Co-mutation-based subtyping of Chinese patients with gastric cancer.

person He Huang The First Hospital of Shanxi Medical University, Shanxi, China info_outline He Huang, Zhimin Wang, Hao Chen, Yafei Cheng, Hui Liu, Shaoxiong Bai, Yaoliang Huo, Yanwu Zeng, Leilei Lu

Authors person He Huang The First Hospital of Shanxi Medical University, Shanxi, China info_outline He Huang, Zhimin Wang, Hao Chen, Yafei Cheng, Hui Liu, Shaoxiong Bai, Yaoliang Huo, Yanwu Zeng, Leilei Lu Organizations The First Hospital of Shanxi Medical University, Shanxi, China, Shanghai OrigiMed Co., Ltd, Shanghai, China Abstract Disclosures Research Funding No funding received None. Background: The incidence of gastric cancer is decreasing globally, in contrast, that of Chinese patients has been increased from 35.9% to 48.3%. In-depth investigation of the characteristic of Chinese gastric cancer patients could help to understand the disease mechanism and improve the treatment of gastric cancer. Methods: Tumor tissues were obtained from 96 Chinese gastric cancer patients and tested by high-coverage ( > 1000x) panel sequencing containing 624 cancer-related genes. Genomic mutations including substitutions, insertions, deletions, rearrangements, fusions, and copy number variations were called. Genes were clustered based on their co-mutation relationship in all samples. Patients with gene mutations in each cluster were classified into corresponding groups. The properties of each patient group were characterized, including genomic features, age, gender, stage, LAUREN type, and grade. Results: The most mutated genes were TP53 , ERBB4 , LRP1B , and CDH1 in our cohort. A higher proportion of female patients had TP53 mutations compared to male patients. Mutations in ALK , KMT2C , and LRP2 were positively associated with microsatellite-instable, and that in ERBB4 and PIK3CA were positively associated with high tumor mutation burden (TMB). Four gene clusters (1, 2, 3, and 4) were identified based on the co-mutation relationship between genes. Cluster 1 included FAT4 , APC , OBSCN , and GATA3 ; Cluster 2 included FGF3 , CCND1 , FGF19 , and GATA4 ; Cluster 3 included BLK , FAM135B , and DLC1 ; and Cluster 4 included ATRX , KMT2D , STAG2 , and EGFR . Patients were classified into the four patient groups (I, II, III, and IV) accordingly when mutations in the corresponding gene cluster were found. Further association analysis indicated that: Group I was associated with adenocarcinoma subtype; Group II was negatively associated with LAUREN type 2; Group III was associated with high TMB; Cluster IV was associated with microsatellite-instable. Conclusions: This study utilized gene co-mutations to classify gastric cancer patients into subgroups with unique clinical characteristics. Further study on the etiology, treatment, and prognosis of these groups of patients is needed.