Primary central nervous system large B-cell lymphoma (PCNS-LBL) has poorer outcomes than systemic diffuse large B-cell lymphoma, with unique challenges due to aggressive biology, limited CNS penetration of therapy, and high toxicity. The MATRix regimen is effective but requires advanced infrastructure, may not suit comorbid patients, and depends on thiotepa—often unavailable in resource-limited settings. Simplified regimens such as methotrexate, rituximab, and temozolomide (MTR) may be more accessible.

To compare a simplified MTR regimen with other treatments for PCNS-LBL.

We analyzed adults with newly diagnosed PCNS-LBL treated at two Brazilian centers. Patients were grouped as: (1) simplified MTR (methotrexate 3–3.5 g/m2 days 1, 10, 20; rituximab 375 mg/m2 days 1, 10, 20; temozolomide 150 mg/m2 days 1–7 per cycle) or (2) alternative regimens. In the simplified MTR group, patients achieving complete response (CR) after the first cycle (∼day 45) and eligible were referred for autologous hematopoietic stem cell transplantation (ASCT). Partial responders or ASCT-ineligible patients in this group received additional MTR cycles at physician discretion. Progression-free survival (PFS) and overall survival (OS) were estimated with Kaplan-Meier.

Twenty patients (median age 61, range 32–89) were included: 8 MTR and 12 controls. Half of the MTR group underwent ASCT after a median of 1 cycle; non- ASCT patients received a median of 3.5 cycles (range 1–5). No treatment-related deaths occurred with MTR; one MTX-related death occurred in controls. CR rates were 62.5% (5/8) in MTR and 58.3% (7/12) in controls (p = 1.0; OR 1.18, 95% CI 0.14–11.40). Overall response rates were 75.0% vs. 66.7% (p = 1.0; OR 1.47, 95% CI 0.15–21.51). Median PFS was 30.1 vs. 8.7 months, with 2-year PFS of 75.0% (95% CI 50.3–100%) vs. 50.0% (95% CI 28.4–88.0%) (log-rank p = 0.7). In 2 years, OS was 83.3% (95% CI: 58.3–100%) for MTR and 58.3% (95% CI: 36.2–94.1%) for other regimens (log-rank p = 0.4).

The simplified MTR regimen demonstrated high efficacy and low toxicity. Compared to the Alliance 50202 phase II trial, which reported CRR of 66% and 2-year PFS of 57% using a more complex MTR-based protocol, our regimen achieved a similar CRR with reduced treatment burden and toxicity. Notably, our approach used lower methotrexate doses (3–3.5 g/m2 vs. 8 g/m2), shorter treatment duration, and omitted intensive cytarabine/etoposide consolidation. Compared to MATRix, which showed CRR of 49% and 7-year PFS of 50% in the IELSG32 trial but carried high hematologic and infectious toxicity (up to 30%), our simplified MTR regimen offered higher CRR (62.5%) and no treatment-related deaths. Similarly, while R-MPV has shown CR and 2-year PFS rates of 60% and 77% (for those who achieved CR), respectively, patients were submitted to whole-brain radiotherapy for consolidation - a strategy linked to long-term neurotoxicity. The favorable toxicity profile observed in the simplified MTR group likely contributed to the lower mortality and the ability to proceed to ASCT, with encouraging sustained responses, supporting the role of early consolidation. In conclusion, simplified MTR is a feasible, effective, and well-tolerated option for PCNS-LBL, enabling early ASCT and showing promising survival, particularly in resource-limited settings or for patients unfit for intensive regimens.