Tumour genotypes account for survival differences in right‐ and left‐sided colon cancers. (9th February 2022)

• Record Type: Journal Article
• Authors: Ward, Thomas M.; Cauley, Christy E.; Stafford, Caitlin E.; Goldstone, Robert N.; Bordeianou, Liliana G.; Kunitake, Hiroko; Berger, David L.; Ricciardi, Rocco
• Is Part Of: Colorectal disease. Volume 24:Number 5(2022)
• Page Start: 601
• Page End: 610
• Publication Date: 2022-02-09
• Subjects: colorectal cancer -- genetics -- sidedness -- survival; Colon (Anatomy) -- Diseases -- Periodicals; Rectum -- Diseases -- Periodicals; 616.34
• DOI: 10.1111/codi.16060
• Languages: English
• ISSNs: 1462-8910

Abstract

Abstract: Aim: We sought to identify genetic differences between right‐ and left‐sided colon cancers and using these differences explain lower survival in right‐sided cancers. Method: A retrospective review of patients diagnosed with colon cancer was performed using The Cancer Genome Atlas, a cancer genetics registry with patient and tumour data from 20 North American institutions. The primary outcome was 5‐year overall survival. Predictors for survival were identified using directed acyclic graphs and Cox proportional hazards models. Results: A total of 206 right‐ and 214 left‐sided colon cancer patients with 84 recorded deaths were identified. The frequency of mutated alleles differed significantly in 12 of 25 genes between right‐ and left‐sided tumours. Right‐sided tumours had worse survival with a hazard ratio of 1.71 (95% confidence interval 1.10–2.64, P = 0.017). The total effect of the genetic loci on survival showed five genes had a sizeable effect on survival: DNAH5, MUC16, NEB, SMAD4, and USH2A. Lasso‐penalized Cox regression selected 13 variables for the highest‐performing model, which included cancer stage, positive resection margin, and mutated alleles at nine genes: MUC16, USH2A, SMAD4, SYNE1, FLG, NEB, TTN, OBSCN, and DNAH5. Post‐selection inference demonstrated that mutations in MUC16 ( P = 0.01) and DNAH5 ( P = 0.02) were particularly predictive of 5‐year overall survival. Conclusions: Our study showed that genetic mutations may explain survival differences between tumour sites. Further studies on larger patient populations may identify other genes, which could form the foundation for more precise prognostication and treatment decisions beyond current rudimentary TNM staging.