Deep Learning-based Artificial Intelligence for the Diagnosis of Small Bowel Obstruction

Active, not recruiting

• Conditions: Bowel Obstruction; Artificial Intelligence
• Interventions: Diagnostic Test: Artificial intelligence

• Registration Number: NCT06481358
• Lead Sponsor: Nagoya University

Brief Summary

The study will compare the diagnostic accuracy and time to diagnosis of computed tomography images of patients with suspected intestinal obstruction seen in the emergency room by residents and surgeons, with and without artificial intelligence.

Detailed Description

DESIGN: This is an diagnostic study. SETTING: We developed a deep learning-based AI technology to automatically extract the intestinal tract from CT images using 5 200 CT images of 158 patients. The CT images of patients who visited the emergency department and were suspected of small bowel obstruction between June 6 and July 26, 2018, were obtained from two tertiary referral centers, which were used as the test samples. Data analysis was completed in December 2023.

PARTICIPANTS: Residents and surgeons participated in the study. INTERVENTIONS: Residents and surgeons were divided into two groups: one group read using the AI technology, and the other group read without the AI technology.

MAIN OUTCOMES AND MEASURES: Participants indicated whether or not small bowel obstruction and obstruction location. The time for diagnosis was also collected. We applied a hierarchical Bayesian model.

Study Timeline

• Study Start: 2022-09-01
• Primary Completion: 2023-06-06
• Status Verified: 2024-06
• Study First Submitted: 2024-06-25
• Study First Submitted QC: 2024-06-25
• Last Update Submitted: 2024-06-25
• Study First Posted: 2024-07-01
• Last Update Posted: 2024-07-01
• Study Completion: 2024-10-31

Recruitment & Eligibility

• Status: ACTIVE_NOT_RECRUITING
• Sex: All
• Target Recruitment: 17

Inclusion Criteria

• Persons with documented consent

Exclusion Criteria

• Persons without documented consent

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Arm && Interventions

Group	Intervention	Description
AI group	Artificial intelligence	Participants read CT images with AI.

Primary Outcome Measures

Name	Time	Method
The diagnosis of the obstruction site	September, 2024	Accuracy of diagnosis of the obstruction site

Trial Locations

Locations (1)

Nagoya University Graduate School of Medicine; 🇯🇵 Nagoya, Aichi, Japan