Background: Central nervous system (CNS) relapse occurs in 2-6% of DLBCL patients (pts) increasing to 10% or more in high-risk groups. Intrathecal (IT) or intravenous high-dose methotrexate (HD MTX) have limited if any prophylactic impact on CNS relapse. To address the role of systemic first-line therapy in pts tolerating intensified strategies (R-ACVBP, R-(Mega)CHOEP, R-CHO(E)P), we compared CNS relapses occurring in a large cohort of pts ≤60 years. Methods: We conducted a retrospective analysis including previously untreated pts with DLBCL by central review, age 18-60 years, from multicenter clinical trials conducted by LYSA and GLA/DSHNHL (Table). We assessed the risk of CNS relapse in matched cohorts based on the aaIPI. Results: A total of 2203 pts were included. Median age was 47 years (18-60). 455 pts were treated with R-ACVBP, 444 with R-(Mega)CHOEP, 1304 with R-CHOP. Distribution of CNS IPI was not significantly different comparing R-ACVBP to R-CHO(E)P groups within aaIPI categories (Table). PFS and OS were comparable according to treatment within aaIPI groups, also adjusted for prognostic factors. No CNS events occured during observation time of 3 years in pts with aaIPI 0. In pts with aaIPI 1, no CNS event occured in the R-ACVPB arm, the 3y-cumulative incidence of CNS relapse for pts treated with R-CHO(E)P group was 1.0% (95%CI 0.3-1.7). In pts with aaIPI 2,3 and intermediate/high CNS IPI, four (1.6%) treated with R-ACVBP experienced relapse in the CNS compared to 15 (3.9%) pts treated with R-(Mega)CHO(E)P (3y-cumulative incidence 1.6% (95%CI 0-3.2) vs. 4.0% (95%CI 2.0-6.0). Conclusions: CNS relapse was extremely rare in younger DLBCL pts with aaIPI 0 or 1; prophylactic measures are not warranted. In pts with aaIPI 2,3 (and intermediate/high CNS-IPI), only 4 (1.6%) CNS relapses were seen with the R-ACVBP while 15 (3.9%) relapses did occur after R-(Mega)CHO(E)P. This analysis underlines the important role of the systemic therapy in controling CNS relapse.