Watson for Ontology (WfO) is a Clinical Decision Support System that provides a suggestion of treatment to clinicians. However, WfO shows a low concordance rate with clinicians to metastatic colorectal cancer because it has a limitation that Watson does not reflect the patient's actual EMR data. In order to improve the decision support capability of WfO, we analyzed the prognostic factors affecting the survival period of metastatic colorectal cancer patients based on EMR data. For analyzing the prognostic factors, Kaplan-Meier and Cox proportional-hazard models were used. As a result, we demonstrated that age, positive rate from harvested lymph nodes, post-operation and chemotherapy were significant factors of prognostic factors for overall survival.

Learning Objective: ● Understand the limitation that Watson for Oncology has and learn the possible solution to overcome.
● Learn the process to extract prognostic factors affecting the survival period of certain disease based on EMR data.


Hyunjin Kwon (Presenter)
Gachon University

Youngho Lee, Gachon University
Jinhyeok Park, Gachon University
Sunjin Sym, Gil Medical Center
Jeong-Heum Baek, Gil Medical Center

Presentation Materials: