Low frequency analysis of opening sound for detection of single leg separation of Björk-Shiley Convexo-Concave heart valves

Abstract

The aim of the study was to evaluate the potential of low frequency analysis of valve opening sounds in order to detect the presence of single leg separation of the outlet strut of Björk-Shiley convexo-concave (BSCC) valves. Single leg separation is believed to precede outlet strut fracture. Phonocardiograms (PCG) of 28 patients with BSCC mitral valves were recorded and filtered to limit the frequency bandwidth between 90 and 1400 Hz (-20 dB). Twenty-four patients had BSCC valves that were presumably unimpaired and four patients had BSCC valves with single leg separation. Spectral analysis consisted of computing the power spectrum for 30 opening sounds for each patient using a fast Fourier transform (FFT) algorithm on an IBM-PC compatible computer. For each patient, the power spectra of the selected opening sounds were averaged and normalized in amplitude on a 0-100% linear scale. A ratio of the high to the low frequency area of the averaged power spectra was computed and evaluated as a diagnostic feature to identify the valves with single leg separation. Six cut-off frequencies serving to distinguish between high and low frequency areas of the spectra were tested: 400, 450, 500, 550, 600, and 650 Hz. The results showed that the high-to-low frequency area ratios tested were all able to identify valves with single leg separation. The correct classification performance was 96.4%. The sensitivity to detect valves with single leg separation was 75% and the specificity was 100%. The positive predictive value was 100% and the negative predictive value 96%. The study confirmed the potential of low frequency analysis of valve opening sounds to detect single leg separation of the outlet strut of BSCC valves implanted in humans.