What type of study is this?

This is a Quantitative Study study.

October 15, 2025Open Access

A Rural-to-Center Artificial Intelligence Model for Diagnosing Helicobacter pylori Infection and Premalignant Gastric Conditions Using Endoscopy Images Captured in Routine Practice

TCTsung‐Hsien ChiangNational Taiwan University YHYen-Ning HsuNational Taiwan University MCMin-Han ChenFudan University

Key Points

AI achieved a diagnostic accuracy of 92.8% for helicobacter pylori during model development, ensuring effective detection.
In implementation, AI's accuracy for atrophic gastritis dropped to 79.9%, highlighting potential challenges in diagnosis.
Endoscopic images from rural hospitals were analyzed using AI, which significantly improved diagnostic access for patients.
No significant differences were found between AI predictions and clinical observations, reinforcing AI reliability.

Abstract

Background and study aim: Diagnosing Helicobacter pylori infection and premalignant gastric conditions typically requires a 13C urea breath test or histological examination, which are often unavailable in remote areas. A rural-to-center artificial intelligence (AI) model was developed and implemented to automatically evaluate upper endoscopic images captured during routine clinical practice. Patients and methods: Endoscopic images were collected from a rural hospital on Matsu Islands and a tertiary center across Taiwan Strait. During model development (2012–2022), AI algorithms were trained, validated, and tested to exclude low-quality and non-gastric images, segment gastric regions, and enhance mucosal features for detecting H. pylori infection and premalignant condition. During model implementation (2023–2024), endoscopic images from a rural hospital were transmitted to the medical center for AI analyses, with results promptly returned. Results: In the development phase, diagnostic accuracies were 92.8% (95% Confidence Interval CI: 88.9%–96.6%) for H. pylori, 88.6% (95%CI: 87.2%–90.0%) for atrophic gastritis, and 88.0% (95%CI: 86.5%–89.5%) for intestinal metaplasia. In the implementation phase, 3518 residents were invited to undergo 13C urea breath testing or histological assessment in rural communities. No significant differences were observed between AI-predicted and clinically observed prevalence: H. pylori (13.9% vs. 12.9%, P=0.55), atrophic gastritis (15.7% vs. 11.9%, P=0.34), and intestinal metaplasia (27.6% vs. 22.4%, P=0.32). Implementation-phase diagnostic accuracies were 91.3% (95%CI: 88.0%–94.6%), 79.9% (95%CI: 72.1%–86.3%), and 63.4% (95%CI: 54.7%–71.6%), respectively. Conclusions: AI enables frontline physicians in resource-limited settings to rapidly assess gastric health using routinely captured endoscopic images, bridging gaps in access and expertise (ClinicalTrials.gov: NCT05762991).

Ask AI

Helpful

Bookmark

View Full Paper