Deep Capsule
Overview
Deep Capsule is a non-invasive deep learning software driven by artificial intelligence (AI) designed to identify and differentiate normal mucosa and small bowel lesions in images from previously acquired capsule endoscopy videos. It assists in the post-procedural review of small bowel capsule endoscopy examinations by detecting and differentiating lesions, stratifying hemorrhagic potential, and bringing clinically relevant frames into focus for physician validation. The software supports clinicians throughout the review process while ensuring that the final interpretation remains in expert hands.
Digestaid, the company behind Deep Capsule, focuses on developing accurate deep learning solutions for the detection of digestive lesions, aiming to revolutionize gastrointestinal practice with AI technologies across digestive and pancreatobiliary tracts, as well as functional tests. They leverage a team of gastroenterologists and engineers to innovate in digestive healthcare.
Reviewed by Pouyan Golshani, MD — Interventional Radiologist
Key Features
- Non-invasive deep learning solution for small bowel mucosa evaluation
- Detects and differentiates lesions (vascular lesions, protruding lesions, ulcers, erosions, blood)
- Stratifies hemorrhagic potential using a validated scale
- Brings clinically relevant frames into focus for physician validation
- Multi-device interoperability (validated with full-length videos from three different capsule endoscopy devices)
- Rapid assisted review (mean AI-assisted reading time of 203 seconds per examination)
- Improved detection of clinically relevant small bowel lesions
- Faster review of long capsule endoscopy examinations
- Reduced risk of overlooking findings across thousands of frames
- Consistent lesion characterization
Use Cases
- Post-procedural review of small bowel capsule endoscopy examinations
- Aid in identifying and differentiating lesions in the small bowel mucosa
- Detection of lesions in patients aged 18 or above with suspected small bowel disease
- Assisting in cases of suspected middle gastrointestinal bleeding
- Supporting inflammatory bowel disease assessment
- Aiding in digestive oncology evaluations
What Physicians Need to Know
Utilize Deep Capsule for efficient post-procedural review of small bowel capsule endoscopy videos, focusing on its ability to highlight and differentiate lesions, and stratify bleeding risk. Always maintain final interpretation in expert hands, using the AI as a powerful assistive tool to reduce reading time and improve diagnostic focus. Be aware that while Deep Capsule excels in small bowel analysis, other AI tools are emerging for colon capsule endoscopy and real-time applications.
Deep Capsule is designed to process previously acquired capsule endoscopy videos and is compatible with major capsule endoscopy systems like Jinshan OMOMu00ae HD, GIVEN PillCamu2122 (SB1, SB2, SB3), and ENDOCAPSULE 10. Ensure your existing capsule endoscopy acquisition workflow can seamlessly export videos for Deep Capsule's review process.
Details
| Category | Gastroenterology AI |
| Pricing | Unknown — unknown |
| Deployment | Web-based application (requires updated browser, 4GB RAM, 1GHz CPU, 200GB storage, 1280x1024 screen resolution, stable broadband internet) |
| Compliance | |
| BAA Available | Unknown AI-estimated |
| HIPAA Compliant | Unknown AI-estimated |
| FDA Status |
Cleared AI-estimated Deep Capsule® (Deep Capsule US) received FDA 510(k) clearance. It is an AI-assisted reading tool designed to aid small bowel capsule endoscopy reviewers for adult patients in whom the capsule endoscopy images were obtained for suspected small bowel bleeding. |
| Integrations | |
| EHR | Not specified |
| Specialties | Gastroenterology |
What the Web Says
Deep Capsule is an AI-assisted reading system designed for small bowel capsule endoscopy (CapE) examinations. It aims to improve the detection and differentiation of lesions, reduce reading times for physicians, and minimize the risk of overlooking clinically relevant findings. The technology has received FDA approval, indicating growing acceptance of AI in ingestible imaging devices.
Overall: PositiveStrengths
- Superior sensitivity in detecting lesions compared to standard-of-care reading.
- Significantly reduces the time required for physicians to review capsule endoscopy examinations (median time of 172 seconds).
- Cost-effective technology, saving valuable resources and time.
- Decreased risk of overlooking clinically relevant lesions.
- Aids in identifying and differentiating lesions in the small bowel mucosa, with particular importance in suspected middle gastrointestinal bleeding, inflammatory bowel disease, and digestive oncology.
- Multi-device interoperability, validated with full-length videos from various capsule endoscopy devices.
Limitations
- Lower positive predictive value in suspected Crohn's Disease (CD) cohorts due to a higher rate of false positives.
- False positives primarily attributed to poor visualization, bowel contents, bubbles, and non-erosive inflammatory findings.
- Not intended to replace clinical decision-making.
- Contraindicated for use in CapE exams that failed to reach the small bowel, did not record any small bowel segment, or when good bowel cleansing was not achieved.
- The algorithm was not trained on data from pregnant women or children, so it should not be used in these demographics.
- Capsule networks, in general, may have higher computational costs and do not always scale well to very deep architectures.
Based on reviews from: Mentoring In IBD, DigestAID, Coherent Market Insights, IEEE Computer Society, MDPI, FDA
Last updated: 2026-08-19
Ratings & Reviews
No reviews yet. Be the first to review this tool!
Rate Deep Capsule
Videos
Product demos, reviews, and walkthroughs for Deep Capsule.
