Shonan Sho1,2, Colin M Court1,2, Paul Winograd1,2, Marcia M Russell1,2, James S Tomlinson1,2,3. 1. Department of Surgery, David Geffen School of Medicine, University of California Los Angeles, Los Angeles, California. 2. Department of Surgery, VA Greater Los Angeles Healthcare System, Los Angeles, California. 3. UCLA Center for Pancreatic Diseases, Los Angeles, California.
Abstract
BACKGROUND AND OBJECTIVES: Approximately 20-40% of stage II/III colorectal cancer (CRC) patients develop relapse. Clinicopathological factors alone are limited in detecting these patients, resulting in potential under/over-treatment. We sought to identify a prognostic tumor mutational profile that could predict CRC recurrence. METHODS: Whole-exome sequencing data were obtained for 207 patients with stage II/III CRC from The Cancer Genome Atlas. Mutational landscape in relapse-free versus relapsed cohort was compared using Fisher's exact test, followed by multivariate Cox regression to identify genes associated with cancer recurrence. Bootstrap-validation was used to examine internal/external validity. RESULTS: We identified five prognostic genes (APAF1, DIAPH2, NTNG1, USP7, and VAV2), which were combined to form a prognostic mutation panel. Patients with ≥1 mutation(s) within this five-gene panel had worse prognosis (3-yr relapse-free survival [RFS]: 53.0%), compared to patients with no mutation (3-yr RFS: 84.3%). In multivariate analysis, the five-gene panel remained prognostic for cancer recurrence independent of stage and high-risk features (hazard ratio 3.63, 95%CI [1.93-6.83], P < 0.0001). Furthermore, its prognostic accuracy was superior to the American Joint Commission on Cancer classification (concordance-index: 0.70 vs 0.54). CONCLUSIONS: Our proposed mutation panel identifies CRC patients at high-risk for recurrence, which may help guide adjuvant therapy and post-operative surveillance protocols.
BACKGROUND AND OBJECTIVES: Approximately 20-40% of stage II/III colorectal cancer (CRC) patients develop relapse. Clinicopathological factors alone are limited in detecting these patients, resulting in potential under/over-treatment. We sought to identify a prognostic tumor mutational profile that could predict CRC recurrence. METHODS: Whole-exome sequencing data were obtained for 207 patients with stage II/III CRC from The Cancer Genome Atlas. Mutational landscape in relapse-free versus relapsed cohort was compared using Fisher's exact test, followed by multivariate Cox regression to identify genes associated with cancer recurrence. Bootstrap-validation was used to examine internal/external validity. RESULTS: We identified five prognostic genes (APAF1, DIAPH2, NTNG1, USP7, and VAV2), which were combined to form a prognostic mutation panel. Patients with ≥1 mutation(s) within this five-gene panel had worse prognosis (3-yr relapse-free survival [RFS]: 53.0%), compared to patients with no mutation (3-yr RFS: 84.3%). In multivariate analysis, the five-gene panel remained prognostic for cancer recurrence independent of stage and high-risk features (hazard ratio 3.63, 95%CI [1.93-6.83], P < 0.0001). Furthermore, its prognostic accuracy was superior to the American Joint Commission on Cancer classification (concordance-index: 0.70 vs 0.54). CONCLUSIONS: Our proposed mutation panel identifies CRC patients at high-risk for recurrence, which may help guide adjuvant therapy and post-operative surveillance protocols.
Authors: Roshini Prakash; Sivan Izraely; Nikita S Thareja; Rex H Lee; Maya Rappaport; Riki Kawaguchi; Orit Sagi-Assif; Shlomit Ben-Menachem; Tsipi Meshel; Michal Machnicki; Shuichi Ohe; Dave S Hoon; Giovanni Coppola; Isaac P Witz; S Thomas Carmichael Journal: Front Neurosci Date: 2019-04-09 Impact factor: 4.677