Does high-resolution endoscopy images detect recurrent/residual polyp?
Research type
Research Study
Full title
Does high-resolution endoscopy images detect recurrent/residual polyp? an online questionnaire study
IRAS ID
325998
Contact name
Pradeep Mundre
Contact email
Sponsor organisation
Bradford Teaching Hospitals NHS Foundation Trust
Duration of Study in the UK
0 years, 3 months, 1 days
Research summary
Research summary: During colonoscopy followup after undergoing resection of large polyps, the current advice/guidelines is to take biopsies from the scar to accurately detect recurrence. However this is based on evidence in era of low resolution endoscopy images. However, with latest high resolution images, we should be able to detect recurrence/or non recurrence with just images without need for additional biopsies hence saving valuable resources. Although this is what experts opinion is, strong evidence for this is lacking. \n\nWe devised an online questionnaire(consisting of clinical questions and anonymized images from stock images of post-resection colonoscopy surveillance procedures that are randomly chosen) of 20 questions hosted on a private (free of charge) online platform called Zoho forms. We intend to involve 50 endoscopists from across the country using personal contacts to complete this questionnaire and record their responses. Consent will be assumed if the questionnaire is completed and submitted.\n\nWe intend to analyse the data once we have 50 completed questionaires, which should take approximately 4 months\nThere is no external funding \nDr Pradeep Mundre will be the Chief investigator and Principal investigator for BTHFT site. \n\nStock Images used in questionnaire from colonoscopy procedures are anonymized and will not have any patient identifiable data. Patients would have already routinely consented for use of the images for teaching and research purposes at the time of procedure.\n\nwe will seek approval from the trust Information Governance team and complete privacy impact form from BTHFT Governance team. \nPatients routine care will not be impacted in any way as its a health professional questionnaire study. \nOnce complete , we intend to published data in a reputable peer reviewed journal, so that we can add to the knowledge and influence change in practice in future to save valuable resources. \nSummary of results:\n Fifty-one UK endoscopists of mixed background and experience provided 1,016 optical assessments of post-resection scars spanning all polypectomy techniques (snare, EMR, ESD and hybrid).\nOptical diagnosis achieved an overall negative predictive value (NPV) of 98.3% with an optical miss rate of 3.4% for excluding recurrence.\nPerformance was maintained across subgroups: high-confidence assessments reached an NPV of 98.7% (miss rate 1.6%), and experts 99.7% (2.1%) versus non-experts 97.3% (5.4%); mean participant-level NPV was 98.5% (95% CI 96.9–100.0).\nEvery subgroup exceeded the ASGE PIVI threshold of ≥90% NPV; overall diagnostic accuracy was modest (59.3%), driven predominantly by false positives rather than missed recurrence.\nThese findings indicate that optical diagnosis reliably excludes recurrence across all scar types and experience levels, supporting wider adoption of selective biopsy strategies.\n
REC name
N/A
REC reference
N/A