High-dose methotrexate with or without whole brain radiotherapy for primary CNS lymphoma (G-PCNSL-SG-1): a phase 3, randomised, non-inferiority trial.

Background: High-dose methotrexate is the standard of care for patients with newly diagnosed primary CNS lymphoma. The role of whole brain radiotherapy is controversial because delayed neurotoxicity limits its acceptance as a standard of care. We aimed to investigate whether first-line chemotherapy based on high-dose methotrexate was non-inferior to the same chemotherapy regimen followed by whole brain radiotherapy for overall survival.

Methods: Immunocompetent patients with newly diagnosed primary CNS lymphoma were enrolled from 75 centres and treated between May, 2000, and May, 2009. Patients were allocated by computer-generated block randomisation to receive first-line chemotherapy based on high-dose methotrexate with or without subsequent whole brain radiotherapy, with stratification by age (<60 vs ≥60 years) and institution (Berlin vs Tübingen vs all other sites). The biostatistics centre assigned patients to treatment groups and informed local centres by fax; physicians and patients were not masked to treatment group after assignment. Patients enrolled between May, 2000, and August, 2006, received high-dose methotrexate (4 g/m(2)) on day 1 of six 14-day cycles; thereafter, patients received high-dose methotrexate plus ifosfamide (1·5 g/m(2)) on days 3-5 of six 14-day cycles. In those assigned to receive first-line chemotherapy followed by radiotherapy, whole brain radiotherapy was given to a total dose of 45 Gy, in 30 fractions of 1·5 Gy given daily on weekdays. Patients allocated to first-line chemotherapy without whole brain radiotherapy who had not achieved complete response were given high-dose cytarabine. The primary endpoint was overall survival, and analysis was per protocol. Our hypothesis was that the omission of whole brain radiotherapy does not compromise overall survival, with a non-inferiority margin of 0·9. This trial is registered with ClinicalTrials.gov, number NCT00153530.

Findings: 551 patients (median age 63 years, IQR 55-69) were enrolled and randomised, of whom 318 were treated per protocol. In the per-protocol population, median overall survival was 32·4 months (95% CI 25·8-39·0) in patients receiving whole brain radiotherapy (n=154), and 37·1 months (27·5-46·7) in those not receiving whole brain radiotherapy (n=164), hazard ratio 1·06 (95% CI 0·80-1·40; p=0·71). Thus our primary hypothesis was not proven. Median progression-free survival was 18·3 months (95% CI 11·6-25·0) in patients receiving whole brain radiotherapy, and 11·9 months (7·3-16·5; p=0·14) in those not receiving whole brain radiotherapy. Treatment-related neurotoxicity in patients with sustained complete response was more common in patients receiving whole brain radiotherapy (22/45, 49% by clinical assessment; 35/49, 71% by neuroradiology) than in those who did not (9/34, 26%; 16/35, 46%).

Interpretation: No significant difference in overall survival was recorded when whole brain radiotherapy was omitted from first-line chemotherapy in patients with newly diagnosed primary CNS lymphoma, but our primary hypothesis was not proven. The progression-free survival benefit afforded by whole brain radiotherapy has to be weighed against the increased risk of neurotoxicity in long-term survivors.