Plain-English translation of NCT07078136 on ClinicalTrials.gov ↗ · Source last updated · Translation generated · How we translate trials
Researchers are testing a new artificial intelligence (AI) system designed to help doctors diagnose small tumors that grow in the wall of the stomach and upper digestive tract. The AI tool analyzes images from two types of endoscopy scans—ultrasound images taken from inside the stomach and regular camera images—combined with your medical information. The goal is to see if this AI system can identify tumors more accurately than doctors working alone.
Currently, doctors correctly identify these types of stomach tumors about 70–75% of the time. Researchers believe an AI system that looks at multiple types of imaging data together could reach 95% accuracy, helping ensure patients get the right diagnosis and treatment plan faster.
You likely qualify if…
You likely don't qualify if…
If you qualify, you will have already had the imaging scans and biopsy needed to diagnose your tumor as part of your normal medical care. For this study, researchers will review the images and medical records you already have on file, and the AI system will analyze those same images. You will also be asked to let doctors independently assess your images to compare how the AI tool performs versus physician diagnosis. There are no additional procedures or visits required beyond what you've already completed for your diagnosis.
AI-generated summary from trial data · Jun 4, 2026 · Not medical advice
China
Sponsor
Huazhong University of Science and Technology
Enrollment target
~130 participants
Started
July 2025
Primary completion
March 2026
This trial's estimated completion date has passed — the record may not be fully up to date.
Age range
18 Years and older
Last updated on clinicaltrials.gov in July 2025.
Reach out to the team running this trial. Response times vary — some teams are faster than others.
Central contact
Bin Cheng
Huazhong University of Science and Technology
Tell us you're interested and we'll help connect you with the research team. We'll walk you through what to expect first — no email needed to get started.