Real-World Validation of an Artificial Intelligence Characterization Support (CADx) System

Sponsor
Changi General Hospital (Other)
Overall Status
Recruiting
CT.gov ID
NCT05034185
Collaborator
National University Hospital, Singapore (Other), Singapore General Hospital (Other), Tan Tock Seng Hospital (Other)
400
1
21.9
18.3

Study Details

Study Description

Brief Summary

Colorectal cancer (CRC) is a leading cause of cancer-related morbidity and mortality worldwide, with rates of CRC predicted to increase. Colonoscopy is currently the gold standard of screening for CRC. Artificial intelligence (AI) is seen as a solution to bridge this gap in adenoma detection, which is a quality indicator in colonoscopy. AI systems utilize deep neural networks to enable computer-aided detection (CADe) and computer-aided classification (CADx). CADe is concerned with the detection of polyps during colonoscopy, which in turn is postulated to help decrease the adenoma miss-rate.

In contrast, CADx deals with the interpretation of polyp appearance during colonoscopy to determine the predicted histology. Prediction of polyp histology is crucial in helping Clinicians decide on a "resect and discard" or "diagnose and leave strategy". It is also useful for the Clinician to be aware of the predicted histology of a colorectal polyp in determining the appropriate method of resection in terms of safety and efficacy. While CADe has been studied extensively in randomized controlled trials, there is a lack of prospective data validating the use of CADx in a clinical setting to predict polyp histology.

The investigators plan to conduct a prospective, multi-centre clinical trial to validate the accuracy of CADx support for prediction of polyp histology in real-time colonoscopy.

Condition or Disease Intervention/Treatment Phase
  • Device: Computer-aided diagnosis (CADx) support tool

Detailed Description

Colonoscopy is currently the gold standard of screening for CRC. A 1% increase in adenoma detection rate (ADR) estimated to be associated with a 3% decreased risk of interval CRC. AI systems can be broadly divided into CADe (for detection) and CADx (for diagnosis, or prediction of polyp histology in the context of colonoscopy). CADe has been extensively studied, with several randomized controlled trials and meta-analysis showing a higher ADR when CADe is used compared to the control groups without CADe.

Besides the ADR, predicted polyp histology is a key component in the performance of colonoscopy as this enables the Clinician to make a decision regarding its management, as described above. In this regards, image-enhanced endoscopy (IEE) is often used to help Clinicians determine if colorectal polyps found on colonoscopy are neoplastic or hyperplastic. The most commonly used non-magnification classification is the NBI International Colorectal Endoscopic (NICE), while the Japan NBI Expert Team (JNET) classification is used where endoscopy systems with optical magnification and the proper training is available. However, these classification systems have varying diagnostic accuracy and interobserver agreement. Previous prospective studies looking at CADx have utilized endocytoscopy and autofluorescence imaging (CAD-AFI) with positive results. However, the major limitation in these CADx studies is that these imaging systems are costly and are not readily available in most centres worldwide. Furthermore, most Clinicians performing colonoscopies have not been trained in these modalities of imaging and will have to rely completely on the CADx function to detect polyps if these imaging modalities are used, without being able to fall back on their experience and training should there be doubts about the accuracy of a CADx diagnosis in a real-world setting.

The Fujifilm 7000 System (Fujifilm Corp., Tokyo) has been in routine clinical use in all tertiary institutions in Singapore. The CAD EYE system was developed by Fujifilm Corp to aid Clinicians in colonoscopy with CADe and CADx functions. The basic functions and handling of the colonoscope, as well as the endoscopy processing unit, are similar to what is currently available in clinical practice, with the added CAD EYE software. The controller has been configured to allow the operator to activate and deactivate the CAD function depending on the need for it. These functions can be turned on and off using a button on the controller by the Clinician. The CADe and CADx functions operate when white light and blue laser imaging (BLI) are used, respectively. This provides a unique opportunity to externally validate the use of the CADx support tool by evaluating its diagnostic accuracy with final polyp histology as the gold standard, while also comparing its performance in a clinical setting against a Clinician using IEE (which is the conventional method of predicting polyp histology in colonoscopy).

