Application of Wavelet Transform and Fractal Analysis for Esophageal pH-Metry to Determine a New Method to Diagnose Gastroesophageal Reflux Disease

Abstract

In this paper, a new method for analysing gastroesophageal reflux disease (GERD) is shown. This novel method uses wavelet transform (WT) and wavelet-based fractal analysis (WBFA) on esophageal pH-metry measurements. The esophageal pH-metry is an important diagnostic tool supporting the physician’s work in diagnosing some forms of reflux diseases. Interpreting the results of 24-h pH-metry monitoring is time-consuming, and the conclusions of such an analysis can sometimes be too subjective. There is no strict procedure or reference values to follow when the impedance measurements are assessed. Therefore, an attempt was made to develop a point of reference for the assessment process, helping to distinguish healthy patients from GERD patients. In this approach, wavelet transform (WT) and wavelet-based fractal analysis (WBFA) were used to aid the diagnostic process. With this approach, it was possible to develop two efficient computer methods to classify healthy and sick patients based on the pH measurement data alone. The WT method provided a sensitivity value of 93.33%, with 75% specificity. The results of the fractal analysis confirmed that the tested signals have features that enable their automatic classification and assignment to a group of sick or healthy people. The article will be interesting for those studying the application of wavelet and fractal analysis in biomedical waveforms. The authors included in the work a description of the implementation of the fractal and wavelet analysis, the descriptions of the results of the analyses, and the conclusions drawn from them. The work will also be of interest to those who study the methods of using machine learning and artificial intelligence in computer-aided, automatic medical diagnostics.