Artificial intelligence facilitates measuring reflux episodes and postreflux swallow‐induced peristaltic wave index from impedance‐pH studies in patients with reflux disease

Abstract

Background/Aim Reflux episodes and postreflux swallow‐induced peristaltic wave (PSPW) index are useful impedance parameters that can augment the diagnosis of gastroesophageal reflux disease (GERD). However, manual analysis of pH‐impedance tracings is time consuming, resulting in limited use of these novel impedance metrics. This study aims to evaluate whether a supervised learning artificial intelligence (AI) model is useful to identify reflux episodes and PSPW index. Methods Consecutive patients underwent 24‐h impedance‐pH monitoring were enrolled for analysis. Multiple AI and machine learning with a deep residual net model for image recognition were explored based on manual interpretation of reflux episodes and PSPW according to criteria from the Wingate Consensus. Intraclass correlation coefficients (ICCs) were used to measure the strength of inter‐rater agreement of data between manual and AI interpretations. Results We analyzed 106 eligible patients with 7939 impedance events, of whom 38 patients with pathological acid exposure time (AET) and 68 patients with physiological AET. On the manual interpretation, patients with pathological AET had more reflux episodes and lower PSPW index than those with physiological AET. Overall accuracy of AI identification for reflux episodes and PSPW achieved 87% and 82%, respectively. Inter‐rater agreements between AI and manual interpretations achieved excellent for individual numbers of reflux episodes and PSPW index (ICC = 0.965 and ICC = 0.921). Conclusions AI has the potential to accurately and efficiently measure impedance metrics including reflux episodes and PSPW index. AI can be a reliable adjunct for measuring novel impedance metrics for GERD in the near future.

Post-reflux swallow-induced peristaltic wave index: a new parameter for the identification of non-acid gastroesophageal reflux-related chronic cough

“…Research suggests that the application of artificial intelligence models may potentially change this current situation by enabling accurate measurements of reflux episodes and PSPWs. 58 However, further promotion and validation of these models are still required. In the future, we will further optimize the method of data collection and design a prospective study to further verify the conclusions of this study.…”

Development and Validation of a Machine Learning System to Identify Reflux Events in Esophageal 24-Hour pH/Impedance Studies

“…Wong et al recently demonstrated successful use of machine learning methods to identify reflux events in pH/impedance studies ( 10 ). They used ResNet-18, a convolutional neural network commonly used for image recognition, to identify reflux events from pH/impedance studies and reported a sensitivity of 84%.…”

“…Furthermore, existing commercial analysis software has not used modern machine learning techniques for development of algorithms or analysis tools to assist with interpretation of these studies. To date, few studies have reported using machine learning methods for interpretation of pH/impedance studies, which include recent studies using a convolutional neural network for study interpretation ( 10 ) or decision tree analysis to identify baseline impedance measurements ( 11 ).…”

On‐therapy impedance‐pH monitoring can efficiently characterize PPI‐refractory GERD and support treatment escalation

“…It has been argued that assessment of PSPW index and MNBI requires additional time for analysis of impedance‐pH tracings 25 : however, given the current limitations of available commercial software for recognition of reflux events, manual analysis of impedance‐pH tracings is regularly warranted 6 and the time required is mainly due to meticulous definition of reflux episodes 11 whereas the additional time needed for calculation of PSPW index and MNBI consists of few minutes only, an extra time justified when the issue is a firm diagnosis of PPI‐refractory GERD, reliable enough to support treatment escalation including surgery. Interestingly, a recent study suggests that artificial intelligence can afford accurate measurement of reflux episodes and PSPWs, thus favoring wider application of impedance‐pH monitoring 26 …”