Immune checkpoint inhibitor (ICI) toxicity in a large prospective cohort of patients with solid tumors: The I-CHECKIT study (SWOG S2013).

2640 Background: Based on clinical trials, the expected rate of ≥ grade 3 iRAEs is ~15% for single agent ICIs and ~30-50% for ipilumimab (ipi) plus nivolumab (nivo). Predictors of irAEs remain poorly defined, limiting clinicians’ ability to anticipate and manage irAEs especially in patients treated in community practices. We conducted a prospective study to quantify the incidence and severity of irAEs across a large, diverse population of patients receiving ICI therapy and to identify clinical predictors of irAE development. Methods: S2013 enrolled adult patients planning to receive standard of care ICI-based therapy in two separate cohorts; only results from the completed first cohort, which received either single drug or combination (combo) ICI therapy, are reported here. Eligibility criteria were few and study included patients with active autoimmune disease, decreased performance status, and any stage of cancer. No chemo, biological or targeted therapy were permitted. Patients were followed for 1 year for the occurrence of irAE. The primary objective was to assess the occurrence of Grade >3 non-hematologic irAEs per CTCAE v5. Evaluable patients had 1 or more toxicity assessment. irAE events were centrally reviewed by the study team. Results: 2084 patients were enrolled and N=2,020 patients were eligible (96.9%). Mean (SD) age was 68.7 (12.5) years, 35.8% were female, 5.2% Black, and 8.0% Hispanic. Most patients received single drug (1,684; 83.4%), while 336 (16.6%) received combo. Majority of single drug patients received pembrolizumab (50.6%), followed by nivo (29.6%), durvalumab (14.0%), and ipi (13.8%). Combinations were mostly ipi plus nivo. Within the first year, 269 (13.3%) patients experienced Grade ≥3 non-hematologic irAE. The most common irAEs were hepatitis, 21.2%; gastrointestinal disorders, 19.3%; and respiratory, 10.0% (Table). IrAE incidence was higher in patients receiving combinations (27.4%) compared to single drug (10.5%; Chi-square p<.001). Conclusions: In this large prospective cohort with unrestrictive enrollment criteria, irAE incidence was similar to that seen in more restrictive clinical trial populations. Analysis of the clinical predictors of irAEs is in progress. Grant: NIH/NCI UG1CA189974. Clinical trial information: NCT04871542 . Grade ≥3 non-hematologic irAEs, by category. irAE n % irAE n % Hepatitis 57 21.2 Metabolism/nutrition 19 7.1 Gastrointestinal 52 19.3 Musculoskeletal 12 4.5 Skin 32 11.9 Cardiac 9 3.3 Respiratory 27 10.0 Hepatobiliary 7 2.6 Endocrine 22 8.2 Nervous system 7 2.6