Device-assisted enteroscopy · Live vision AI
Explore deeper. Detect with confidence.
AI assistance for detecting and characterizing lesions during device-assisted enteroscopy throughout the gastrointestinal tract.
DEVICE-ASSISTED ENTEROSCOPY · LIVE AI ACTIVE
PROTRUDING LESION97%
DAE · FRAME 02418
SEGMENTANTEGRADE · JEJUNUMLIVE DETECTION
PROTRUDINGVASCULARULCER / EROSIONBLOOD
FRAME ANALYSIS157 FPSMULTI-LESION ANALYSISFINDING READY FOR ENDOSCOPIST CONFIRMATION
01 · Intended use
Real‑time, multi‑lesion detection during DAE.
Deep Bowel® is a deep-learning technology developed for application during device-assisted enteroscopy examinations.
It automatically detects protruding lesions, vascular lesions, ulcers, erosions and blood in real time across the different segments of the digestive tract evaluated during the procedure.
AI
The clinical challenge
The small bowel remains difficult territory.
Device-assisted enteroscopy enables deep diagnostic and therapeutic exploration, but subtle lesions can appear briefly during technically demanding navigation.
“The opportunity is to support attention precisely where difficult access and subtle findings meet.”
02
How it works
Live assistance from navigation to clinical action.
Deep Bowel® supports the endoscopist during examination while keeping navigation, interpretation and treatment in expert hands.
Enteroscope navigation
The clinician advances through the gastrointestinal tract using device-assisted enteroscopy.
Live frame analysis
The AI continuously evaluates the endoscopic image during examination.
Lesion detection
Protruding and other relevant mucosal lesions are highlighted when identified.
Finding characterization
Detected abnormalities are categorized and localized for focused inspection.
Clinical decision
The endoscopist validates the finding and determines diagnostic or therapeutic action.
AI capabilities
Designed for deep, live endoscopic exploration.
97.3%
Research frame-level accuracy
Device-assisted enteroscopy
A pioneer AI approach developed specifically for endoscopic exploration beyond conventional reach.
Protruding lesion detection
Research validation distinguished enteric protruding lesions from normal mucosa and other findings.
Panendoscopic potential
Designed to support evaluation across esophageal, gastric, enteric and colonic mucosa.
Real-time-ready analysis
The research CNN processed approximately 157 frames per second in its validation dataset.
Clinician-centered assistance
Findings remain in live endoscopic context for expert confirmation and action.
03
Scientific validation
Research developed specifically for device‑assisted enteroscopy.
Artificial Intelligence and Device‑Assisted Enteroscopy: Automatic Detection of Enteric Protruding Lesions Using a Convolutional Neural Network
Cardoso P, Mascarenhas Saraiva M, Afonso J, et al. · 2022 · Proof-of-concept validation
Read publication ↗Clinical benefits
Deeper access. More findings in focus.
01
More consistent detection of subtle protruding lesions
↗02
AI assistance during technically demanding deep-enteroscopy exploration
↗03
Support for panendoscopic evaluation throughout the gastrointestinal tract
↗04
Findings remain immediately available for biopsy or endoscopic therapy
↗Deep Bowel®
Go deeper. See more. Act with confidence.
Talk to our team ↗Deep Bowel® is part of DigestAID’s research and development pipeline. Availability and regulatory status may vary by market.