Identification of Predictors of CRS and Neurotoxicity Duration after Axicabtagene Ciloleucel Therapy

Chimeric antigen receptor T-cell (CAR-T) therapy has transformed the treatment of high-risk B-cell malignancies, but it remains associated with high rates of cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). While prolonged toxicities are also observed in a subset of patients, to date the factors associated with the duration of CRS and ICANS are unknown. Here, we provide the first characterization of factors associated with time to CRS/ICANS resolution in patients receiving the FDA-approved CD19 CAR T-cell product axicabtagene ciloleucel (axi-cel). Eighty-four patients (pts) treated at the Fred Hutch Bezos Family Immunotherapy Clinic with axi-cel per the FDA label outside of a clinical trial between 1/2018 and 4/2022 were included. CRS and ICANS were graded (G) using ASTCT criteria. Cause-specific Cox regression models were used to estimate associations between 60+ pre/post-infusion patient, disease-related, laboratory variables (e.g., platelet count, creatinine, CRP, ferritin), and time to CRS/ICANS resolution. Time to CRS/ICANS resolution was defined as the time from axi-cel infusion left-truncated at the time of CRS/ICANS onset until CRS/ICANS symptom resolution (first day of grade 0 for at least three consecutive days), whichever toxicity happened last, censoring for death from any cause. Baseline characteristics and toxicity outcomes are shown in Table 1. We observed any-grade CRS in 72 pts (86%) and G≥3 CRS in 7 pts (8%). Fifty-two pts (62%) developed any-grade ICANS, and 16 pts (19%) G≥3 ICANS. The median time to CRS/ICANS resolution was 10 days (95%CI: 9,11). Two pts (2%) died prior to CRS/ICANS resolution. Using univariate cause-specific Cox regression, the following pre-infusion factors were associated with longer time to CRS/ICANS resolution: lower platelet count at day 0 (HR=2.45, 95%CI: 1.02-5.88, p=0.044), higher PTT at day 0 (HR=0.05, 95%CI: 0.00-0.86, p=0.039), higher preLD creatinine (HR=0.06, 95%CI: 0.01-0.50, p=0.009), and peak CRS severity (HR = 0.73, 95%CI: 0.55-0.97, p=0.029). Cumulative incidence curves of time to CRS/ICANS resolution for select predictors are shown in Figure 1. In a multivariable Cox model after LASSO variable selection including day 0 PTT, ferritin, and peak CRS severity, lower day 0 platelet count remained independently associated with longer time to CRS/ICANS resolution (HR=6.96, 95%CI:1.09-44.5, p=0.04; model discrimination: C-index=0.65). This is to our knowledge the first study aimed at identifying factors associated specifically with the duration of CRS and ICANS after CD19 CAR T-cell therapy. We found that pts with lower platelet count, high PTT, and altered renal function were at high risk of developing prolonged toxicities. Future clinical trials could use these factors to risk-stratify pts and target populations more likely to benefit from toxicity mitigation strategies.