Alanna Ebigbo1, Robert Mendel2,3, Tobias Rückert2, Laurin Schuster2, Andreas Probst1, Johannes Manzeneder1, Friederike Prinz1, Matthias Mende4, Ingo Steinbrück5, Siegbert Faiss4, David Rauber2,6, