AI predicts cancer patient survival just by reading physician notes

A team of researchers from the University of British Columbia and BC Cancer have created a ground-breaking AI model that can inform oncologists of individual cancer patient survival rates—with greater than 80% accuracy—all from the analysis of the initial consultation document. Lead author Dr. John-Jose Nunez, and his team, including researchers from BC Cancer and UBC's departments of Computer Science and Psychiatry, created a model that can more accurately predict an individual patient's survival rate by taking into account unique clues within their initial consultation document, as opposed to the traditional method of categorizing cancer patient survival rates by generic factors such as cancer site and tissue type. The model is applicable to all cancers, unlike past models that were limited to certain types.It uses natural language processing, a branch of AI that understands complex human language, to analyze oncologist notes after a patient's initial consultation visit. This analysis was demonstrated to predict six-month, 36-month, and 60-month survival with an accuracy of more than 80%. The findings were recently published in JAMA Network Open. In this prognostic study of 47,625 patients with cancer, the researchers trained and evaluated traditional and neural models to predict whether patients would survive six, 36, or 60 months based on their initial oncologist consultation. They evaluated their model on an internal holdout set of data, and found the best models achieved accuracy and BAC above 0.800 and AUC above 0.900. This performance was comparable with or superior to that of prior work, the authors noted, which has predicted survival only for specific types of cancer or used data that are more difficult to obtain. New possibilities “Our results suggest it is possible to predict the survival of patients with cancer without having to construct structured data sets, or limiting the predictions to specific types or locations of cancer,” the authors wrote. “The availability of structured data may vary. Given the widespread availability of initial oncologist consultation documents, this opens up the possibility of more easily training and using such models across cancer types at different cancer centres.” “The AI essentially reads the consultation document similar to how a human would read it,” said Dr. Nunez, a psychiatrist and clinical research fellow with the UBC Mood Disorders Centre and BC Cancer, in a UBC media release. “These documents have many details like the age of the patient, the type of cancer, underlying health conditions, past substance use, and family histories. The AI brings all of this together to paint a more complete picture of patient outcomes.”

AI predicts cancer patient survival just by reading physician notes

Register for free access or login.

Related Topics