Development of 3D training models for the identification and classification of colorectal polyps.

Abstract

Colorectal cancers develop from pre-malignant polyps that can be removed during colonoscopy. Detection, assessment, and removal of polyps has a major role in bowel cancer prevention and is an important part of bowel cancer screening programmes. Trainee colonoscopists must acquire skills to recognise and classify colorectal polyps. Accurate classification is based on morphology, surface pit and capillary patterns. It is difficult to teach assessment skills because static polyp images are often of poor quality and cannot show all areas of interest. Based on anonymised, endoscopic reference images, 3D polyp models were created in ZBrush, demonstrating a variety of morphological forms. The models had detailed pit patterns to show the capillary structure, a key predictor of pathology. The models were subsequently uploaded to the online 3D repository and model viewer, Sketchfab, to create an interactive training resource for trainee colonoscopists. The digital models were evaluated by a panel of expert colonoscopists who scored them for realism and potential as aids for training. There was agreement that the digital polyp models would be useful for teaching. Polyp morphology was rated as realistic however representation of pit patterns received a mixed response, highlighting areas for further development.