Summary of US Food and Drug Administration Chimeric Antigen Receptor (CAR) T-Cell Biologics License Application Approvals From a Statistical Perspective

Abstract

The approval of tisagenlecleucel and axicabtagene ciloleucel in 2017 marked a milestone in the development of oncology therapies. Since 2017, the breakthrough in treatment or even cure of previously intractable diseases represented by this new class of cancer treatments has continued with subsequent chimeric antigen receptor T (CAR T)-cell approvals. To date, the US Food and Drug Administration has approved five autologous CAR T-cell products for seven indications. A feature of autologous CAR T-cell products that differentiates them from traditional oncology drugs is that they need to be manufactured specifically for each patient. This feature has implications in study design, statistical analyses, and interpretation of study results. In this article, we share our experiences in the statistical review of CAR T-cell products and provide considerations for the design and statistical analyses of CAR T-cell trials. We also describe how the newly adopted estimand framework for clinical trials can help clarify nuanced issues in CAR T-cell trial design.

Reply to W.H. Tong et al

“…When a new anticancer therapy is administered, US Food and Drug Administration review teams recommend censoring DOR [duration of response] at the last adequate disease assessment before new anticancer therapy. The reasoning for such censoring is to isolate the treatment effect of chimeric antigen receptor T-cell therapy from that of subsequent therapies 4 We agree that independent censoring during the Kaplan-Meier analysis is a strong assumption. Exact reasons for SCT were not collected, and although the occurrence of SCT was likely independent of disease status in some cases (eg, on the basis of physicians' preferences at the time of study conduct when little was known about long-term chimeric antigen receptor T-cell therapy effects and the durability of its response), it was likely related to a patient's status in others.…”

“…Censoring SCT in Kaplan-Meier analysis is one statistical approach to answer to this question. The recent article by Lin et al 4 outlines US Food and Drug Administration guidance for the handling of subsequent therapies (including SCT):…”

“…In alignment with health authorities, the analysis in the pivotal ELIANA trial focused on evaluating the effect of tisagenlecleucel on RFS in a world where SCT statistical approach to answer to this question. The recent article by Lin et al 4 outlines US Food and Drug Administration guidance for the handling of subsequent therapies (including SCT):…”

Should endovascular stenting be used routinely as first-line treatment for malignant superior vena cava syndrome?—a critical review in the context of recent advances in oncological treatments

“…At a median follow-up of 51.1 months, the 4-year overall survival rate was 44%, suggesting that a proportion of patients in this poor prognostic group may achieve long-term survival with CAR-T therapy (63). It is important to note that CAR-T therapy requires a cell-manufacturing process that takes a median of 15 to 33 days (64). Bridging therapy can be indicated for SVCS patients planned for CAR-T therapy to avoid symptomatic progression.…”

Statistics in Oncology

“…Autologous CAR T products are different from traditional oncology drugs in needing to be manufactured specifically for each patient. The review by Lin et al 2 provides important statistical considerations for this unique class of products and highlights discussions at the trial design stage, at tumor response readjudication during the biologics license application stage, and during the prescribing information negotiations. The need for autologous CAR T products to be specifically manufactured for each patient has implications for multiple aspects of the study design and analysis.…”