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

To validate the end-to-end algorithm, separate expert clinicians were used to annotate a test set of 1774 recordings collected from 373 patients through an ongoing multi-site study. For the binary Good Signal / Poor Signal classification step, the algorithm output was compared to the annotations of signal quality qualitatively. For the final murmur detection step, algorithm output is compared to annotations of systolic and diastolic murmur presence using the primary analysis measures of sensitivity and specificity. As secondary analyses, sensitivity and specificity were calculated for PCG stratified by recording position and device, while sensitivity was calculated for PCGs with murmurs stratified by grade.