Individualizing Apheresis Processing Volumes Using a Predictive Algorithm: Analysis of the CAR T-Cell Collection Protocol Utilized in ACIT001/EXC002, a Phase Ib/II Trial of Decentralized Production of CAR T Cells for Treatment of Relapsed/Refractory Aggressive NHL and ALL.

Abstract

Chimeric antigen receptor (CAR) T-cell therapy has become an established treatment for relapsed or refractory B-cell malignancies. Apheresis collection of peripheral blood mononuclear cells (PBMCs) is a critical first step in CAR T-cell manufacturing. However, the total blood volume (TBV) required to achieve adequate T-cell yield is difficult to predict. As a result, standard practice often involves processing two to four TBVs, which may lead to unnecessary time on machine (TOM) and increased risk of procedure-related complications such as citrate toxicity, hypovolemia, and thrombocytopenia. As part of the ACIT001/EXC002 clinical trial evaluating point-of-care CAR T-cell manufacturing for relapsed/refractory non-Hodgkin lymphoma (NHL) or acute lymphoblastic leukemia (ALL), we developed a pragmatic protocol to individualize TBV processing. This protocol uses a preliminary TBV estimate based on an assumed CD3-positive T-cell collection efficiency (CE) of 50%, followed by a midpoint product assessment to measure actual CE and adjust the remaining TBV and TOM. Using this approach, the preliminary estimate was adequate or conservative in 72% of procedures, and over 75% of patients required less than the standard two to four TBVs. The target CD3-positive T-cell yield was achieved in 93% of patients. These findings demonstrate that a midpoint-adjusted TBV calculation protocol can minimize unnecessary TOM while maintaining reliable T-cell collection, thereby improving resource utilization and patient safety. TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT03938987.