Conference or Workshop Item #18225

Abstract

Multichannel Intraluminal Impedance-pH (MII-pH) monitoring is designed to detect intraluminal bolus movement without the use of radiation and allows for detection of Gastroesophageal reflux (GER). Automatic analysis of MII-pH data are available however since the recordings are complex and filled with artifacts; a thorough and time-consuming review of the recordings, episode by episode, is still required. The proposed method was designed to segment GER events in a set of 100 episodes of two minutes interval of MII data based on a decision tree approach. An amount of 24 hours of MII-pH data belonging to eight patients were recorded, digitized and stored along with standardized timings of GER events that had been characterized by two gastroenterologist experts. The performance of the algorithm was evaluated using 100 individual GER events. The algorithm has been shown to perform correctly in over 95 of cases. © 2019 IEEE.