Artificial Intelligence, Colonoscopy
Conditions
Brief summary
In this study, the investigators proposed a prospective study about the effectiveness of speech and image recognition-based system in improving reporting quality during colonoscopy for colonoscopy report quality in endoscopists. The participants would be divided into two groups. For the collected colonoscopy videos, group A would record their observations with the assistance of the artificial intelligence system. The artificial intelligence assistant system can automatically capture bowel segment images and prompt abnormal lesions. Group B would complete the endoscopy report without special prompts. After a period of washout period, the two groups switched, that is, group A without AI assistance and group B with AI assistance to complete the colonoscopy report. Then, the completeness of the colonoscopy report, the completeness of capturing anatomical landmarks and detected lesions, the completeness of structured description, the accuracy of lesion reporting, the time for reporting and the satisfaction with the reporting system are compared with or without AI assistance.
Interventions
DIAGNOSTIC_TESTArtificial intelligence assistant system
The artificial intelligence assistant system can automatically capture bowel segment images and prompt abnormal lesions based on speech recognition and deep learning.
Sponsors
Renmin Hospital of Wuhan University
Study design
Allocation
RANDOMIZED
Intervention model
CROSSOVER
Primary purpose
DEVICE_FEASIBILITY
Masking
NONE
Eligibility
Sex/Gender
ALL
Age
18 Years to 70 Years
Healthy volunteers
Yes
Inclusion criteria
Patients: 1. Male or female ≥18 years old; 2. Able to read, understand and sign an informed consent; 3. The investigator believes that the subjects can understand the process of the clinical study, are willing and able to complete all study procedures and follow-up visits, and cooperate with the study procedures; 4. Patients requiring colonoscopy. Doctors: 1. Males or females who are over 18 years old; 2. After qualified medical education and obtaining the Physician's Practice License.
Exclusion criteria
Patients: 1. Have drug or alcohol abuse or mental disorder in the last 5 years; 2. Pregnant or lactating women; 3. Patients with known multiple polyp syndrome; 4. patients with known inflammatory bowel disease; 5. known intestinal stenosis or space-occupying tumor; 6. known colon obstruction or perforation; 7. patients with a history of colorectal surgery; 8. Patients with a previous history of allergy to pre-used spasmolysis; 9. Unable to perform biopsy due to coagulation disorders or oral anticoagulants; 10. High-risk diseases or other special conditions that the investigator considers the subject unsuitable for participation in the clinical trial. Doctors: 1\. The researcher believes that the subjects are not suitable for participating in clinical trials.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| The integrity of colonoscopy report | One month | Report integrity with or without AI-assisted. Calculation method = number of information recorded / total number of information need to record x 100% |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| The integrity of capturing anatomical landmarks | One month | The integrity in captured bowel landmrak images with or without AI-assisted. Calculation method = number of anatomical landmarks in captured images / total number of anatomical landmarks x 100% |
Other
| Measure | Time frame | Description |
|---|---|---|
| The integrity of report lesion | One month | Report lesion integrity with or without AI-assisted. Calculation method = number of report lesions / total number of lesions x 100% |
| The satisfaction with the reporting system | One month | The satisfaction with the reporting system with or without AI-assisted |
| Accuracy of lesion reporting | One month | Accuracy of lesion report with or without AI-assisted. Calculation method = number of lesions with correct description / total number of lesions descriptionx 100% |
| The completeness of structured description | One month | The completeness of structured description with or without AI-assisted. Calculation method = number of structured descriptions / total number of structured descriptions need to record x 100% |
| The time for reporting | One month | The time for reporting with or without AI-assisted |
Countries
China
Contacts
Primary ContactHonggang Yu, MD
Outcome results
None listed