Context: Relapses involving the central nervous system (CNS) are uncommon among patients with diffuse large B-cell lymphoma (DLBCL) but carry very poor prognosis. Prophylaxis is heterogeneously prescribed around the time of frontline therapy, with no clear standard of care in terms of recipient selection or route of administration. Design: We retrospectively evaluated 1162 adult patients across 21 US academic centers with DLBCL or similar histologies who received single-route CNS prophylaxis as part of frontline chemoimmunotherapy between 2013-2019. Primary endpoint was CNS relapse rate. Secondary endpoints included PFS, OS, and treatment-related toxicities following frontline and salvage therapies. Patients: Prophylaxis was administered intrathecally (IT) in 894 (77%) and using systemic high-dose methotrexate (HD-MTX) in 236 (20%); 32 patients (3%) switched route due to toxicity and were assessed separately. By CNS-International Prognostic Index (IPI), 18% were considered low-risk, 51% moderate, and 30% high. Double-hit lymphoma (DHL) was confirmed in 243 of 866 evaluable patients (21%). Results: Sixty-four patients (5.7 %) had CNS relapse, after median 7.1 months from diagnosis, including 15 of 64 (23%) within the first 6 months. There was no significant difference in CNS relapse between IT and HD-MTX recipients (5.4 vs 6.8%, p=0.40), including after propensity score matching to account for differences between respective recipient groups. Weighting by CNS-IPI, expected versus observed CNS relapse rates were nearly identical (5.8 vs 5.7%). Testicular involvement was associated with high risk of CNS relapse (11.3%) despite most having lower CNS-IPI scores. DHL did not significantly predict for CNS relapse after single-route prophylaxis, including with adjustment for treatment regimen and other factors. Median OS following CNS relapse was 6.2 months (IQR 2.9-24.6) and was significantly shorter among patients achieving less than CR to first salvage therapy (HR 3.39, 95% CI 1.76-6.54, p=0.0003). Additional analyses of outcomes following CNS relapse will be presented. Conclusions: This large study of CNS prophylaxis recipients with DLBCL found no significant difference in CNS relapse rates between routes of administration. Relapse rates among high-risk subgroups remain elevated; treatments following CNS relapse were heterogeneous, with poor outcomes for the vast majority of patients. Reconsideration of SCNSL prophylaxis and treatment strategies is of critical need.
- central nervous system
- real-world evidence