The investigators plan to conduct a prospective, multi-centre clinical trial to validate the accuracy of CADx support for prediction of polyp histology in real-time colonoscopy.

Study Design

Study Type:
Observational
Anticipated Enrollment :
400 participants
Observational Model:
Cohort
Time Perspective:
Prospective
Official Title:
Real-World Validation of an Artificial Intelligence Characterization Support (CADx) System for Prediction of Polyp Histology in Colonoscopy: A Prospective Multicentre Study
Actual Study Start Date :
Mar 3, 2021
Anticipated Primary Completion Date :
Oct 3, 2022
Anticipated Study Completion Date :
Dec 30, 2022

Arms and Interventions

Arm Intervention/Treatment
Patients with one or more polyps detected

During colonoscopy, the Clinician inspect for the presence of polyps as per routine clinical practice with the CAD EYE function turned off. When a polyp is encountered, the Clinician will make a prediction on the histology based on the white light and BLI features of the polyp with and without optical magnification, as per routine clinical practice. Following this, the CAD EYE function will be switched on and the Clinician will take note of the CADx prediction for the same polyp, which will be either "neoplastic" or "hyperplastic". In addition, other polyp features such as the size and location will be recorded, which is similar to what is performed in routine clinical practice. The polyp will be resected and sent for pathological examination, which will form the "gold standard" for the diagnosis of polyp histology.

Device: Computer-aided diagnosis (CADx) support tool
The CADx support tool operates when the Clinician switches the preconfigured CAD EYE function on using a button on the controller while the scope system is in BLI mode. This is performed after the Clinician first makes an optical prediction of polyp histology using IEE as described. The CADx support tool will make a prediction of polyp histology as "hyperplastic" or "neoplastic".

Outcome Measures

Primary Outcome Measures

  1. To evaluate the diagnostic performance of the CADx support tool compared to optical prediction of polyp histology by the Clinician in real-time colonoscopy in a clinical setting [1 year]

    Polyp histology used as gold standard

Secondary Outcome Measures

  1. To determine the diagnostic performance of CADx versus optical prediction of polyp histology by endoscopist in the subgroup analysis [1 year]

    Subgroups include bowel preparation, size of polyp and location

Eligibility Criteria

Criteria

Ages Eligible for Study:
40 Years and Older
Sexes Eligible for Study:
All
Inclusion Criteria:
  1. Patients who have an indication for colonoscopy and who have at least one polyp detected during colonoscopy

  2. 40 years of age and above

  3. Consent obtained for the study

Exclusion Criteria:
  1. Less than 39 years of age

  2. Declined participation in study

  3. Patients with no polyps detected during colonoscopy

  4. Patients with inflammatory bowel disease

  5. Patients with known unresected colorectal cancer

Contacts and Locations

Locations

Site City State Country Postal Code
1 Changi General Hospital, National University Hospital, Singapore General Hospital and Tan Tock Seng Hospital Singapore Singapore 529889

Sponsors and Collaborators

  • Changi General Hospital
  • National University Hospital, Singapore
  • Singapore General Hospital
  • Tan Tock Seng Hospital

Investigators

  • Principal Investigator: Dr James Li, Changi General Hospital, Singapore Health Services

Study Documents (Full-Text)

None provided.

More Information

Publications

Responsible Party:
James Li Weiquan, Consultant, Changi General Hospital
ClinicalTrials.gov Identifier:
NCT05034185
Other Study ID Numbers:
  • CIRB Ref: 2021/2001
First Posted:
Sep 5, 2021
Last Update Posted:
Jul 25, 2022
Last Verified:
Jul 1, 2022
Individual Participant Data (IPD) Sharing Statement:
No
Plan to Share IPD:
No
Studies a U.S. FDA-regulated Drug Product:
No
Studies a U.S. FDA-regulated Device Product:
No
Keywords provided by James Li Weiquan, Consultant, Changi General Hospital
Additional relevant MeSH terms:

Study Results

No Results Posted as of Jul 25, 2022