Noninvasive Vascular Blood Sound Monitoring Through Flexible Microphone

Abstract

An arteriovenous vascular access is used to provide hemodialysis to people suffering from acute or chronic kidney disease. Monitoring of vascular access is essential to reduce the risk of sudden clotting caused by vascular stenosis, to maintain patency of the access, and to avoid hospitalization or central catheter placement. Current imaging technologies are unsuitable for widespread monitoring of vascular access at a frequency greater than monthly and cannot detect rapidly progressing access dysfunction. Early detection can identify patients for imaging and enable surgical treatment planning. In this chapter, we review the signal processing methods of analyzing the sounds of blood flow (bruits) using phonoangiography. This method relies on acoustic features produced from turbulent blood flow. We discuss the custom signal processing techniques used for de-noising of the signal and feature extraction to reduce the dimensionality of wavelet-processed audio signals. To move toward a device suitable for use in dialysis centers, we developed a multichannel-flexible piezoelectric microphone to record bruits from multiple locations. This piezoelectric microphone in combination with the signal processing technique could localize stenosis within 1 cm of the actual location and differentiate degree of stenosis into three grades (mild, moderate, and severe) in vitro. These results were confirmed with statistical analysis and the performance of the feature was also tested with a threshold-based stenosis detection classifier to mimic a clinical screening scenario aimed at identifying at-risk patients.