Overview
Colorectal cancer prevention relies on high-quality colonoscopy and accurate optical characterization of colorectal polyps. Optical diagnosis may allow optimization of management strategies such as resect-and-discard for diminutive polyps, potentially improving efficiency in routine practice.
COLO-PREDICT is a prospective, single-center, observational study designed to evaluate the impact of a structured digital training program (Colo-ID, a digital application for colonic polyp characterisation training) on the accuracy of optical histology prediction of colorectal polyps in real-life clinical practice.
All consecutive adult patients undergoing colonoscopy with at least one detected polyp will be included. Optical prediction of polyp histology will be recorded during the procedure. All polyps will be resected and sent for histopathological analysis according to standard practice. No modification of patient management will occur as part of the study.
The primary objective is to assess the concordance between optical prediction and histopathology before and after implementation of the digital training program. Secondary objectives include evaluation of prediction performance according to polyp characteristics and assessment of potential implications for clinical practice.
Description
Colorectal cancer (CRC) prevention depends on the detection and removal of precursor lesions during colonoscopy. Optical diagnosis of colorectal polyps using enhanced imaging technologies has been proposed to improve procedural efficiency and to potentially support management strategies such as resect-and-discard for diminutive lesions. However, variability in optical characterization accuracy remains a major limitation in routine practice.
COLO-PREDICT is a prospective, single-center, observational study conducted in real-life clinical practice. The study aims to evaluate the impact of a structured digital training program (Colo-ID, a digital application for colonic polyp characterisation training) on the accuracy of optical histologic prediction of colorectal polyps.
Adult patients undergoing colonoscopy with at least one detected polyp will be consecutively included. During each procedure, the endoscopist will record an optical prediction of polyp histology based on standard classification systems and routine imaging modalities. All polyps will be resected and submitted for histopathological examination in accordance with current standard of care. No modification of patient management, surveillance interval, or therapeutic strategy will be implemented as part of the study.
The digital training program consists of a structured educational intervention focused on optical characterization principles, image-based training, and performance assessment in a simulated environment. The study evaluates optical prediction performance before and after implementation of this training program in routine practice.
The primary endpoint is the rate of concordance between optical histologic prediction and final histopathology. Secondary endpoints include diagnostic performance parameters (sensitivity, specificity, positive predictive value, negative predictive value), performance according to polyp size and location, and assessment of potential implications for clinical practice efficiency.
Data will be collected prospectively and analyzed using appropriate statistical methods to compare performance before and after training. The study is conducted in accordance with ethical and regulatory requirements and does not introduce additional risk for patients, as all clinical decisions follow standard practice.
Eligibility
Inclusion Criteria:
- Adults aged ≥18 years
- Patients undergoing colonoscopy in routine clinical practice at the participating center
- At least one colorectal polyp detected during colonoscopy
- Optical histology prediction recorded during the procedure
- Histopathological analysis available for resected polyps (reference standard)
Exclusion Criteria:
- Patients aged \<18 years
- Refusal/opposition to use of data (if applicable under local regulations)
- Missing key study data (e.g., no recorded optical prediction or no histopathology result)
