The Molecular Landscape of Gastric Cancers for Novel Targeted Therapies from Real-World Genomic Profiling

Background: Panel-based comprehensive genomic profiling is used in clinical practice worldwide; however, large real-world datasets of patients with advanced gastric cancer are not well known. Objective: We investigated what differences exist in clinically relevant alterations for molecularly defined or age-stratified subgroups. Methods: This was a collaborative biomarker study of a real-world dataset from comprehensive genomic profiling testing (Foundation Medicine, Inc.). Hybrid capture was carried out on at least 324 cancer-related genes and select introns from 31 genes frequently rearranged in cancer. Overall, 4634 patients were available for analyses and were stratified by age (>= 40/< 40 years), microsatellite instability status, tumor mutational burden status (high 10 >= /low < 10 Muts/Mb), Epstein-Barr virus status, and select gene alterations. We analyzed the frequency of alterations with a chi-square test with Yate's correction. Results: Genes with frequent alterations included TP53 (60.1%), ARID1A (19.6%), CDKN2A (18.2%), KRAS (16.6%), and CDH1 (15.8%). Differences in comprehensive genomic profiling were observed according to molecularly defined or age-stratified subgroups. Druggable genomic alterations were detected in 31.4% of patients; ATM (4.4%), BRAF V600E (0.4%), BRCA1 (1.5%), BRCA2 (2.9%), ERBB2 amplification (9.2%), IDH1 (0.2%), KRAS G12C (0.7%), microsatellite instability-high (4.8%), NTRK1/2/3 fusion (0.13%), PIK3CA mutation (11.4%), and tumor mutational burden-high (9.4%). CDH1 alterations and MET amplification were significantly more frequent in patients aged < 40 years (27.7 and 6.2%) than in those aged >= 40 years (14.7 and 4.0%). Conclusions: Real-world datasets from clinical panel testing revealed the genomic landscape in gastric cancer by subgroup. These findings provide insights for the current therapeutic strategies and future development of treatments in gastric cancer.