Clinical relevance of consolidation radiotherapy and other main therapeutic issues in primary central nervous system lymphomas treated with upfront high-dose methotrexate

Purpose: To evaluate the optimal dose of methotrexate (MTX) and the efficacy of other drugs, intrathecal chemotherapy (CHT), and radiotherapy (RT) in primary brain lymphomas. Methods and Materials: Two hundred eighty-eight immunocompetent patients with histologically documented, previously untreated primary brain lymphomas, receiving CHT containing high-dose MTX (greater than or equal to1 g/m(2)) with or without RT were selected from 19 prospective series. The impact on survival of the MTX dose (<3 g/m(2) vs. greater than or equal to3 g/m(2)), the main drugs, intrathecal CHT, and combination CHT (mono-CHT vs. poly-CHT) was assessed, according to the intention-to-treat principle. The role of post-CHT irradiation (immediate vs. delayed RT) was evaluated in 119 patients with a complete response to CHT. The whole brain and tumor bed dose (< 40 Gy vs. greater than or equal to 40 Gy) was assessed in 70 irradiated complete responders. Results: No difference in overall survival (OS) was detected between mono-CHT and combination CHT (p = 0.38). MTX greater than or equal to3 g/m(2) (P = 0.04), thiotepa (p = 0.03), and intrathecal CHT (p = 0.03) improved the OS, and nitrosoureas (p = 0.01) correlated with a worse survival. In multivariate analysis, limited to patients receiving MTX greater than or equal to3 g/m(2), only the addition of cytarabine improved the OS; nitrosoureas reduced MTX efficacy. Of the 119 complete responders, 70 received immediate RT. A RT dose of greater than or equal to 40 Gy to the whole brain or tumor bed did not improve OS. The 3-year OS was similar between the immediate and delayed RT groups. In multivariate analysis, RT delay had no negative impact on survival. Conclusions: MTX greater than or equal to3 g/m(2) seems to improve survival in primary brain lymphoma patients. The efficacy of additional drugs, except for cytarabine, remains unproved. Randomized trials are needed to confirm that RT withdrawal yields no detrimental effect in complete responders. (C) 2001 Elsevier Science Inc.