Predicting ICANS by means of plasma CAR‐T cell derived extracellular vesicles in patients undergoing infusion of anti‐CD19 CAR‐T cells

Abstract

Introduction: Immune effector cell-associated neurotoxicity syndrome (ICANS) is a life-threatening adverse effect of anti-CD19 chimeric antigen receptor (CAR) T-cell therapy that usually occurs within 5–7 days after cell infusion. Although several clinical and biochemical parameters have been associated with ICANS, it is still a matter of debate how to predict its onset at the patient level. We here tested the hypothesis that CAR-T cell derived extracellular vesicles (EV) carrying the engineered CAR protein and produced early after CAR-T cell activation can be used as predictive biomarker of ICANS. Purposely, we measured plasma CAR+ EV in lymphoma patients underwent anti-CD19 CAR-T cell therapy. Methods: Seventy-one patients with aggressive r/r B-cell lymphomas were admitted to the advanced cell therapy unit of IRCCS AOU of Bologna (NCT04892433) for anti-CD19 CAR-T cell infusion. Included patients received tisa-cel (n = 27), axi-cel (n = 34), or brexu-cel (n = 10) after a median number of 3 prior lines of treatment (2–11); median age was 62 years (19–76) and no patients had CNS disease at the time of CAR-T cell infusion. Twenty out of 71 patients (28%) had ICANS of any grade: 5 patients (7%) ICANS grade 1, 7 patients (10%) ICANS grade 2 and 8 patients (11%) ICANS grade ≥3 (3 patients ICANS grade 3, 3 patients ICANS grade 4 and 2 patients ICANS grade 5 with diffuse cerebral edema). ICANS was classified according to Lee et al. The median time from CAR-T cell infusion to ICANS onset was 5 days (3–12). Available plasma samples at day +1 after CAR-T cell infusion were analyzed for CAR+ EV by FACS analysis. Data analysis was performed with Prism software v9.1.3 (GraphPad). Results: CAR+ EV were already detectable +1 day after CAR-T cell infusion in 58 patients. The median onset of ICANS was at day +5 (3–12). Patients with ICANS of any grade showed higher CAR+ EV level compared to no-ICANS ones (p < 0.0001). CAR+ EV anticipated the median ICANS onset of 2 to 11 days. CAR+ EV ROC analysis showed that a concentration >187.5 CAR+ EV/μl at day +1 after infusion predicts ICANS onset with sensitivity of 100% and specificity of 83.33% (p < 0.0001). Conclusions: These findings lead us to hypothesize that the plasma level of CAR+ EV mirrors target engagement by CAR-T cells, and their massive release is related to ICANS. Thus, CAR+ EV level could be considered a putative early predictor of ICANS onset; further analyses in larger cohorts are warranted to confirm this finding. Keywords: Cellular therapies, Diagnostic and Prognostic Biomarkers Conflicts of interests pertinent to the abstract. P. L. Zinzani Consultant or advisory role: Secura Bio, Celltrion, Gilead, Janssen-Cilag, BMS, Servier, Sandoz, MSD, AstraZeneca, Takeda, Roche, EUSA Pharma, Kyowa Kirin, Novartis, ADC Therapeutics, Incyte, BeiGene Other remuneration: Speakers bureau: Celltrion, Gilead, Janssen-Cilag, BMS, Servier, MSD, AstraZeneca, Takeda, Roche, EUSA Pharma, Kyowa Kirin, Novartis, Incyte, BeiGene F. Bonifazi Consultant or advisory role: Novartis, Gilead