Predicting chimeric antigen receptor apheresis process volume

The recent work by O'Reilly et al. [ ] provides an elegant tool for chimeric antigen receptor (CAR)–T-cell apheresis planning by generating a formula that predicts the process volume required to achieve a certain target dose of T cells. The formula is mathematically much more sophisticated than the referenced one that we had previously proposed. Our “formula” is solely the estimate that at least 40% of T cells presented at the connector will be collected into the product bag [ ]. O'Reilly et al. comment that the process volumes suggested by the “40% formula” regularly result in the collection of more than the required dose. Indeed, the 40% formula underpromises by design, as it is based on worst-case estimates of collection efficiency (CE). Past experience teaches that, although the mean CE for lymphocytes of 60% or so is achieved with current apheresis technology, CE is quite variable [ ]. A CE of 40%, and, accordingly, the targeted cell dose, will be exceeded by 90% of aphereses if the proposed volume can be processed. Although overcollection may be economically undesirable, from the patient's perspective, the meaningful risk is undercollection. Moreover, the formula by O'Reilly et al. is mathematically unable to predict small process volumes: where high T-cell concentrations are present in blood or modest T-cell doses are targeted, the formula predicts negative process volumes. Therefore, the minimum recommended process volume is 7 L. Experience teaches, however, that the mean CAR-T apheresis process volume in adults is less than 5 L [ ] and that 7 L is not reasonably achievable in most children. Targeting a T-cell dose of 10⁹, the dose required for manufacturing of Liso-cel, the formula by O'Reilly et al. proposes a process volume of 7 L for all patients with a T-cell concentration in the peripheral blood of at least 200/µL, i.e., the sometimes-mentioned lower threshold for CAR-apheresis eligibility.