Preinfusion Risk Factors for the Development of Severe Immune Effector Cell Associated Neurotoxicity Syndrome Following Chimeric Antigen Receptor T-Cell Infusion in Pediatric Patients

Introduction: Immune Effector Cell Associated Neurotoxicity Syndrome (ICANS) is a common and challenging toxicity associated with CART for B-cell acute lymphoblastic leukemia (B-ALL). Although cytokine release syndrome (CRS) has been associated with ICANS, the only preinfusion risk factor consistently identified across clinical trials is bone marrow disease burden (DB). To improve prognostication and risk stratification, we sought to identify preinfusion clinical risk factors and biomarkers associated with the development of severe ICANS.

Methods: We conducted a retrospective cohort study of all patients who received investigational or commercial CD19- or CD22-directed, 4-1BB-based CART for B-ALL at our tertiary care institution between April 2012 - June 2023. The primary outcome was development of severe ICANS, defined as grade 3 or higher neurological adverse events based on CTCAE categories of encephalopathy, speech impairment, movement disorder, vision changes, cranial nerve disorder or seizure, or cerebral edema. Grading was performed prospectively for clinical trial patients and retrospectively for commercial tisagenlecleucel (tisa-cel) patients.

We hypothesized the following baseline risk factors for severe ICANS: DB, age, CNS disease status, history of transient neurologic dysfunction (methotrexate toxicity, seizure, posterior reversible encephalopathy syndrome (PRES)) or structural injury (intracranial hemorrhage, stroke). DB was dichotomized as high (marrow blasts 25%) or low.

Logistic regression models were used to calculate odds ratios (ORs) and their confidence intervals (CIs). Analyses were a priori stratified by DB.

To identify potential biomarkers, pre-infusion peripheral blood serum was examined using a proximity extension assay (Olink, Boston, USA); 192 proteins were measured using the Immune Response and Immuno-Oncology panels. Normalized protein expression (NPX), a log-2-fold transformed score was compared between patients who did and did not develop severe ICANS or seizure using differential expression analysis.

Results: The study population included 442 patients treated across 8 clinical trials or with commercial tisa-cel. Comparable with previously reported estimates, 187/442 (42%) patients developed any ICANS, and 37/442 (8.4%) developed severe ICANS. Encephalopathy was the most common phenotype in 144 (33%) patients. Seizure was reported in 21 (4.8%), and cerebral edema occurred in 2 (0.4%).

In univariate analysis neurologic exposures were considered individually. Only high DB (OR 6.46, p<0.001, 95% CI 3.1,14.1) and history of PRES (OR 6.23, p<0.012, 95% CI 1.27, 24.9) were associated with a higher risk of severe ICANS. In multivariate analysis neurologic exposures were combined as transient or structural. Among low DB patients (n=311), no factors were significantly associated with severe ICANS. However, among high DB patients (n=102), history of a transient neurologic dysfunction was associated with a higher risk of developing severe ICANS (OR 4.56, p=0.030, 95% CI 1.4,14.8).

At the pre-infusion timepoint, no serum biomarkers were associated with a higher risk of developing severe ICANS or seizure. Eight proteins (PRDX5, DCTN1, HEXIM1, PPP1R9B, EIF4G1, MGMT, DDX58, HCLS1) were associated with a lower risk of developing severe ICANS or seizure. Analyses were then stratified by DB. In high DB patients, higher levels of the protein CD27 were associated with the development of severe ICANS or seizure (p=0.03). In low DB patients, no proteins were associated with a higher risk of severe ICANS or seizure.

Conclusions: We present the largest study to date of ICANS across multiple clinical trials of CART for children and young adults with B-ALL. We demonstrate that ICANS is common, occurring in 42% of patients, with 8% experiencing severe ICANS. The most significant risk factor for ICANS is DB. Amongst patients with high DB, previous history of a transient neurologic event further increased the risk of severe ICANS. In patients with low DB, none of the studied exposures, including prior neurological dysfunction/injury or CNS disease status, increased the severe ICANS risk. We also identify CD27 as a potential predictive biomarker in patients with high DB. These results have important implications for toxicity prediction and for the design of toxicity mitigation trials in pediatric patients receiving CART.