Type, timing, and patient characteristics associated with immune-related adverse event development in patients with advanced solid tumors treated with immune checkpoint inhibitors.

Abstract

e15160 Background: Development (DV) of immune-related adverse events (IRAEs) to immune checkpoint inhibitor (ICI) have been associated (asso/) with (w/) favorable treatment (tx) response, but heterogeneity in clinical presentation remains a concern. We aim to characterize the type, timing, and risk factors asso/ w/ IRAE DV in ICI-treated solid tumor (ST) patients (pts). Methods: The characteristics (char), tx course, and clinical outcomes of ST pts who received at least 2 ICI doses at our institution from 1/1/2011 to 4/29/2017 were reviewed. IRAEs were identified and graded (GR) based on CTCAE (v.4.0). Fisher’s exact test and multivariable regression analyses (MVA) were performed to study the differences between pts w/ versus without (w/o) IRAE. Results: Of the 344 pts identified, 111 pts (32%) had IRAE(s) w/ 145 total events (GR 1: 31%, GR 2: 57%, GR 3: 12%), of which 45% required systemic steroids (SS), 27% lead to ICI discontinuation (DC), and 17% lead to hospitalization (HP, Table). Median time to any IRAE was 12 weeks (w, range 1-149.5w). The most common IRAEs were endocrine (30%), gastrointestinal (GI, 28%), and skin (24%) related. On univariable analysis, antibiotic use, melanoma, and combination ICI use were asso/ w/ IRAE DV; melanoma remained significant on MRA (odds ratio, OR: 3.58 [1.68, 7.64], p = 0.0010). Pts w/ IRAE had higher disease control rate (complete/partial response + stable disease) than pts w/o IRAE (OR: 2.20 [1.29, 3.75], p = 0.0038). Conclusions: Our real-world data showed higher IRAE incidence when compared to those reported in trials, although no GR 4/5 IRAEs occurred. Melanoma was asso/ w/ IRAE DV on MRA, possibly due to the type, dose, and combination of ICIs. These findings warrant further investigation to better identify those at the highest risk to develop clinically significant IRAE. [Table: see text]