CD22 CAR T-Cells Manufactured in Prodigy System and in Bag Culture Yield Equivalent Responses for B-ALL

Introduction While all currently FDA approved CAR T-cells have been manufactured using bag culture, there is growing interest in closed-system bioreactors to facilitate decentralized manufacturing, improve access, and decrease cost. It is critical to understand how new platforms impact CAR T-cell functionality, yet direct comparisons between platforms are lacking. In the context of CD22 CAR T-cells for B-cell acute lymphoblastic leukemia (B-ALL), we compare clinical outcomes across two manufacturing platforms. Methods We retrospectively compared outcomes for children and young adults with B-ALL treated with CD22 CAR T-cells at Stanford (NCT04088890, NCT04088864) and the National Cancer Institute (NCI) (NCT02315612). Both used an identical construct and CD4/CD8 selected apheresis, but Stanford used a CliniMACS Prodigy protocol and the NCI used bag culture. Stanford's expansion dose was 1e6 cells/kg and NCI's was 0.3e6 cells/kg. Our primary aim was to compare best response at Day 28, based on morphology and minimal residual disease (MRD) by flow cytometry. We also analyzed CAR T-cell expansion, toxicities, and inflammatory markers. Mann-Whitney tests and Fisher exact tests were used to compare groups. Results Across 57 evaluable patients, 41 (71.9%) received BC-based cells, and 16 (28.1%) received Prodigy-based cells. BC patients were younger (median 15.2 vs 23 years in Prodigy; p=0.0034), had lower baseline lymphocyte count (median 500 vs 1605 cells/mcL in Prodigy; p=0.0010), and higher baseline bone marrow (BM) involvement (76% with > 5% blasts in BM vs 44% in Prodigy; p=0.031). Prevalence of non-central nervous system (CNS) extramedullary disease (EMD) was similar, but the Prodigy group had more CNS2/3 disease (2.4% BC, 25% Prodigy; p=0.019). There were no differences in prior therapies.Complete response (CR) rates were similar, with 80% CR in the BC group and 75% in the Prodigy group. (Fig 1) Peak expansion in those with high BM disease burden did not differ significantly, regardless of EMD. For those with low BM disease burden and EMD, the BC group had greater expansion (median 4075 cells/mcL in BC vs 13.9 cells/mcL in Prodigy; p=0.015), without differences in CR rates. (Fig 2)While systemic inflammation in the BC group was higher, including peak ferritin (64978 vs 5940 ng/mL for Prodigy; p=0.0018) and peak C-reactive protein (median 170 vs 76 mg/L for Prodigy; p=0.0073), there were no significant differences in incidence or severity of cytokine release syndrome, neurotoxicity, or hemophagocytic lymphohistiocytosis-like toxicity. Conclusion While different expansion patterns were observed, CD22 CAR T-cells manufactured in the Prodigy were as safe and effective as those made by bag culture. As commercial CAR T-cells move to automated manufacture, CAR T-cell phenotype, expansion, and functionality especially in patients with EMD may require further study.