Randomised trial of voice-guided AI navigation for biliopancreatic endoscopic ultrasound
In a single-centre randomised trial of 266 patients in China, a real-time AI system that recognises scanning stations and prompts the endoscopist by voice raised the completeness of guideline-recommended biliopancreatic endoscopic ultrasound stations from a median of 0.81 to 1.00, while lengthening procedures by about two minutes.
- Model
- EUS-GUIDEQ
- Field
- Medicine
- Date
- 2026-10-28
- Human collaborators
- Rui Wang, Xiaoyan Wang
What was found
EUS-GUIDEQ was developed from 27,072 annotated images from 189 patients, with station classification accuracy of 94.2% on 1,500 internal images and 78.8% on 2,062 multicentre external images. Between December 2023 and May 2025 patients were randomised 1:1 to AI-assisted or standard examination, with 132 and 134 analysed. The primary endpoint, completeness of standardised scanning stations adjudicated by blinded experts from recorded video, was a median of 1.00 (interquartile range 0.94 to 1.00) against 0.81 (0.75 to 0.88), P < 0.001, with benefit across participating endoscopists. Procedure time was 12.70 against 10.55 minutes, and no serious procedure-related adverse events occurred.
Novelty check
Europe PMC, 2026-10-29: TITLE_ABS:"endoscopic ultrasound" AND TITLE_ABS:"artificial intelligence" AND TITLE_ABS:"randomized" returned 4 records: this trial, a 2026 study of AI and endosonographers' detection of solid pancreatic lesions, a 2026 mediastinal EUS navigation system and a 2023 detection study. AI guidance for EUS is not new; no earlier randomised trial of real-time station guidance for biliopancreatic EUS was found.
Caveats and known objections
Developed and tested by the same investigators at the Third Xiangya Hospital of Central South University, with co-authors from other hospitals in Hunan and Wuhan; no vendor or funder is named in the abstract. Prospective and randomised but single-centre, and the endpoint is a process surrogate, scanning completeness, not diagnostic yield or patient outcome; the authors call for multicentre, multi-vendor validation. External station accuracy fell to 78.8%, and the gain came with about two more minutes per procedure. Autonomy is ai-assisted: the endoscopist performed the examination and the system prompted.
No independent check is recorded on this entry yet.
Reading the primary source closely enough to say whether it supports the claim counts as a check, and you are credited on the entry.
Or on GitHub: submit a check challenge the grade send a correction or send a pull request
Flag this for triage
Signals order the review queue and nothing else. They are never published, and they never move a grade: that takes a citation.
Entry history (1 event)
- AddedEntered the registry graded Peer reviewed and AI-assisted.
Entries are never deleted. A grade that does not hold up is downgraded on the record, with the reason beside it.
Graded peer reviewed for verification and ai-assisted for autonomy. What these mean.
Cite this entry
whataifound.org. (2026). Randomised trial of voice-guided AI navigation for biliopancreatic endoscopic ultrasound. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2026-10-28-eus-guideq-endosonography-rctBibTeX
@misc{whataifound-centralsouthuniversity-2026-rct,
title = {Randomised trial of voice-guided AI navigation for biliopancreatic endoscopic ultrasound},
author = {{whataifound.org}},
year = {2026},
howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
note = {Result by Central South University (Third Xiangya Hospital). Verification: Peer reviewed. Autonomy: AI-assisted.},
url = {https://whataifound.org/finding/2026-10-28-eus-guideq-endosonography-rct}
}