Artificial Intelligence in screening colonoscopy and error reduction

Abstract

Artificial Intelligence (AI) has the potential to change many aspects of healthcare practice. Image discrimination and classification has many applications within medicine. Machine learning algorithms and complicated neural networks have been developed to train a computer to differentiate between normal and abnormal areas. Machine learning is a form of AI that allows the platform to improve without being programmed. Computer Assisted Diagnosis (CAD) is based on latency, which is the time between the captured image and when it is displayed on the screen. AI-assisted endoscopy can increase the detection rate by identifying missed lesions. An AI CAD system must be responsive, specific, with easy-to-use interfaces, and provide fast results without substantially prolonging procedures. AI has the potential to help both, trained and trainee endoscopists. Rather than being a substitute for high-quality technique, it should serve as a complement to good practice. AI has been evaluated in three clinical scenarios in colonic neoplasms: the detection of polyps, their characterization (adenomatous vs. non-adenomatous) and the prediction of invasive cancer within a polypoid lesion.