Choosing the optimal diagnostic test for patients with chest pain

In a report led by E2H co-founders, an international team of investigators used a novel machine learning approach to analyze the results of the landmark PROMISE trial. PROMISE aimed to identify the optimal diagnostic strategy for patients with symptoms suggestive of coronary artery disease. The team used a novel computational approach to learn a phenomapping representation of the trial population and learn signatures of personalized benefit with either anatomical or functional testing. Using artificial intelligence, an algorithm was trained to learn generalizable and reproducible features associated with relative benefit from either approach. Building on this approach, the team developed ASSIST, an externally validated tool that personalizes the selection of the optimal diagnostic approach for patients with suspected coronary disease.

The full results can be accessed at: https://academic.oup.com/eurheartj/article/42/26/2536/6242724?login=true 

The ASSIST tool is available at: https://www.cards-lab.org/assist 

Read more at: https://medicine.yale.edu/news-article/machine-learning-based-clinical-decision-support-tool-for-chest-pain/