苹果与橘子之比较：ZUMA-7、TRANSFORM 与 BELINDA 试验

In December 2021, three phase III trials investigating Chimeric Antigen Receptor (CAR) T-cell for large B-cell lymphoma were published, only one of which showed no treatment effect on Event-Free Survival (EFS). All compared anti-CD19 CAR T-cell as second-line treatment with immunochemotherapy plus autologous stem cell transplant if an adequate response to chemotherapy was achieved. In this letter, we discuss the methodological reasons that partially explain the discrepant results observed between the ZUMA-7, TRANSFORM and BELINDA trials. A raw comparison shows that BELINDA simultaneously had the worst experimental arm and the best control arm among the three trials. This could be partially related to differences in the event definition and time of assessment.

Stable Disease was considered an event as early as W9 in TRANSFORM, W12 in BELINDA and only W21 in ZUMA-7. Since tisa-cel had the longest manufacturing time, the time window may have been too short to assess its full potential compared with axi-cel and liso-cel. In comparison, a patient with stable disease in ZUMA-7 would not be considered an event until W21.

On the other hand, a second salvage regimen was allowed before considering stable disease as an event only in the BELINDA control arm which could have delayed EFS. Many of these issues could be avoided if progression-free survival was preferred to EFS and if the time to manufacture CAR-T cells was shortened, as long delays can result in a higher tumor volume and more refractory diseases at the time of infusion.