Introduction: Ciltacabtagene autoleucel (cilta-cel) is a B-cell maturation antigen–directed chimeric antigen receptor-T cell therapy approved in the United States and European Union for adults with lenalidomide-refractory multiple myeloma as early as first relapse. Cilta-cel is manufactured using autologous T cells, which results in variability in the initial starting material and subsequent manufactured product, including drug products that do not meet health authority specifications (referred to as out-of-specification [OOS]). Quality attributes of T cells obtained from each patient (via apheresis) are the predominant contributors to OOS manufacturing outcomes (Baguet et al. Blood Adv. 2024;8:337-342). Moreover, manufacturing of in-specification drug products is associated with the enrichment of naive and central memory T-cell subpopulations, which are more abundant in material collected from patients with relatively fewer prior lines of therapy (pLOTs; Ayala et al. J Exp Med. 2024;221:e20230903). This study examined cilta-cel manufacturing outcomes across multiple pLOTs.

Methods: Commercial cilta-cel manufacturing data were analyzed to identify determinants of manufacturing outcomes, define biomarkers predictive of manufacturing success (defined as in Sanoyan et al. BMC Cancer. 2023;23:345), and assess results following OOS-driven remanufacturing or recollection.

Results: Rates of first-time manufacturing success significantly improved when using T cells from patients with fewer pLOTs (93% with ≥1 pLOT to 89% with ≥4 pLOTs; N=~3000). CD3+ cell recovery at the initiation of manufacturing was less favorable in patients with ≥4 versus fewer pLOTs. Furthermore, the initial cellular CD4:CD8 ratio improved by 54% with fewer pLOTs (average CD4:CD8 ratio, 1.0 with 1 pLOT vs 0.65 with ≥4 pLOTs); these data also suggest that CD4:CD8 ratio >0.5 may contribute to better in-specification, first-time manufacturing outcomes. Recovery of viable cells before lentiviral vector transduction and cumulative population doubling level of cells improved (up to 7%) with relatively fewer pLOTs (P<0.05 for both). Markers of metabolic activity, including glucose consumption and lactate production, were also significantly improved with fewer pLOTs (average improvement of 10% for material from patients with 1 pLOT versus more pLOTs).

Across cilta-cel OOS outcomes from all pLOTs (n=317), 63% were released through an expanded access program without remanufacturing. Of the remaining 37%, 8.6% were canceled. An additional 28.6% (n=90) underwent remanufacturing, 60% (n=54) resulted in OOS or cancellation, and 40% (n=36) resulted in within-specification outcomes. Notably, material from patients with ≥4 pLOTs undergoing remanufacturing resulted in the same or additional OOS results or cancellations as the first manufacturing OOS outcome (67% of attempts [n=36]). Remanufacturing from recollection of new apheresis material was requested in 14% of initial OOS cases as a result of insufficient remaining material or because of concerns about the effects of prior therapies, suggesting long-lasting impact of prior treatments and/or disease characteristics. Remanufacturing attempts using new apheresis material (n=45) resulted in similar or worse outcomes in 31% of cases, and 20% of cases resulted in cancellation of the order. Furthermore, cancellation and OOS outcome rates for remanufacturing with ≥4 pLOTs were 39.5% and 35.5%, respectively (n=29). Overall, the manufacturing success rate improved to 99% when using cells from patients with ≥1 pLOT (6.5% received an OOS product) compared with 97% for ≥4 pLOTs (9.2% received an OOS product).

Conclusions: Rates of first and overall manufacturing success were higher when using cells collected from patients with ≥1 versus ≥4 pLOTs. T-cell attributes associated with manufacturing outcomes were also improved in starting material collected from patients with ≥1 versus ≥4 pLOTs. Remanufacture of product often results in additional OOS outcomes, which may further delay patient treatment.

This content is only available as a PDF.
2025
Sign in via your Institution