*EMBARGOED All research presented at the 2022 ACG Annual Scientific Meeting and Postgraduate Course is strictly embargoed until Sunday, October 23, 2022, at 12:00 pm EDT.
Poster D0209 – Voice Enabled Artificial Intelligence for Detection of Pathologic Gastroesophageal Reflux Disease and Barrett’s Esophagus
Tuesday, October 25, 2022 | 10:00 AM – 12:00 PM ET | Location: Crown Ballroom
Author Insight from Amrit K. Kamboj, MD, Mayo Clinic
What’s new here and important for clinicians?
Gastroesophageal reflux disease (GERD) can lead to voice alterations such as hoarseness. In this study, we sought to identify specific voice biomarkers associated with pathologic GERD and Barrett’s esophagus (BE). We captured a short voice recording in patients undergoing a clinically indicated upper endoscopy and/or pH monitoring study. Patients were carefully chosen and those with other medical conditions that could result in voice disturbance were excluded. We compared patients with GERD (GERD+) and BE to those without GERD (GERD-) and to a vocally normal group of individuals with normal voice as judged by speech pathology. Random forest models were trained and demonstrated good to excellent ability to discern BE and GERD+ from the GERD- and vocally normal groups, respectively, for both men and women. Voice differences were in part related to voice signal periodicity which mirrors voice quality. These results suggest that voice biomarkers may be useful as a non-invasive tool in the detection of pathologic GERD and BE.
What do patients need to know?
At present, there is no widely available, non-invasive screening tool that can detect gastroesophageal reflux disease (GERD) and Barrett’s esophagus (BE). Since GERD can lead to voice alterations such as hoarseness, we sought to identify specific voice biomarkers that are associated with GERD and BE using a short voice recording from patients with GERD and BE and those without GERD and those with normal voices as judged by speech pathology. Using machine learning tools, we showed that there were voice quality differences between those with BE and GERD as compared to those with normal voices and those without GERD. Our results suggest that voice biomarkers may be a useful tool to detect GERD and BE. However, additional studies are necessary to further study voice differences in patients with and without GERD and BE before determining whether voice biomarkers can serve as a reliable non-invasive tool in the detection of these conditions.
Author Contacts
Amrit K. Kamboj, MD (First Author) & Cadman Leggett, MD (Senior Author), Mayo Clinic
kamboj.amrit [at] mayo.edu & leggett.cadman [at] mayo.edu
Media Interview Requests
To arrange an interview with any ACG experts or abstract authors, please contact Becky Abel of ACG via email at mediaonly [at] gi.org or by phone at 301-263-9000.