Wavelet analysis of valve closing sound detects malfunction of bileaflet mechanical valve

Abstract

Several studies have reported the asynchronous closure of normal bileaflet valves (NBVs), resulting in a split in its closing sound; however, the clinical significance of this split has never been studied in malfunctioning bileaflet valves (MBVs). The study comprised 218 valves in 184 patients, including normal monoleaflet valves (n = 10), NBVs (n = 198), and MBVs (n = 10). Valve function was confirmed by cinefluoroscopy prior to analysis of the valve sound by the Morlet continuous wavelet transform (CWT). The split interval (SI) for each heartbeat was measured, and the coefficient of variation (CV) of its mean (valve SI) was calculated as a parameter for the fluctuation of the SI. The CWT of monoleaflet valves showed a single spike, whereas NBVs exhibited a clear split. There was no significant difference in valve SI between the aortic and mitral positions; however, the mean of the CV was significantly greater in the mitral position (n = 90, 0.507 +/- 0.254) than in aortic position (n = 108, 0.353 +/- 0.228, P = 0.000045). The split was not found in six (aortic; three, mitral; three) of ten patients with MBVs. The other four patients had a distinct split, but the CV was significantly lower for MBVs (0.138 +/- 0.105) than for NBVs (0.343 +/- 0.221, P = 0.042). Receiver-operating characteristics analysis demonstrated the cutoff line of the CV to be 0.112 for detecting malfunctioning aortic valves with the highest accuracy of 86.1%. This new system using the Morlet CWT can detect MBVs. It will be a useful modality for screening the function of bileaflet valves.

Detection of Mechanical Heart Valve Thrombosis Using Support Vector Machine

“…They evaluated statistically the split time and split time change obtained from mechanical heart valve sounds with 184 normal bileaflet, 10 tilting disc and 10 dysfunction. Also they found the limit values for the split time that determine the dysfunction [17]. Zhang et al concluded that a combination of spectral and time-scale features of heart sounds recorded from paravalvular leakage, valve obstruction and normal patients can be used to detect these dysfunctions [18].…”