Primary central nervous system lymphoma (PCNSL) is increasing in incidence. Prospectively validated standardized treatment strategies for PCNSL are lacking. Moreover, therapeutic strategies have recently evolved without clear understanding of their impact on survival. This analysis aimed to study the impact of clinico-demographic characteristics and variability in patterns of care on survival outcomes in patients with PCNSL treated at five institutions in the Chicago Area.


Patients with PCNSL (excluding primary ocular) were retrospectively reviewed in an IRB approved, multicenter longitudinal registry. Data on disease and patient characteristics and outcomes were assessed in patients diagnosed between 2000-2012. Survival data was analyzed using Kaplan Meier method and Log-Rank. Multivariate analysis was performed using Cox proportional hazard regression.


We identified 99 patients with PCNSL with a median age of 63 years (24-88); 46% of these patients were less than 60 years of age. 54% of patients were male. 34 of 74 tested patients had elevated LDH (46%), 30 of 96 evaluated patients had deep structure involvement (31%), 56 of 67 tested patients had elevated CSF protein concentrations (84%), and 22 of 49 tested patients were EBER+ (45%). Of the 80 patients assessed for immunocompromise, 11 (14%) were HIV+ and 11 (14%) were on immunosuppressives leading up to their PCNSL diagnosis. Treatment data was acquired in 94 patients of which 90 patients were actively treated while 4 patients remained untreated; treatment data could not be obtained in 5 patients. In actively treated patients (n=90), the following therapeutic strategies were used: a high dose methotrexate (HD-MTX) containing regimen ± rituximab (R) + whole brain radiation (WBRT; n=28), a HD-MTX containing regimen ± R (n=57), WBRT alone (n=3) or R alone (n=2). 55 patients got R as part of their treatment. 27 patients received intrathecal chemotherapy (IT). Response data was available in 75 patients with an overall response in 60 patients (overall response rate of 80%; 55% with complete response), stable disease in 2, and progression in 13.

Median follow-up was 10.6 months (0.1-104.6). Age, sex, LDH, CSF protein concentrations, location of tumor, EBV status, HIV status and type of therapy were factors evaluated in univariate and multivariate analyses for impact on survival. Neither age, sex, high LDH, high CSF protein concentrations, deep structure involvement, nor EBV status impacted OS on Log-Rank analysis. HIV+ status was negatively associated with OS (31.4 vs 1.6 mo in HIV- vs HIV+ respectively, P <.05). OS was similar with the addition of WBRT to HD-MTX ± R systemic therapy vs systemic therapy alone. Similarly, the use of IT did not appear to impact survival. OS improved in patients receiving R (28.4 mo vs NR, P =.01). In multivariate Cox regression, HIV was a risk factor for inferior OS (HR 16.5; P <.002) whereas R use was associated with improved OS adjusting for HIV status as a covariate (HR 0.138, P <.005).


Therapeutic strategies for PCNSL vary in the Chicago Area although HD-MTX is commonly used. Therapy has evolved to incorporate rituximab and omit WBRT/IT without adverse effect on OS, a strategy that is undergoing evaluation in prospective trial.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.