Solvitur ambulando

Sidana et al¹ report on their real-world experience (RWE) of the use of ciltacabtagene autoleucel (cilta-cel) for patients with relapsed myeloma with the goal of answering the question of how to use cilta-cel. Reading the publication of the cilta-cel registrational trial (CARTITUDE-1) data provides the safety and efficacy foundation in a selected population that drove the approval for use for patients with relapsed myeloma: deep responses with high response rates, prolonged median progression-free survival (PFS), and an overall acceptable risk-benefit profile.^2,3 What we do not learn from the publication is how to actually use the drug in our clinics for all who are candidates. In this issue of Blood, Sidana et al have tried to solve that disconnect in this real-world analysis; in Latin, Solvitur ambulando, or “It is solved by walking.” They teach us the “how.”

Sixteen academic medical centers in the United States collaborated to form the Multiple Myeloma Immunotherapy Consortium for this real-world evidentiary study, which analyzed patients with relapsed multiple myeloma treated with cilta-cel. This publication is critically important for 2 reasons, validation of the trial experience and filling the information gap. Firstly, it provides confirmation of the original clinical trial data in the subgroup of real-world patients who would have met eligibility for the CARTITUDE 1 study (n = 108, 44% of RWE data). Secondly, and more importantly, 56% of the patients in this real-world study would not have been eligible for CARTITUDE 1 due to, most commonly, prior use of B-cell maturation antigen (BCMA)-targeted therapies (BCMA-TTs), poor performance status, and comorbidities. This is where we learn how to use cilta-cel in our entire patient population.

CARTITUDE-1 excluded those who had prior BCMA-TT. In this study, 14% had prior BCMA-TT, and the overall response, complete response, and PFS were inferior to those patients who had not had prior BCMA-TT. This was borne out both in direct comparisons (see Figure 4 in the article by Sidana et al) and in the multivariable analysis (Table 3 in Sidana et al). However, it is not that simple. Sidana et al then compared those who were treated within 6 months with a BCMA-TT or 6 months and beyond and found a significant interaction. Those patients that were treated within 6 months with a prior BCMA-TT had lower overall response rate (54% vs 94%) and lower complete response rate (8% vs 19%) than those patients treated beyond 6 months from prior BCMA-TT. There are not enough patients to determine which type of prior BCMA-TT has a greater impact: antibody-drug conjugates, bispecific antibodies, or other chimeric antigen receptor T cell (CAR-T) therapies. Future real-world analysis and prospective studies should focus on the questions of timing and prior BCMA-TT.

Sidana et al assessed manufacturing success and the impact of using out of specification (OOS) CAR-T products. There was a 1% manufacturing failure, and 19% of patients had an OOS product. The OOS cohort had modestly reduced efficacy outcomes compared with the standard CAR-T products. Risk factors for having an OOS product were a prior history of autologous stem cell transplant, prior bispecific antibody therapy, higher baseline ferritin, and higher baseline C-reactive protein. This argues that an OOS product is unlikely to be the cause of any difference in outcome but rather is the effect of adverse baseline characteristics.

What factors impacted outcomes? High-risk cytogenetics, poor performance status, presence of extramedullary disease, and prior BCMA-TT are all associated with a decrease in PFS. Each of these factors negatively affect any of our current therapies. It is more important in our decision-making to understand what factors do not impact outcomes. In this study, neither age, history of plasma cell leukemia, response to bridging therapy, nor type of lymphodepleting chemotherapy negatively impacted PFS.

Toxicity remains a significant challenge in the use of cilta-cel. Treatment-emergent adverse events in this real-world analysis, perhaps surprisingly, were very similar to that of CARTITUDE-1. There is a high rate of CRS that can be serious and is rarely fatal. Neurologic toxicity is uncommon and can be serious and rarely fatal. Similar rates of late neurologic toxicity such as cranial nerve palsy, parkinsonism, and other rare disorders were observed. Overall, the non-disease-related mortality was 10%, similar to CARTITUDE-1.

This real-world analysis suffers the problems of many real-world analyses. The data collection was retrospective in nature. There is no control population. Findings are hypothesis generating and not definitive. This group employed several mechanisms to improve the analysis, including uniform case data collection and intent-to-treat analyses from time of leukapheresis.

Prospective randomized studies are necessary for new drug approval and as the gold standard for the evaluation of new therapies in myeloma. Complementary prospective studies in unique subsets further support these randomized studies. And real-world analysis by dedicated investigators using uniform data sets can help solve some of the unanswered questions. In the future, we should expand eligibility criteria for clinical trials to ensure they more align with those who will receive the drug after approval, obviating the need for these real-world analyses.

Conflict-of-interest disclosure: J.L.K. declares consultancies for Sanofi, Bristol Myers Squibb (BMS), Ascentage, Sebia, AbbVie, and Genentech and research funding from AbbVie, BMS, Novartis, Heidelberg Pharmaceuticals, Genentech, Takeda, and Janssen.