Real-world analyses to evaluate the role of TIGIT as a target in first-line (1L) gastric, gastroesophageal junction, and esophageal adenocarcinomas (GC/GEJC/EAC).

Abstract

4033 Background: Anti-programmed death protein 1 (PD-1) immune checkpoint inhibitors (ICIs) are approved for GC/GEJC/EAC.T-cell immunoreceptor with Ig and ITIM domains (TIGIT) is a potential anti-tumor ICI target.Understanding how PD-L1 ( CD274 ), effector T-cell (Teff), and TIGIT RNA expression levels correlate with real-world outcomes of 1L treatment can inform the potential for TIGIT as a target. Methods: Deidentified real-world data for patients (pts) with metastatic GC/GEJC/EAC were included from the Tempus AI, Inc. clinicogenomic database. Data were analyzed for pts with available longitudinal clinical data and biopsies (whole transcriptome RNA-sequencing [seq], 648 gene panel DNA-seq, and PD-L1 immunohistochemistry). Gene expression comparisons, including between TIGIT and a Teff gene set (geometric mean of CD8A , GZMA , GZMB , IFNG , EOMES , and PRF1 ), used the Wilcoxon test, Kruskal-Wallis test, and Spearman correlation. Real-world time to next treatment or death (rwTTNTD; maximum follow-up 24 months) was assessed for 1L ICI + chemotherapy (chemo) or chemo using the Kaplan-Meier method and compared by expression level (high [≥ median] vs low [ < median]) of TIGIT , Teff gene set, and TIGIT normalized by Teff gene set. Results: Among 545 pts (1L ICI + chemo, n = 50; 1L chemo, n = 124) TIGIT expression was most highly correlated with Teff ( R = 0.80, P < 0.001) and FOXP3 expression ( R = 0.78, P < 0.001), followed by CD274 expression ( R = 0.57, P < 0.001). Positive correlations with PD-L1 combined positive score were observed for TIGIT and Teff expression ( P < 0.001 for both) but not TIGIT normalized to Teff levels ( P > 0.05). Biopsies from liver metastases, which tend to show a poor response to ICIs, had lower immune signatures vs stomach tumor biopsies ( P < 0.001). High vs low expression of TIGIT and Teff were associated with numerically longer rwTTNTD for ICI + chemo, with HRs (95% CI) of 0.82 (0.43–1.54) for TIGIT (n = 29 vs 21) and 0.78 (0.42–1.48) for Teff (n = 28 vs 22). Corresponding HRs (95% CI) for chemo were 1.04 (0.70–1.54) for TIGIT (n = 53 vs 71) and 1.26 (0.85–1.87) for Teff (n = 64 vs 60). When TIGIT expression was normalized to Teff levels, HRs (95% CI) for rwTTNTD for high vs low expression were 0.92 (0.49–1.71) for ICI + chemo (n = 23 vs 27) and 0.81 (0.54–1.20) for chemo (n = 57 vs 67). Conclusions: TIGIT expression was highly correlated with FOXP3 and Teff gene set expression in GC/GEJC/EAC tumors. The improved rwTTNTD seen in pts with high TIGIT expression may be driven by these patients also having high Teff expression. When TIGIT is normalized to Teff, pts with high TIGIT expression lose this added benefit. Combining anti-TIGIT and anti–PD-1 treatments may lead to enhanced T cell activation, which could bring benefit to pts with 1L GC/GEJC/EAC. Larger studies will help confirm these findings.