Optical diagnosis of colorectal polyp images using a newly developed computer-aided diagnosis system (CADx) compared with intuitive optical diagnosis

Abstract

Background: Optical diagnosis of colorectal polyps (CRPs) remains challenging. Imaging enhancement techniques such as narrow band imaging and blue light imaging (BLI) can improve optical diagnosis. We developed and prospectively validated a computer-aided diagnosis system (CADx) using high definition white light (HDWL) and BLI images, and compared it with the optical diagnosis of expert and novice endoscopists. Methods: The CADx characterized CRPs by exploiting artificial neural networks. Six experts and thirteen novices optically diagnosed 60 CRPs based on intuition. After a washout period of four weeks, the same set of CRPs was permuted and optically diagnosed using BASIC (BLI Adenoma Serrated International Classification). Results: The CADx had a diagnostic accuracy of 88.3% using HDWL images and 86.7% using BLI images. The overall diagnostic accuracy, combining HDWL and BLI (multimodal imaging), was 95.0% and significantly higher compared to experts (81.7%, p=0.031) and novices (66.5%, p<0.001). Sensitivity (95.6% vs. 61.1% and 55.4%) was also higher for CADx, while specificity was higher for experts compared to CADx and novices (94.1% vs 93.3% and 92.1%). For endoscopists, diagnostic accuracy did not increase using BASIC, neither for experts (Intuition 79.5% vs BASIC 81.7%, p=0.140) nor for novices (Intuition 66.7% vs BASIC 66.5%, p=0.953). Conclusion: The CADx had a significantly higher diagnostic accuracy than experts and novices for the optical diagnosis of CRPs. Multimodal imaging, incorporating both HDWL and BLI, improved the diagnostic accuracy of the CADx. BASIC did not increase the diagnostic accuracy of endoscopists compared to intuitive optical diagnosis.

Applications of Artificial Intelligence for the Diagnosis of Gastrointestinal Diseases

“…In a prospective study by van der Zander et al [ 66 ], CADx diagnostic performances using high-definition white-light (HDWL) and blue-light imaging (BLI) were compared to those of optical diagnosis by expert and novice endoscopists using the BLI Adenoma Serrated International Classification (BASIC) [ 67 ]. The accuracy of AI was 88.3% using HDWL images, 86.7% using BLI images, and 95.0% using both, performing better than experts (81.7%, p = 0.03) and novices (66.7%, p < 0.001).…”

Artificial intelligence for polyp characterization: a challenging road ahead

“…dedicated convolutional neural network (CNN)-based artificial intelligence (AI) systems. In this issue of Endoscopy, van der Zander et al present a study on a novel CNN-based AI system for polyp characterization [6]. The authors showed that, by combining high-definition white-light and blue-light imaging (so-called multimodal imaging), the AI overall diagnostic accuracy was impressive (95.0 %), and higher than that of both experts (81.7 %) and novices (66.7 %).…”

“…As reported in the study by Mori et al [7], within a clinical framework and according to the Preservation and Incorporation of Valuable endoscopic Innovations (PIVI) statements [1], the main outcome of any AI validation study should be clinically oriented. For this purpose, the study by van der Zander et al [6] appears largely underpowered. Second, a tool for everyday clinical practice should demonstrate good discriminatory performance for every polyp histology.…”

“…It has been argued that AI technology has the ability to decrease unwarranted variations and to improve the quality standards of all endoscopists. However, when dealing with AI systems, the relevance of the endoscopist's background (defined as "intuition" by van der Zander et al [6]) and the required training before AI use are likely to represent key factors in need of careful evaluation and definite standardization. Furthermore, there are concerns that nonexperts might passively accept the AI prediction without challenging it, thus reducing the endoscopist to a mere technician.…”

“…As polyp characterization involves quick processing of several images, search for specific visual clues, and recognition of well-defined visual patterns, it represents a fertile ground for Artificial intelligence for polyp characterization: a challenging road ahead dedicated convolutional neural network (CNN)-based artificial intelligence (AI) systems. In this issue of Endoscopy, van der Zander et al present a study on a novel CNN-based AI system for polyp characterization [6]. The authors showed that, by combining high-definition white-light and blue-light imaging (so-called multimodal imaging), the AI overall diagnostic accuracy was impressive (95.0 %), and higher than that of both experts (81.7 %) and novices (66.7 %).…”