MP18-13 EARLY EXPERIENCE WITH WATSON FOR ONCOLOGY, A CLINICAL DECISION-SUPPORT SYSTEM FOR PROSTATE CANCER TREATMENT RECOMMENDATIONS

Abstract

INTRODUCTION AND OBJECTIVES: Urological oncologists have difficulties to provide optimal personalize care due to rapid alteration of scientific research results, medical development, and treatment guidelines. IBM Watson for Oncology (WFO) is an artificial intelligence clinical decision-support system that provides assistance to oncologists with evidence based treatment recommendations. In the present study, we examined the level of concordance between treatment recommendations according to WFO and actual patient received treatment in prostate cancer cases. METHODS: We enrolled 201 patients with prostate cancer who diagnosed and received initial treatment in the department of urology between Jan 2018 and June 2018. WFO provided treatment recommendations using clinical data in three categories: recommended, for consideration, and not recommended. Prostate cancer treatments evidence-based concordant if the actual prostate cancer treatment were included in 'recommended' and 'for consideration' by WFO. RESULTS: The overall prostate cancer treatment concordance was 73.6% (n=148) (Recommended 53.2%; n=107, For consideration 20.4%; n=41). Prostate cancer treatments were varied by age and stage. Especially, patients with 75 years and older age and stage IIA, IV groups were less likely to be less concordant than other stage and age groups. Logistic regression analysis revealed ECOG PS ≥1 and 75 years and older age were significantly less likely to be concordant (p=0.001 and p=0.026, respectively). CONCLUSIONS: The treatment recommendations by WFO and actual patient received treatment showed relatively high concordance rate in the prostate cancer patients. Lowering concordance rate of prostate cancer treatment was seemed to be affected by higher performance status, older age, and advanced cancer stage.