Reduction in Platelet Counts May Predict the Development of Immune Effector Cell Associated Neurotoxicity Syndrome in Patients Receiving CD19 CAR-T Cell Therapy for Relapsed/Refractory Large B Cell Lymphoma

Introduction: Anti-CD19 CAR-T cell therapy is an effective treatment for relapsed/refractory large B cell lymphoma (LBCL). However, significant toxicity remains of concern. Immune effector cell associated neurotoxicity syndrome (ICANS), presenting with symptoms like confusion, altered consciousness, seizures, and, rarely, cerebral edema, affects 10-60% of patients. There is a recognized association between the inflammatory state and coagulation. Small cohort studies suggest a potential link between abnormal laboratory values of coagulation parameters and the ICANS development. The objective of the current study was to evaluate the utility of coagulation parameter dynamics as predictors of ICANS. Methods: This international retrospective study encompassed patients from two tertiary care academic centers. The following laboratory coagulation data, collected from patients following CD19 CAR-T cell infusion, were analyzed: platelet (PLT) counts, prothrombin time (PT), activated partial thromboplastin time (PTT), international normalized ratio (INR), D-dimer levels, and fibrinogen levels. The measurements were taken daily before lymphodepletion and continued for two weeks post-infusion. To assess the statistical significance of differences between populations, two-sided two-sample t-tests were conducted using MATLAB R2023b. Test values for each day were categorized based on the ICANS grade. The “no-ICANS” group included values of patients without ICANS and those up to 2 days before the ICANS onset; the “ICANS group” included the values obtained in patients on the day immediately preceding the ICANS onset. A similar analysis was performed for high-grade ICANS (grades 3-4). The ‘ttest2‘ function was used to compute p-values, indicating whether significant differences existed between the groups, with results interpreted based on a 5% significance level. Results: The study included 265 patients treated with CD19-targeting CAR-T cells (tisagenlecleucel: n=85, 32%; axicabtagene ciloleucel: n=161, 61%; brexucabtagene autoleucel: n=19, 7%). The median age at CAR-T cell therapy infusion was 64 ± 13 years (range 20-86), with 67% being females and 33% males. All-grade cytokine release syndrome (CRS) was observed in 218 patients (82%), with high-grade CRS recorded in 26 (10%). Any-grade ICANS was documented in 92 (34%) patients and 36 (14%) of them had high-grade ICANS. Median values of PLT, fibrinogen, PT, PTT, INR, and D-dimer were 67 vs. 123, 258 vs. 377, 14.3 vs. 13.7, 29 vs. 31, 1.14 vs. 1.09, and 2.02 vs. 0.97, respectively, for patients with ICANS compared to those without ICANS. Fibrinogen and PLT levels were significantly lower on days with high-grade ICANS (grades 3-4) relative to other days [fibrinogen: 335 vs. 375 mg/dL (p = 0.003); PLT: 95 vs. 133 x 109/L, (p = 1.8 x 10-11)]. This was also true when comparing PLT counts measured on days with ICANS of any grade to those recorded on other days, with significant differences observed in PLT counts (90 vs. 136 x 109/L, (p = 9.87 x 10-31). In more than 75% of high-grade cases, PLT values were below 100 x 109/L on the day before the ICANS onset. Additionally, a reduction in PLT values was found to correlate to the high-grade ICANS onset on the following day, with mean reduction of 10% on the pre-ICANS day, while a mean change on other days amounted to +0.4% (p=0.014). Conclusions: Lower PLT and fibrinogen levels were consistently associated with ICANS development. The PLT count below 100 x 109/L preceded the ICANS onset in over 75% of high-grade cases. Overall, these findings may indicate a potential predictive value of PLT count reduction for ICANS onset.