Exploratory analysis of the impact of prior immune checkpoint inhibitor (ICI) on trastuzumab deruxtecan (T-DXd; DS-8201) clinical outcomes and biomarkers (BM) in DESTINY-Gastric01 (DG-01), a randomized, phase 2, multicenter, open-label study in patients (pts) with HER2+ advanced gastric or gastroesophageal junction adenocarcinoma.

Abstract

322 Background: T-DXd is a HER2-targeting antibody-drug conjugate. In primary analyses of DG-01 (NCT03329690), T-DXd showed clinically meaningful improvement in objective response rate (ORR) and overall survival (OS) vs chemotherapy in the primary cohort of pts with HER2+ advanced gastric cancer (AGC; IHC3+ or IHC2+/ISH+; Shitara K et al. NEJM 2020;382:2419). In the context of emerging first-line ICI-based regimens in AGC, we conducted a post hoc analysis of T-DXd clinical outcomes and baseline BMs in pts with or without prior ICI. Methods: DG-01 primary cohort enrolled pts who progressed on ≥2 prior lines with centrally confirmed HER2+ AGC. Efficacy and safety were assessed in pts with or without prior ICI (defined as any ICI-related treatment received prior to T-DXd). Exploratory BM analysis in the T-DXd arm assessed RNA sequencing (RNAseq) of tumor biopsies and ctDNA of liquid biopsies taken immediately before T-DXd treatment. A linear regression model and Wilcoxon rank-sum test were used to identify differentially expressed genes and signature scores between pts with or without prior ICI. Results: Prior ICI was received in 35.2% (44/125) of pts treated with T-DXd and 27.4% (17/62) treated with chemotherapy. The table shows efficacy in pts with or without prior ICI. Safety data in the T-DXd arm were similar between pts with or without prior ICI, including adjudicated drug-related interstitial lung disease/pneumonitis; 9.1% (4/44) of pts with prior ICI and 9.9% (8/81) of pts without prior ICI. Among 34 T-DXd-treated pts with RNAseq data, 12 received prior ICI. Searching for differentially expressing genes (fold change ≥2, P≤0.01) identified multiple genes with a higher trend for expression in tumors with prior ICI, including immune-oncology (IO)-related molecules, CXCL9 and PD-1 (log2 fold change 2.3 and 1.4, P=7.3e-05 and 2.8e-03, respectively). Furthermore, gene signature analysis showed a concurrent higher trend of multiple chemokine gene expressions in the prior ICI group. Conclusions: This exploratory analysis suggests that treatment with T-DXd results in better outcomes than chemotherapy regardless of prior ICI. Given a relatively small sample size, the relationship between changes in IO-related signatures and efficacy of T-DXd warrants further evaluation in larger cohorts. Clinical trial information: NCT03329690. [Table: see text]