A comparison of two manufacturing methods in the phase I COBALT study of CD19CAR T for LBCL

As demand for CAR T products increases, finding solutions to manufacturing bottlenecks becomes critical. While most current FDA-approved CAR T products are manufactured using traditional bag-based manufacturing methods, semi-automated manufacturing platforms can simplify and expedite autologous CAR T product delivery to patients. We performed the phase I COBALT study (NCT02431988) of 2^nd-generation CD19 CAR T cells for relapsed/refractory (r/r) large B cell lymphoma (LBCL). Here, we compared a manual, bag-based, IL2-supplemented manufacture process (process-A) with a CD4/8-pre-selected, semi-automated, interleukin (IL)-7/IL-15-supplemented Miltenyi CliniMACS Prodigy-based manufacturing process (process-B), with a focus on the drug product and manufacturing feasibility and logistics. GMP scale-up runs using leucapheresis products from people with LBCL showed that process-B delivered the target CAR T dose more consistently than process-A, with lower viral vector usage, less grade A clean room time, and less hands-on staff time required per product. On study, ten patient-specific products were manufactured (5 with process-A; 5 with process-B). 6 of 10 products reached the target dose (2 of 5, process-A; 4 of 5, process-B), and 9 of 10 patients were infused. Higher early CAR T expansion was observed in patients treated with process-B products. In this analysis within the COBALT study, process-B compares favorably with process-A in reproducibly reaching the target CAR T dose in people with r/r LBCL, and appears to be associated with better CAR T expansion in vivo.