Skip to content

A study to determine how artificial intelligence assistance affects the accuracy of a physician's diagnosis of colorectal polyps.

Influence of Computer Aided Diagnosis System on Endoscopist's Diagnostic Accuracy in Diagnosing Colorectal Polyps -A prospective observational study using still images- - Influence of Computer Aided Diagnosis System on Endoscopist's Diagnostic Accuracy in Diagnosing Colorectal Polyps -A prospective observational study using still images-

Status
Active, not recruiting
Phases
Unknown
Study type
Interventional
Source
JPRN
Registry ID
JPRN-UMIN000049564
Enrollment
14
Registered
2022-11-20
Start date
2022-09-01
Completion date
Unknown
Last updated
2026-06-29

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Colorectal polyps

Interventions

The lesion images will be endoscopic images of polyps for which the pathological diagnosis has been confirmed in patients who underwent colonoscopy at the Digestive Organs Center of Showa University N
Not a good sample&quot
Blur image&quot
will be displayed and no support result will be output. The researcher will display a predesigned sample size of anonymized lesion images on a stand-alone computer, and while viewing the lesion image

Sponsors

ShowaUniversityNorthernYokohamaHospital
Lead Sponsor

Eligibility

Sex/Gender
All

Inclusion criteria

Inclusion criteria: Epithelial lesions of the colon measuring 10 mm or less in diameter imaged at the Digestive Disease Center, Showa University Northern Yokohama Hospital from July 1, 2021 to the end of January, 2022 2. lesions taken from patients who have not expressed their refusal to participate in this study 3. lesions taken from patients who were 20 years of age or older at the time of examination

Exclusion criteria

Exclusion criteria: 1. lesions for which no NBI images have been captured 2. lesions for which the NBI images taken are too blurred to be annotated 3. lesions for which only images showing multiple lesions were taken 4. lesions imaged from patients with inflammatory bowel disease 5. lesion imaged from a patient with polyposis

Design outcomes

Primary

MeasureTime frame
To test whether the sensitivity of physicians to neoplastic lesions is improved by the use of AI with statistical significance.

Secondary

MeasureTime frame
1. physician's specificity, accuracy, PPV, and NPV for neoplastic lesions 2. sensitivity, specificity, accuracy, PPV, and NPV of neoplastic lesions less than 5 mm when the physician diagnosed with high confidence 3. percentage of lesions that the physician could diagnose with high confidence

Countries

Japan

Contacts

Public ContactMasashi Misawa

ShowaUniversityNorthernYokohamaHospital Digestive Disease Center

misawaanny@gmail.com0459497000

Outcome results

None listed

Source: JPRN (via WHO ICTRP) · Data processed: Jul 3, 2026