AI in GIM Diagnosis
Study Details
Study Description
Brief Summary
This study will use artificial intelligence (AI) for diagnosing gastric intestinal metaplasia.
Condition or Disease | Intervention/Treatment | Phase |
---|---|---|
|
N/A |
Detailed Description
The patients with previously diagnose gastric intestinal metaplasia (GIM) and have the surveillance gastroscopy will be enrolled. The routine surveillance program will be performed additional to taking photo at both GIM and normal mucosa at least 5 pictures in each. Biopsy will be done to confirm the diagnosis of GIM and normal mucosa. All pictures will be inserted to AI algorithm based on the convolutional neural network (CNN). Then, the AI program will be validated in daily endoscopy compared with pathology. Accuracy, sensitivity and specificity can be calculated by 2x2 table.
Study Design
Arms and Interventions
Arm | Intervention/Treatment |
---|---|
Experimental: GIM patient The patients with GIM will be assessed at both GIM and normal mucosa during endoscopy. |
Diagnostic Test: Artificial intelligence
The AI algorithm based on the convolutional neural network (CNN) will be used for analysis the pictures of gastric intestinal metaplasia and normal mucosa. Then AI will be used as a diagnostic tool for GIM during routine endoscopy by using pathology as a gold standard.
|
Outcome Measures
Primary Outcome Measures
- Accuracy of AI for GIM diagnosis [1 year]
Accuracy, sensitivity, specificity can be calculated by 2x2 table (pathology is a gold standard)
Eligibility Criteria
Criteria
Inclusion Criteria:
-
More than 18 years of age
-
Able to sign a consent form
Exclusion Criteria:
-
History of gastric surgery
-
Coagulopathy
-
Pregnancy/Breast feeding
Contacts and Locations
Locations
Site | City | State | Country | Postal Code | |
---|---|---|---|---|---|
1 | Rapat Pittayanon | Pathum Wan | Bangkok | Thailand | 10330 |
Sponsors and Collaborators
- King Chulalongkorn Memorial Hospital
Investigators
None specified.Study Documents (Full-Text)
None provided.More Information
Publications
None provided.- RP018