Clinical Decision Support For High-Risk Stage Ii Colon Cancer: A Real-World Study Of Treatment Concordance And Survival

BACKGROUND: Prognostic and pathologic risk factors typically guide clinicians and patients in their choice of surveillance or adjuvant chemotherapy when managing high-risk stage II colon cancer. However, variations in treatment and outcomes in patients with stage II colon cancer remain.OBJECTIVE: This study aimed to assess the survival benefits of treatments concordant with suggested therapeutic options from Watson for Oncology, a clinical decision support system.DESIGN: This is a retrospective observational study of concordance between actual treatment and Watson for Oncology therapeutic options.SETTING: This study was conducted at a top-tier cancer center in China.PATIENTS: Postoperative treatment data were retrieved from the electronic health records of 306 patients with high-risk stage II colon adenocarcinoma.MAIN OUTCOME MEASURES: The primary outcomes measured were the treatment patterns plus 3- and 5-year overall and disease-free survival for concordant and nonconcordant cases.RESULTS: Overall concordance was 90%. Most nonconcordant care resulted from adjuvant chemotherapy use (rather than surveillance) in patients with high-level microsatellite instability and >= 70 years old. No difference in overall survival (p= 0.56) or disease-free survival (p= 0.19) was observed between concordance groups. Patients receiving adjuvant chemotherapy had significantly higher 5-year overall survival than those undergoing surveillance (94% vs 84%,p= 0.01).LIMITATIONS: This study was limited by the use of retrospective cases drawn from patients presenting for surgery, the lack of complete follow-up data for 58% of patients who could not be included in the analysis, and a survival analysis that assumes no unmeasured correlation between survival and censoring.CONCLUSIONS: Watson for Oncology produced therapeutic options highly concordant with human decisions at a top-tier cancer center in China. Treatment patterns suggest that Watson for Oncology may be able to guide clinicians to minimize overtreatment of patients with high-risk stage II colon cancer with chemotherapy. Survival analyses suggest the need for further investigation to specifically assess the association between surveillance, single-agent and multiagent chemotherapy, and survival outcomes in this population.