Newswise — The editors of The American Journal of Gastroenterology proudly announce the publication of a special issue, “AI Clinical Applications in GI and Hepatology.” This issue was thoughtfully assembled to explore the burgeoning challenges, clinical promise, and ethical considerations of AI use in GI.

“…We present a collection of studies examining today’s AI capabilities (or lack thereof) across gastroenterology, with a few examples of early groundbreaking work that will transform what it means to be a gastroenterologist,” writes special editorial author Ryan W. Stidham, MD, MS. Topics include computer-aided diagnosis (CAD) in endoscopy; large language models to standardize, educate, and automate; magnetically driven robotic colonoscopy; patient perspectives and comfort; ethical and legal considerations; and AI vs. GI experts.

Access to any articles from this issue, or past issues, is available upon request. The College is also able to connect members of the press with study authors or outside experts who can comment on the articles.

Current Issue:

The Magnetic Flexible Endoscope: Phase 1 First-in-Human Clinical Trial
Researchers tested a novel magnetic flexible endoscope (MFE) for colonoscopy that is controlled externally with magnets, robotic control, and real-time image processing. Five patients had a standard colonoscopy with sedation, after which sedation was ceased and the MFE was inserted and advanced through the colon while the patients were awake. They reported no discomfort, pain, trauma, or other adverse events, and there were no software or system failures.

Computer-Aided Colonoscopy Alert Fatigue and Its Effect on Adenoma Detection
This study examined whether alert fatigue affects the effectiveness of computer-aided detection (CADe) in improving adenoma detection rates (ADR) during colonoscopies. Researchers found that CADe significantly increased ADR in early-day procedures but not later in the day. In procedures without CADe, ADR remained stable throughout the day, suggesting that CADe can enhance ADR initially but its effectiveness may diminish due to alert fatigue.

Use of Natural Language Processing to Objectively Identify Hepatic Encephalopathy in Multiple Cohorts
This study identified simple phrases that reliably indicate a hepatic encephalopathy (HE) episode, using natural language processing on inpatient charts. Five key features identified HE with 85.8% sensitivity and 82.1% negative predictive value (NPV), 100% sensitivity/NPV in an internal validation cohort, and 94.1% sensitivity/84.2% NPV in an external cohort—asterixis, altered mental status, confusion, and initiation/continuation of lactulose or rifaximin.

Patient Perspectives on Artificial Intelligence in Gastroenterology: A Multicenter Survey of Knowledge, Concerns, and Beliefs
Results from a patient survey on artificial intelligence in gastroenterology found that participant knowledge of AI was mostly limited and their trust in AI was moderate. Most respondents believed AI could enhance physician care, but felt physicians should make final decisions and wanted to be informed when AI is used. Concerns included AI reliability, data privacy, and healthcare costs, especially among racial minorities and those with lower socioeconomic status.

About the American College of Gastroenterology
Founded in 1932, the American College of Gastroenterology (ACG) is an organization with an international membership of over 21,000 individuals from 86 countries. The College’s vision is to be the preeminent organization supporting health care professionals who provide compassionate, equitable, high-quality, state-of-the-art, and personalized care to promote digestive health. The mission of the College is to provide tools, services, and accelerate advances in patient care, education, research, advocacy, practice management, professional development and clinician wellness, enabling our members to improve patients’ digestive health and to build personally fulfilling careers that foster well-being, meaning and purpose. Learn more at www.gi.org.