An assessment model for efficacy of patients with relapsed or refractory diffuse large B-cell lymphoma following autologous CD19 chimeric antigen receptor T-cell therapy.

Abstract

Abstract Background Currently, there is no predictive model for the efficacy of autologous CD19 chimeric antigen receptor T-cell therapy (CAR-T) in relapsed or refractory diffuse large B-cell lymphoma (r/r DLBCL). This study aims to construct a comprehensive model that takes into account numerous influencing factors to predict the efficacy of CD19 CAR-T therapy. Methods A total of 80 r/r DLBCL patients receiving CD19 CAR-T therapy from two centers were included in the study. Multivariable logistic regression analysis model was constructed using data from CAR-T clinical trials as the derivation cohort and real-world data as the validation cohort. Results The model was optimized based on the results of clinical practice and further developed into an index model, which demonstrated excellent predictive utility in both the derivation cohort (C-index = 0.891) and the external validation cohort (C-index = 0.797). Calibration curve, decision curve analysis, and clinical impact curve confirmed the clinical utility of the predictive model. The risk stratification based on the index model can indicate differences in progression-free survival (PFS) and overall survival (OS). Conclusions the established predictive model for assessing the efficacy of CD19 CAR-T treatment in r/r DLBCL is accurate and clinically useful.