Longitudinal immune monitoring in patients treated with CD19 and BCMA CAR-T cells using standardized 8‑color flow cytometry panel reveals substantial differences among products kinetics and persistence

Abstract Background: Chimeric antigen receptor T cells (CAR-T) represent a groundbreaking therapy in hematologic malignancies. Currently approved products target CD19 (axi-cel, brexu-cel, tisa-cel, liso-cel) and BCMA (cilta-cel, ide-cel). CAR-T expansion was linked to both clinical efficacy and toxicity, however, comparative data across CAR-T products remain limited, particularly longitudinal data on CAR-T kinetics and persistence of both CAR-T and non-CAR-T subsets and its clinical relevance. Aims: To assess and compare longitudinal kinetics of CAR-T and non-CAR-T immune subsets among different CD19 and BCMA CAR-T products and to determine clinical relevance of immune monitoring in CAR-T treated patients. Methods: In this multicenter study, 727 peripheral blood (PB) samples from 123 patients treated with axi-cel (sample/patient counts: 377/71), cilta-cel (177/26), tisa-cel (96/14), and brexu-cel (77/12) were analyzed. Samples were collected at baseline (Day 0 [D0]), then D7, D14, D21, D28, and monthly to Month 15 (M15). Specifically designed and standardized 8-color flow cytometry panel (CD3, CD4, CD8, CD19, CD20, CD45, CD16+CD56, and CAR) was used in routine practice. Results: Diagnoses included B-cell non-Hodgkin lymphoma (B-NHL) (71% [89/123]), multiple myeloma (MM) (21% [26/123]), and B-cell acute leukemia (ALL) (7% [8/123]). Median age was 63, 71, and 40 years for B-NHL, MM, and ALL respectively. Axi-cel was used in 80% (71/89) of B-NHLs. CRS occurred in 91%, 76%, and 75% with grade ≥2 observed in 54%, 28%, and 24%. ICANS occurred in 46%, 40%, and 38%,with grade ≥2 observed in 30%, 8%, and 12% of B-NHL, MM, and ALL cases. CR/VGPR was achieved in 70%, 96%, and 100% cases and 26%, 0%, and 12% progressed (PD). Median follow-up of was 364 days. Median CD45+ leukocytes per sample was 1.74 × 10⁵ and median limit of detection (LOD) was 0.0117%. First, CAR-T product expansion was compared among the products. Axi-cel and brexu-cel counts (cells/µL PB) peaked on D7, while tisa-cel and cilta-cel peaked on D14. Cilta-cel exhibited markedly higher peak counts (median: 599.1), compared to axi-cel (91.4), tisa-cel (86.8), and brexu-cel (35.3). Moreover, persistence in PB was assessed for axi-cel and cilta-cel, which had sufficient sampling. Axi-cel persisted at low levels often near LOD (median 0.8/µL) in 45% (17/38) to 41% (7/17) of patients from M4 to M13-15, while cilta-cel was undetectable from M3 onwards (16/16). Axi-cel patients showed limited B-cell recovery (5% to 35% from M4 to M13-15), while B-cells recovered in 100% of cilta-cel patients on M3 and sustained. Then, CD4/CD8 kinetics in the CAR-T pool were compared between axi-cel and cilta-cel. Axi-cel had initial low CD4+ proportions followed by CD4+ increase in persistence (D7–D21 30.9% vs M4–M15 85.2%; p=0.001), whereas cilta-cel showed early CD4+ dominance declining over time (D7 77.8% vs D28 43.0%; p<0.001). In order to assess the association of CAR-T cell expansion and clinical outcomes, we used our largest cohort of B-NHL patients treated with axi-cel (N=71). As anticipated, CAR-T expansion at D7 was linked to toxicity and efficacy. Increased number of CD8+ CAR-T cellsassociated with CRS ≥2 (p=0.048), with CD4+ showing statistical trend (p=0.06). Total CAR-T counts were elevated in CRS and ICANS ≥2 (p<0.012). CD4+ and CD8+ CAR-T counts were higher in CRS ≥2 (p<0.02), and CD8+ counts increased in ICANS ≥2 (p=0.005). Patients achieving CR had higher D7 CD8+ CAR-T counts (p=0.049); total and CD8+ counts at D14 were also elevated (p<0.024). Using a cutoff (Maxstat) of 129 CAR-T/µL at D7, low expansion predicted worse PFS (1-year PFS: 59% vs. 100%, HR 0.12, p=0.015) and lower CR rates (57% vs 82%, p=0.08). Conclusion: This is the first study comprehensively comparing CAR-T and non-CAR-T cell kinetics across different CAR-T products targeting CD19 and BCMA using a specialized flow cytometry panel. Distinct peak kinetics and persistence patterns were observed, with cilta-cel showing higher peaks and rapid disappearance by M3, accompanied by B-cell recovery, while axi-cel tends to persist for longer time period in a subset of patients, with prolonged B cell aplasia. In axi-cel patients, greater CAR-T expansion was associated with higher toxicity but also improved response and PFS.