Computer-aided diagnosis improves characterization of Barrett’s neoplasia by general endoscopists (with video)
Jelmer B. Jukema, Carolus H. J. Kusters, Martijn R. Jong, Kiki Fockens, Tim Boers, Joost van der Putten, Roos E. Pouw, Lucas C. Duits, Bas L. Weusten, Lorenza Alvarez Herrero, Martin H. Houben, Wouter B. Nagengast, Jessie Westerhof, Alaa Alkhalaf, Rosalie C. Mallant–Hent, Pieter Scholten, Krish Ragunath, Stefan Seewald, Peter Elbe, Francisco Baldaque Silva, Maximilien Barret, Jacobo Ortiz Fernández‐Sordo, Guiomar Moral Villarejo, Oliver Pech, Torsten Beyna, Nahid S.M. Montazeri, Fons van der Sommen, Peter H. de With, A J de Groof, Jacques Bergman, Alaa Alkhalaf, Lorenza Alvarez Herrero, Francisco Baldaque‐Silva, Maximilien Barret, Jacques Bergman, Torsten Beyna, Raf Bisschops, Tim Boers, Wouter L. Curvers, Pierre H. Deprez, Lucas C. Duits, Peter Elbe, José Miguel Esteban López-Jamar, Gary W. Falk, Kiki Fockens, Gregory G. Ginsberg, Albert J. de Groof, Rehan Haidry, Martin H. Houben, Anthony Infantolino, Prasad G. Iyer, Martijn Jong, Pieter-Jan de Jonge, Jelmer B. Jukema, Arjun D. Koch, Srinadh Komanduri, Vani J. Konda, Koen C. Kusters, Philippe Leclerq, Cadman L. Leggett, Charles J. Lightdale, Rosalie C. Mallant–Hent, Guiomar Moral Villarejo, Jacobo Ortiz Fernández‐Sordo, Oliver Pech, Ian Penman, Roos E. Pouw, Joost van der Putten, Krish Ragunath, Pieter Scholten, Stefan Seewald, Amritha Sethi, Michael Smith, Fons van der Sommen, Arvind J. Trindade, Sachin Wani, Irving Waxman, Jessie Westerhof, Bas L. Weusten, Peter H. de With, Herbert C. Wolfsen
Abstract
BACKGROUND AND AIMS: Characterization of visible abnormalities in patients with Barrett's esophagus (BE) can be challenging, especially for inexperienced endoscopists. This results in suboptimal diagnostic accuracy and poor interobserver agreement. Computer-aided diagnosis (CADx) systems may assist endoscopists. We aimed to develop, validate, and benchmark a CADx system for BE neoplasia. METHODS: The CADx system received pretraining with ImageNet and then consecutive domain-specific pretraining with GastroNet, which includes 5 million endoscopic images. It was subsequently trained and internally validated using 1758 narrow-band imaging (NBI) images of early BE neoplasia (352 patients) and 1838 NBI images of nondysplastic BE (173 patients) from 8 international centers. CADx was tested prospectively on corresponding image and video test sets with 30 cases (20 patients) of BE neoplasia and 60 cases (31 patients) of nondysplastic BE. The test set was benchmarked by 44 general endoscopists in 2 phases (phase 1, no CADx assistance; phase 2, with CADx assistance). Ten international BE experts provided additional benchmark performance. RESULTS: Stand-alone sensitivity and specificity of the CADx system were 100% and 98% for images and 93% and 96% for videos, respectively. CADx outperformed general endoscopists without CADx assistance in terms of sensitivity (P = .04). Sensitivity and specificity of general endoscopists increased from 84% to 96% and 90% to 98% with CAD assistance (P < .001). CADx assistance increased endoscopists' confidence in characterization (P < .001). CADx performance was similar to that of the BE experts. CONCLUSIONS: CADx assistance significantly increased characterization performance of BE neoplasia by general endoscopists to the level of expert endoscopists. The use of this CADx system may thereby improve daily Barrett surveillance.