Derivation and validation of a novel score for early prediction of severe <scp>CRS</scp> after <scp>CAR‐T</scp> therapy in haematological malignancy patients: A multi‐centre study

Abstract

Chimeric antigen receptor T (CAR-T) cell therapy is highly effective in inducing complete remission in haematological malignancies. Severe cytokine release syndrome (CRS) is the most significant and life-threatening adverse effect of this therapy. This multi-centre study was conducted at six hospitals in China. The training cohort included 87 patients with multiple myeloma (MM), an external validation cohort of 59 patients with MM and another external validation cohort of 68 patients with acute lymphoblastic leukaemia (ALL) or non-Hodgkin lymphoma (NHL). The levels of 45 cytokines on days 1-2 after CAR-T cell infusion and clinical characteristics of patients were used to develop the nomogram. A nomogram was developed, including CX3CL1, GZMB, IL4, IL6 and PDGFAA. Based on the training cohort, the nomogram had a bias-corrected AUC of 0.876 (95% CI = 0.871-0.882) for predicting severe CRS. The AUC was stable in both external validation cohorts (MM, AUC = 0.907, 95% CI = 0.899-0.916; ALL/NHL, AUC = 0.908, 95% CI = 0.903-0.913). The calibration plots (apparent and bias-corrected) overlapped with the ideal line in all cohorts. We developed a nomogram that can predict which patients are likely to develop severe CRS before they become critically ill, improving our understanding of CRS biology, and may guide future cytokine-directed therapies.