Incidence and characterization of neurological adverse events among a retrospective cohort of patients receiving immune checkpoint inhibitors.

Abstract

2655 Background: The rising usage of immune checkpoint inhibitors (ICIs) has increased the incidence of rare adverse events in routine practice. Neurological adverse events (nAEs) are a heterogeneous group of outcomes that are infrequent and typically diagnosed by exclusion. They more often involve the peripheral (e.g., neuropathy, myasthenia gravis) than the central nervous system (e.g., encephalitis, hypophysitis). Evidence remains extremely limited with no large cohorts and incidence reported through small case series. This study aimed to analyze data from all clinically significant nAEs to characterize their risk factors, incidence, severity, and management. Methods: We created a retrospective registry of all patients who received at least one dose of an ICI for any indication between 2/1/2011 and 4/7/2022 at a comprehensive cancer center and its outreach clinics. Study personnel reviewed clinical documentation from all patients to collect data on treatment-emergent adverse events; attribution and severity were estimated using National Cancer Institute guidelines and the Common Terminology Criteria for Adverse Events. The investigators created a secure, cloud-based registry (REDCap), validated it with data quality rules, and resolved all discrepancies; clinical research specialists at Vasta Global captured most of the data. Categorical variables were compared using Fisher’s exact test. Statistical significance defined as p < 0.05. The study had institutional IRB approval. Results: In the total cohort of 3,137 patients, there were 47 nAEs (1.49%). The majority of these nAEs were peripheral (31, 0.99%; grade 3/4 in 14 patients, 0.45%) as opposed to central (16, 0.50%; grade 3/4 in 8, 0.25%). Pre-existing neurological comorbidities were present in five patients who had central nAEs (stroke = 3, dementia = 2) and in one patient with a peripheral nAE (stroke = 1). All high-grade (3-4) central nAEs occurred among renal (4, 50%), melanoma (3, 38%), or lung (1, 12.5%) primaries which was statistically significant when compared to other grade 3/4 adverse events (p < 0.0001) or no adverse events (p < 0.0001). Median days from first ICI to nAE was 55 IQR [22-175] and average treatment with prednisone equivalent was 2.5 mg/kg, std 4.6. Conclusions: The incidence of nAEs (1-2%) appeared higher than prior estimates in this retrospective patient registry which included patients who did not have neurology consultation or specific billing codes. All high-grade nAEs occurred in a tumor type known to favor metastases to the brain. With similar patterns among low grade events, an underlying mechanism of molecular mimicry seems more likely. Both time to presentation and treatment remained highly heterogenous. Further research is needed on treatment-emergent nAEs to determine if they are potentiated by neurologic comorbidities or subclinical brain metastases.