Acoustic Detection of Coronary Occlusions

Acoustic Detection of Coronary Occlusions before and after Stent Placement Using an

Electronic Stethoscope

Andrei Dragomir 1, Allison Post 1, Yasemin M. Akay 1, Hani Jneid 2,3, David Paniagua 2,3,Ali Denktas 2,3, Biykem Bozkurt 2,3 and Metin Akay 1,*1 Department of Biomedical Engineering, University of Houston, Houston, TX 77204, USA;

Andrei.Drag@gmail.com (A.D.); Allison.Post@central.uh.edu (A.P.); ymakay@uh.edu (Y.M.A.)2 Winters Center for Heart Failure Research, The Michael E. DeBakey VA Medical Center, Houston, TX 77030,USA; jneid@bcm.edu (H.J.); dpaniagua@bcm.edu (D.P.); ali.denktas@bcm.edu (A.D.);bbozkurt@bcm.edu (B.B.)

3 Cardiovascular Research Institute, Baylor College of Medicine, Houston, TX 77030, USA

* Correspondence: makay@uh.edu; Tel.: +1-832-842-8860

Academic Editors: Raúl Alcaraz Martínez and Kevin H. Knuth

Received: 27 April 2016; Accepted: 23 July 2016; Published: 29 July 2016

Abstract: More than 370,000 Americans die every year from coronary artery disease (CAD).Early detection and treatment are crucial to reducing this number. Current diagnostic and disease-monitoring methods are invasive, costly, and time-consuming. Using an electronic stethoscope and spectral and nonlinear dynamics analysis of the recorded heart sound, we investigated the acoustic signature of CAD in subjects with only a single coronary occlusion before and after stent placement, as well as subjects with clinically normal coronary arteries. The CAD signature was evaluated by estimating power ratios of the total power above 150 Hz over the total power below 150 Hz of the FFT of the acoustic signal. Additionally, approximate entropy values were

estimated to assess the differences induced by the stent placement procedure to the acoustic signature of the signals in the time domain. The groups were identified with this method with 82% sensitivity and 64% specificity (using the power ratio method) and 82% sensitivity and 55% specificity (using the approximate entropy). Power ratios and approximate entropy values after stent placement are not statistically different from those estimated from subjects with no coronary occlusions. Our approach

demonstrates that the effect of stent placement on coronary occlusions can be monitored using an electronic stethoscope.