T-ICAHT: Grading and Prognostic Impact of Thrombocytopenia After CAR T-cell Therapy
Kai Rejeski, Jaime Sanz, Teng Fei, Monica Nair, Hamza Hashmi, Abraham Avigdor, Ofrat Beyar‐Katz, Veit Bücklein, Kevin J. Curran, Sigrun Einarsdottir, Jonathan H. Esensten, Netta Glaubach, Noa Golan-Accav, Marina Gómez-Llobell, Iris Halamis, Orit Itzhaki, Frederick L. Locke, Sham Mailankody, Ronit Marcus, Marcela V. Maus, M. Lia Palomba, Jae H. Park, Marcelo C. Pasquini, Sandeep Raj, Sridevi Rajeeve, Gilles Salles, Michael Scordo, Gunjan L. Shah, Avichai Shimoni, Marion Subklewe, Tobias Tix, Saad Z. Usmani, Ori Ben Valid, Yannis K. Valtis, Tsila Zuckerman, Nirali N. Shah, Miguel‐Angel Perales, Roni Shouval
Abstract
ABSTRACT: Immune effector cell-associated hematotoxicity (ICAHT) was recently introduced as a distinct toxicity category of chimeric antigen receptor (CAR) T-cell (CAR-T) therapy. Although a grading system based solely on neutrophil counts was proposed (hereafter termed N-ICAHT), the prevalence and prognostic impact of thrombocytopenia remain poorly defined. In this multicenter observational study, we systematically examined patterns of thrombocytopenia in 744 patients treated with commercial CD19 CAR-T for B-cell non-Hodgkin lymphoma (B-NHL). We developed a grading system termed T-ICAHT, with thresholds that closely aligned with N-ICAHT, based on depth, duration, and timing of thrombocytopenia. In the core NHL data set, 43% of patients developed any-grade early T-ICAHT (days 0-30), with 23% developing severe (grade ≥3) manifestations. Late T-ICAHT (days 31-100) was observed in 42% (grade ≥3, 13%). Although T-ICAHT and N-ICAHT gradings showed some correlation, considerable discordance was noted. On multivariate analysis, bridging therapy, poor performance status, and high HEMATOTOX scores were associated with increased risk of severe early T-ICAHT. Patients with higher T-ICAHT grades showed increased platelet and red blood cell transfusion burden and more bleeding events. T-ICAHT grades were inversely associated with overall survival (OS), with landmarked 2-year estimates ranging from 67% (grade 0) to 48% (grade 1-2) and 35% (grade ≥3). In multivariable Cox regression analysis, the independent prognostic capacity of T-ICAHT for OS was confirmed. Finally, we validated T-ICAHT's clinical and prognostic utility in 3 external cohorts spanning an additional 599 pediatric and adult patients (NHL, multiple myeloma, and B-cell acute lymphoblastic leukemia), confirming its broad applicability. These findings support integrating T-ICAHT into the ICAHT framework to standardize thrombocytopenia grading in CAR-T recipients.