Abstract
AIM: We proposed an integrated endoscopic strategy and compared it with conventional methods to assess whether it improves the diagnostic accuracy of Helicobacter pylori infection status. METHODS: A retrospective cohort (between 1 August 2022 and 30 April 2024; n = 163) was used to compare white light endoscopy (WLE) and magnifying endoscopy (ME) and to prespecify the integrated endoscopy (IE) algorithm. IE first uses ME to grade the proportion of fundic crypt openings and then applies WLE features to classify H. pylori status. A prospective cohort (between 1 May 2024 and 30 December 2024; n = 221) applied the locked IE algorithm and compared its performance with WLE and ME. The primary outcome was overall diagnostic accuracy; secondary outcomes were accuracy for each infection status category. RESULTS: In the retrospective cohort, IE achieved higher overall accuracy than WLE and ME (79.8% vs 69.9% and 71.8%, respectively), with improved accuracy for negative (85.4%) and eradicated status (72.5%). ME was more accurate than WLE for negative (82.9% vs 61.0%) and eradicated status (62.7% vs 52.9%), whereas WLE was more accurate for positive status (87.3% vs 71.8%). In the prospective cohort, IE again showed the highest overall accuracy (86.4% vs 70.6% for WLE and 72.9% for ME), with higher accuracy for negative (90.7%) and eradicated status (92.4%), while WLE maintained higher accuracy for positive status (89.8% vs 78.4% for IE). CONCLUSION: IE that combines ME-based fundic crypt opening assessment with WLE features may improve endoscopic classification of overall, negative, and eradicated H. pylori infection status compared with WLE or ME alone. These findings support IE as a promising diagnostic approach that warrants further multicenter confirmation. (ClinicalTrials.gov, NCT06397066.).