The significance of the analysis on scalographic pattern for detecting malfunctioning bileaflet valve with the wavelet analysis.

ABSTRACT

The authors developed the wavelet analysis system which can detect the splitting of bileaflet mechanical heart valve (BLV) into two spikes on the scalogram, and reported that either consecutive single spike or the split behavior can detect malfunctioning BLV (MBV). The latest study on 12 BLVs suggested that the comparison between two spike areas showed higher potential to detect MBV than the split behavior. The aim of the current study is to review 226 files of BLV sound and to select the suitable scalographic property to differentiate the function of BLV with the split. Eight of 30 MBV files showed consecutive single spike, and the rest of 22 MBV files showed two spikes. Two spike areas can be compared by the following three ratios; the anterior spike area/posterior spike area (Aa/La), its reverse ratio (Pa/Aa) and the smaller spike area/the larger spike ratio (Sa/La). Therefore, the current study compared those three ratios to pursue the suitable ratio to compare two spike areas and its sensitivity to differentiate valve function by the ROC analysis. As a result, the Sa/La was suitable for comparing two spike areas, and only this ratio showed high accuracy to differentiate the function of BVL with the split, and its cutoff value was <0.665. Conclusively, the key for detecting MBV was either consecutive single spike or the mean of Sa/La < 0.665. However, this cutoff point is still tentative due to small number of malfunctioning valves, and other key might be available in future.