Evaluation of novel LCI CAD EYE system for real time detection of colon polyps

Background Linked color imaging (LCI) has been shown to be effective in multiple randomized controlled trials for enhanced colorectal polyp detection. Recently, artificial intelligence (AI) with deep learning through convolutional neural networks has dramatically improved and is increasingly recognized as a promising new technique for enhancing colorectal polyp detection. Aim This study aims to evaluate a newly developed computer-aided detection (CAD) system in combination with LCI for colorectal polyp detection. Methods First, a convolutional neural network was trained for colorectal polyp detection in combination with the LCI technique using a dataset of anonymized endoscopy videos. For validation, 240 polyps within fully recorded endoscopy videos in LCI mode, covering the entire spectrum of adenomatous histology, were used. Sensitivity (true-positive rate per lesion) and false-positive frames in a full procedure were assessed. Results The new CAD system used in LCI mode could process at least 60 frames per second, allowing for real-time video analysis. Sensitivity (true-positive rate per lesion) was 100%, with no lesion being missed. The calculated false-positive frame rate was 0.001%. Among the 240 polyps, 34 were sessile serrated lesions. The detection rate for sessile serrated lesions with the CAD system used in LCI mode was 100%. Conclusions The new CAD system used in LCI mode achieved a 100% sensitivity per lesion and a negligible false-positive frame rate. Note that the new CAD system used in LCI mode also specifically allowed for detection of serrated lesions in all cases. Accordingly, the AI algorithm introduced here for the first time has the potential to dramatically improve the quality of colonoscopy.

Automated summary

This study evaluated a newly developed computer-aided detection (CAD) system in combination with Linked Color Imaging (LCI) for colorectal polyp detection. The system achieved 100% sensitivity per lesion and a negligible false-positive frame rate, with a 100% detection rate for sessile serrated lesions. The AI algorithm introduced here has the potential to significantly improve the quality of colonoscopy.