Predicting CAR-T outcomes in R/R DLBCL: a multicenter real-world study of a 5-index model

The clinical management of relapsed or refractory diffuse large B‑cell lymphoma (R/R DLBCL) has been transformed by chimeric antigen receptor T‑cell therapy (CAR‑T), yet a significant challenge remains in predicting which patients will derive long‑term benefit.

This multicenter retrospective real‑world study aimed to validate a previously developed efficacy prediction model for CD19 CAR‑T in Chinese patients with R/R DLBCL. A total of 92 patients with DLBCL who received CD19 CAR‑T across four Chinese centers from August 1, 2021, to November 30, 2024 were included. The 5‑index prediction model (incorporating double‑expressor lymphoma status, TP53 alterations, ECOG performance status ≥2, bulky disease ≥5 cm, and prior therapy lines ≥4) was applied to predict treatment outcomes. The primary endpoints were overall response rate (ORR), complete response (CR) rate, progression‑free survival (PFS), and overall survival (OS).

The median follow‑up was 14.6 months. The C‑index for the 5‑index model was 0.767, indicating good predictive performance. The model effectively stratified patients into different risk groups, with significant differences observed in PFS (P < 0.0001) and OS (P = 0.0007) across groups. The model outperformed traditional prognostic indices such as IPI and R‑IPI.

The 5‑index risk model demonstrated robust predictive ability in a real‑world setting, providing a reliable basis for personalized treatment decisions in Chinese DLBCL patients undergoing CAR‑T. Future work will focus on further optimizing the model and conducting multi‑regional validation.