2-L4 (posteroanterior view), more women became osteoporotic by BMD while receiving letrozole (4.1% v 0%; P = .064). CONCLUSION: After 5 years of adjuvant tamoxifen, subsequent letrozole causes a modest increase in bone resorption and reduction in bone mineral density in the spine and hip compared to placebo. Further follow-up is necessary to evaluate the long-term clinical implications of this difference.

International study on low-grade primary central nervous system lymphoma.

OBJECTIVE: The aim of this study was to characterize the clinical presentation, course, and outcome of low-grade primary central nervous system lymphoma. METHODS: Cases were assessed in a retrospective series collected from 18 cancer centers in 5 countries. RESULTS: Forty patients (18 men, 22 women; median age, 60 years [range, 19-78]) were identified. Involvement of a cerebral hemisphere or deeper brain structures was seen in 37 patients, only leptomeningeal involvement in 2 patients, and spinal cord disease in 1 patient. Chemotherapy/radiotherapy was conducted in 15 patients, radiotherapy alone in 12, chemotherapy alone in 10, and tumor resection alone in 2, whereas 1 patient received no treatment. The median progression-free, disease-specific, and overall survival were 61.5 (range, 0-204), 130 (range, 1-204), and 79 (range, 1-204) months, respectively. Only age 60 years or older was associated with shorter progression-free (p = 0.009), disease-specific (p = 0.015), and overall survival (p = 0.001) in multivariate analysis. INTERPRETATION: Low-grade primary central nervous system lymphoma differs from the high-grade subtype in its pathological, clinical, and radiological features. It has a better long-term outcome than primary central nervous system lymphoma in general with age 60 years or older adversely affecting survival.

The treatment of primary central nervous system lymphoma in 122 immunocompetent patients: a population-based study of successively treated cohorts from the British Colombia Cancer Agency.

BACKGROUND: The objective of this study was to evaluate the clinical outcome of a population-based cohort of immunocompetent patients with primary central nervous system lymphoma (PCNSL) treated with 3 different strategies over 13 years. METHODS: One hundred twenty-two consecutive patients (median age, 66 years) with PCNSL were identified. Three treatment strategies were employed: 1) whole-brain irradiation with (from January, 1990, to June, 1991) or without (from April, 1995, to December, 1999) cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP)-type chemotherapy (n=50 patients); 2) combined-modality therapy, including 1 g/m2 methotrexate plus whole-brain irradiation (from July, 1991, to March, 1995; n=34 patients); and 3) 8 g/m2 methotrexate alone (from January, 2000, to March, 2003) with whole-brain irradiation reserved for those with progressive disease (n=38 patients). Treatment failure was defined as progressive disease, disease recurrence, death from toxicity or lymphoma, or toxicity that necessitated a change in primary treatment. RESULTS: The median failure-free survival was 7 months, and the median overall survival (OS) was 17 months. The median OS was similar in all 3 eras. In this population-based analysis, one-third of patients did not receive the treatment strategy of the era. Therefore, the data also were analyzed by treatment received. On multivariate analysis (including era of treatment), 3 factors-age > 60 years, lactate dehydrogenase > normal, and omission of methotrexate-were associated significantly with poorer OS (hazard ratio: 2.3, 2.2, and 2.3, respectively). CONCLUSIONS: Outcomes for a general population with PCNSL remained constant despite different treatment strategies over three eras. For the two-thirds of patients who could receive potentially curative treatment, age, lactate dehydrogenase level, and receipt of > or = 1 g/m2 methotrexate appeared to be important determinants of OS.