Harvard, UPenn researchers use AI to predict mortality for cancer patients

A new algorithm may accurately predict whether cancer patients starting chemotherapy will still be alive after a month of treatment, according to a study published in JAMA.

A team of researchers from Boston-based Harvard Medical School and Philadelphia-based Perelman School of Medicine at the University of Pennsylvania developed the algorithm using machine learning, a type of artificial intelligence in which a computer learns over time as opposed to being programmed like typical software.

To train the algorithm, the researchers obtained EHR data from nearly 27,000 patients who received chemotherapy at the Dana-Farber/Brigham and Women's Cancer Center in Boston between 2004 and 2014. They also collected the patients' dates of death using Social Security data.

From there, the researchers validated the algorithm using information from a separate dataset of 9,114 patients, whose diagnoses included breast, colorectal and lung cancers.

The algorithm correctly predicted 30-day mortality for all the patients in the validation set, across primary cancers, stages and chemotherapies. The same algorithm also performed well when predicting 180-day mortality, according the study authors.

In their conclusion, the study authors noted that although the algorithm accurately predicted short-term mortality among patients starting chemotherapy during the study, further research is needed to determine the feasibility of applying it to the clinical setting.

More articles on artificial intelligence:
NIH to establish AI, machine learning committee
Nuance adds Mark Sherwood as CIO: 4 things to know
12 healthcare use cases for natural language processing

© Copyright ASC COMMUNICATIONS 2020. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.


Featured Webinars

Featured Whitepapers