A Deep Learning Algorithm for Automated Cardiac Murmur Detection Via a Digital Stethoscope Platform

John S Chorba; Avi M Shapiro; Le Le; John Maidens; John Prince; Steve Pham; Mia M Kanzawa; Daniel N Barbosa; Brent E White; Jason Paek; Sophie G Fuller; Grant W Stalker; Sara A Bravo; Dina Jean; Subramaniam Venkatraman; Patrick M McCarthy; James D Thomas

Deep learning approaches have shown impressive results towards problems in the medical field in recent years, utilizing imaging data such as radiologic studies 14 and echocardiograms 15 to develop interpretative algorithms. Interest in using deep learning for stethoscope sound analysis has also expanded in recent years, leading to applications in lung 16 and heart 17 sound classification. Indeed, an independently developed algorithm, focused on the binary distinction between pathology and normal, has tested favorably in a pediatric cohort, generating confidence in this approach 18 . Deep learning has also shown promise in utilizing auxiliary data unintended to be part of the original data set 19 , illustrating its ability to turn seemingly trivial data into useful information. Based on the traditional training of cardiologists to identify and triage valvular pathology from auscultatory characteristics, we hypothesized that a deep learning approach could be developed to perform similarly, if not better, than these specialty providers.