Background: The rarity and difficulty of conducting large trials limit the evaluation of various treatment options for primary central nervous system lymphoma (PCNSL). In this study, we sought to describe the demographics, diagnoses, management, and outcomes of patients with PCNSL at a single institution.
Methods: This is a retrospective study of 153 patients with PCNSL between 1986 and 2010. Prognostic factors identified by univariate and multivariable survival analyses were used by recursive partitioning analysis to generate a prognostic model.
Results: The median age was 61 and Karnofsky performance status (KPS) was 70. The progression-free survival was 9.3 months; the overall survival was 27 months. The diagnosis of PCNSL was established mainly by stereotactic brain biopsy (80%), cerebrospinal fluid analysis (7.2%), and vitrectomy (2.6%). Methotrexate-based chemotherapy with or without consolidation whole-brain radiation therapy offered better response rate and survival in the initial treatment than whole-brain radiation therapy alone. However, this observation was not present in the subsequent salvage treatments. Multivariable Cox proportional hazards regression identified age and KPS as the only prognostic indicators. Recursive partitioning analysis categorized the patients into 3 groups. Patients with KPS >70 had a favorable outcome compared with patients with KPS ≤70. This held true especially for patients age 60 and younger.
Conclusions: The Cleveland Clinic experience with management of PCNSL demonstrated successful disease control with methotrexate-based regimens. The survival and prognostic indicators approximate those reported previously and provide independent validation for a simple yet powerful prognostic model using age and KPS to predict survival.