Abstract
Although chimeric antigen receptor (CAR)-T cell therapies are highly effective for B-cell lymphoma, they frequently cause cytokine release syndrome (CRS). High-grade CRS is serious and may require intensive care, yet reliable early predictive markers remain elusive. To identify risk factors for high-grade CRS, we retrospectively analyzed B-cell lymphoma patients who received CD19 CAR-T cell therapy. Of 106 patients analyzed, CRS occurred in 93 (88%), Grade ≥ 2 CRS in 28 (26%), and Grade ≥ 3 CRS in 6 (6%). Reticulocyte counts at infusion were significantly lower in patients who developed Grade ≥ 2 CRS (1.85 vs. 2.80 × 10(4)/µL, p = 0.02). Multivariate analysis identified low reticulocyte count (< 15,000/µL; HR 2.21; 95% CI 1.01-4.86; p = 0.048), high metabolic tumor volume (> 100 mL), and use of axicabtagene ciloleucel as independent risk factors for Grade ≥ 2 CRS. Stratification by the reticulocyte cutoff showed higher 30-day cumulative incidence of CRS in patients with low counts, for both Grade ≥ 2 (42.9% vs. 19.7%, p = 0.012) and Grade ≥ 3 CRS (17.9% vs. 1.3%, p < 0.001). KyoTox-CRS, a risk-scoring system integrating these factors, effectively stratified these CRS risks. Early prediction of high-grade CRS based on the reticulocyte count at infusion may help to guide optimal risk-based management of CAR-T cell therapy.