Intention to treat versus modified intention-to-treat analysis in B-cell maturation antigen and CD19 chimeric antigen receptor trials: A systematic review and meta-analysis

Introduction Chimeric antigen receptor T-cell therapy (CART) has revolutionised treatment of haematological malignancies; however, current reporting uses a modified intention-to-treat analysis (mITT) which over-estimates efficacy. We assessed what proportion of CD19 and B-cell maturation antigen (BCMA) CART trials report the number of patients not receiving CART after being enrolled by performing meta-analysis of the mITT and intention-to-treat (iTT) overall response rate (ORR). Methods PubMed/MEDLINE, EMBASE and Cochrane databases were searched. All prospective clinical trials of CD19 and BCMA-targeting CART enrolling two or greater patients from 1st January 2013 to 1st November 2020 were included. Results A total of 28 BCMA CART and 74 CD19 CART trials were identified. These included 10 BCMA CART (35.7%) and 52 (70.2%) CD19 CART trials reporting total number of patients enrolled and number of patients treated with CART. For this cohort of trials, the mITT ORR for BCMA CART was 78.0% (95% confidence interval (CI) = 67.0–89.0%), and the iTT ORR was 70.0% (95% CI = 59.0–80.0%). For CD19 leukaemia CART, the mITT ORR was 87.2% (95% CI = 83.1–91.2), and the iTT ORR was 74.9 (95% CI = 64.8–85.0). For CD19 lymphoma CART, the mITT ORR was 70.7% (95% CI = 63.9–77.5), and the iTT ORR was 58.7% (95% CI = 49.7–67.7). Conclusion Across BCMA and CD19 CART trials, there is a difference of up to 8–12% in the ORR between modified and iTT analyses and a paucity of information regarding reasons why patients did not receive the intended study treatment.