T lymphocyte collection is essential for CAR T-cell engineering in refractory hematologic malignancies but needs to be optimised. No guidelines have been established for predicting the amount of T lymphocytes to be collected. The quantity of lymphocytes and especially T cells collected depends on the pre-cytapheresis lymphocyte blood level (pcLBL) and the number of blood volumes (BVs) processed. Our aim was to define and standardise a simple method for predicting the number of T lymphocytes collected, taking into account the number of BVs processed and the pcLBL regardless of the procedures defined by different companies.

We used data from our large retrospective series, which included 407 collection sessions using the same cytapheresis method in 400 patients mainly being followed up for non-Hodgkin’s lymphoma (NHL) or multiple myeloma (MM). We initially analysed the performance of lymphocyte collections using collection efficiencies (CE1 and CE2), which are indices that determine the ability to collect as many cells as possible, and also assessed the percentage of neutrophils collected. Finally, we evaluated whether the number of T cells collected could be easily predicted by multiplying the pcLBL and number of BVs by an average factor.

In our series, CE1 and CE2 for total lymphocytes and T cells were between 76 ± 15% and 69 ± 15%, thus confirming adequate cell collection. A low percentage of neutrophils was collected (9 ± 12%). Confirmation of adequate cell collection led us to consider the relationship between pcLBL and T-cell collection. We then demonstrated that the amount of T cells collected correlated with pcLBL, and could be predicted by multiplying pcLBL by 2.5 for each BV processed.

Easy prediction of T-cell collection is an important tool that can help apheresis and haematology teams monitor collection sessions, regardless of the companies involved and CAR T-cell technology.