Neuroimaging Findings during Immune Effector Cell Associated Neurotoxicity Syndrome (ICANS) - a CAR T Cell Neurotoxicity Imaging Virtual Archive (CARNIVAL) Study

Introduction Neuroimaging findings in immune effector cell associated neurotoxicity syndrome (ICANS) have not been systematically described. To address this gap in knowledge, we created a retrospective database (CARNIVAL) of neuroimaging studies obtained in patients with hematologic malignancies who received CAR T cell therapy (CAR T) at 6 pediatric institutions within the CARNATION consortium. Objectives The primary objective of this study was to characterize the range of neuroimaging findings that are associated with acute ICANS. The secondary objective was to determine whether particular neurologic symptoms are associated with specific imaging findings. Methods Neuroimaging studies were uploaded to a cloud-based platform (AMBRA). Patient data were captured in RedCAP. All patients with hematologic malignancies who had brain MRIs in the acute observation period post CAR T (day 1-28) were included in this analysis. Brain MRIs obtained at baseline (most proximal pre-CAR T study), acute, and follow up (any time >28 days post CAR T) studies were reviewed by two pediatric neuroradiologists. Imaging features were categorized and correlated with CAR product, ICANS grade and core symptoms, non-canonical ICANS symptoms (such as headache, tremor, ataxia), central nervous system (CNS) involvement, CNS radiation, and neurologic comorbidities. Imaging findings were categorized semi-quantitatively to define ICANS-specific patterns. Results Acute MRIs were available for 84 patients treated from 2014-2022. 41 of these patients had baseline MRIs and 14 had follow up MRIs. In patients with ICANS, the acute MRI showed ICANS-related abnormalities in 45%, while 28% had non-ICANS related abnormalities and 27% were normal. The acute ICANS-related abnormalities resolved completely in 50% of the patients who had follow up imaging. ICANS grade strongly predicted the likelihood of acute abnormalities on brain MRI, ranging from 7% for patients with no ICANS, to 42% for grade 2, and 100% for grade 4 (P<0.001). Acute MRI abnormalities were not associated with patient age, CRS grade, prior radiotherapy, or neurologic comorbidities. ICANS-specific and non-specific imaging pattern incidence, reversibility, and association with ICANS grade, CAR product and patient characteristics are quantified. Conclusion CARNIVAL provides a unique resource for studying the neuroimaging findings in CAR-T-associated toxicity and is open to additional institutions worldwide. CARNIVAL encompasses the largest dataset and provides the first comprehensive overview of neuroimaging findings in pediatric CAR-T patients and yields insights that can optimize patient care, resource utilization, and further research advances into pathophysiologic mechanisms of ICANS.