Clinical Impact of Cytokine Release Syndrome on Prolonged Hematotoxicity after Chimeric Antigen Receptor T Cell Therapy: KyoTox A-Score, a Novel Prediction Model

Prolonged hematotoxicity is the most common long-term adverse event in chimeric antigen receptor T cell therapy (CAR-T). To evaluate the impact on prolonged cytopenia of inflammatory status after CAR T infusion, we performed a single-center retrospective study and analyzed patients with B cell lymphomas after CAR-T. Among 90 patients analyzed at 90 days after infusion, the cumulative incidence was 57.5% for prolonged neutropenia, 36.7% for anemia, and 49.8% for thrombocytopenia. Patients who experienced cytokine release syndrome (CRS) had significantly higher incidence and longer duration of prolonged cytopenia. In addition, we found that among patients with grade 1 CRS, those with a longer duration of CRS-related symptoms (>5 days; grade 1b in modified CRS grading m-CRS) had a significantly higher incidence and longer duration of prolonged cytopenia than those whose CRS-related symptoms resolved within 5 days (grade 1a m-CRS). Multivariate analysis revealed that a higher m-CRS grade (grade 1b or 2; hazard ratio HR, 2.42), higher peak CRP (≥10 mg/dL; HR, 1.66), longer duration of elevated CRP (≥10 days; HR, 1.83), and a decrease in serum inorganic phosphorus concentration (≥30% from baseline; HR, 1.95) were associated with significantly higher cumulative incidence of prolonged neutropenia, as well as anemia and thrombocytopenia. Using these factors, we developed a new predictive scoring model for prolonged hematotoxicity, the KyoTox a-score, which can successfully stratify the incidence and duration of cytopenia independent of the existing model, CAR-HEMATOTOX, which is based on laboratory data at lymphodepletion. Thus, this newly developed post-CAR-T inflammation-dependent score is accurate and useful for predicting prolonged hematotoxicity.