Diagnosis of aortic valve stenosis by correlation analysis of wavelet filtered heart sounds

Traditional auscultation performed by general practitioners (GP) remains problematic and often gives significant results only in a late stage of the heart valve disease. Valve stenoses and insufficiencies are nowadays diagnosed with accurate but expensive ultrasonic devices. This study is aimed to develop a new heart sound analysis method for diagnosing aortic valve stenoses (AVS) based on a wavelet and correlation technique approach. Heart sounds recorded from 81 patients (43 AVS patients, 11 healthy controls-REF and 27 patients with other valve diseases-OVD) with an electronic stethoscope were wavelet filtered and envelopes were calculated. Three correlations on the basis of these envelopes were performed: (1) within AVS, (2) within mixed group AVS+REF and (3) within mixed group AVS+OVD leading to the mean correlation coefficients r_AVS, T_AVSREF/ and r_AVSOVD/. The results show that r_AVS is significantly higher than r_AVSREF/ and r_AVSOVD/ (p<0.01). These differences in sound patterns between the groups show that the introduced method provides an automatic diagnostics of AVS. We developed a new method of heart sound classification that seems to be suited for diagnosing valve diseases and application in a low cost and easy to use system for GPs.