e14553 Background: Cutaneous immune-related adverse events (cirAEs) from immune checkpoint inhibitor (ICI) therapy are increasing as these drugs are more widely used. Population-level studies are lacking. In this retrospective cohort study, we analyzed the primary tumor’s effect on cirAE incidence and downstream utilization of systemic immunosuppression in a national healthcare database (TriNetX). Methods: Index event was defined as day of ICI initiation; outcomes were restricted to 2 years from index. cirAEs were defined as 42 dermatoses identified in a comprehensive review of the cirAE literature (Sibaud 2018) and expert opinion. Systemic immunosuppression was classified as steroidal or non-steroidal. Primary outcomes included aggregate and cancer-specific incidence of cirAEs across the four most common cancer indications for ICI therapy (melanoma, lung, urinary, and gastrointestinal). Secondary outcomes included utilization of new steroidal or non-steroidal systemic immunosuppression. For each analysis, we excluded patients with an outcome prior to the index event. Risk ratios (RR) were calculated after 1-to-1 propensity score matching, adjusting for age at index, sex, race, ethnicity, and ICI target, with the lung cancer group as reference. Results: We identified 27,481 eligible subjects. Aggregate incidence of cirAEs across all cancer types was 23.41%, with non-specific rashes, pruritus, and drug eruptions as the most common diagnoses. Among all patients, new steroidal and non-steroidal immunosuppression use following ICI initiation was 16.1% and 4.1%, respectively. After adjusting for covariates, melanoma (RR 1.60, 95% CI 1.43-1.79, p < 0.001) was associated with a higher risk of cirAE, while urinary and gastrointestinal tract cancers were not significantly different from the reference. Melanoma (RR 0.64, 95% CI 0.55-0.74), urinary tract (RR 0.71, 95% CI 0.62-0.81) and gastrointestinal tract (RR 0.46, 95% CI 0.39-0.53) cancers were all less likely to require steroidal immunosuppression (all p < 0.001). However, patients with urinary tract cancer (RR 4.03, 95% CI 3.04-5.35) and melanoma (RR 2.44, 95% CI 1.75-3.40) were more likely to receive non-steroidal systemic immunosuppression (both p < 0.001). Conclusions: Our results reinforce that ICI recipients commonly develop skin toxicities and provide robust evidence that melanoma is independently associated with their incidence. Furthermore, we found that patients with lung cancers received steroidal systemic immunosuppression most frequently, an intervention associated with poorer overall survival (Riudavets 2020). In contrast, patients with melanoma were more likely to receive non-steroidal immunosuppression. This study is the first population-based analysis of an irAE across tumor types and is proof-of-concept for future investigations into clinical risk factors in a large dataset.
