Estimating Residual Stenosis for an Arteriovenous Shunt Using a Flexible Fuzzy Classifier

Abstract

This article determines whether the diameter of a vascular lumen is clinically valuable for hemodialysis therapy. A novel method is proposed that uses a flexible fuzzy classifier (FFC) implementing a curve approximation function to estimate residual stenosis in an arteriovenous shunt (AVS). The auscultation method provides a noninvasive technique for monitoring the degree of AVS stenoses. The Burg autoregressive method is used to identify the characteristic spectra for changes in frequency and amplitude between frequencies of 25 and 800 Hz. Fractional order self-synchronization error formulation is then used to quantify the degree of stenosis between normal access and AVS stenosis in the frequency spectra. To estimate the diameter of a vascular lumen, the FFC is used to create a nonlinear prediction model. For the 42 long-term patients studied, the results of this examination show that the proposed prediction model estimates residual stenosis efficiently.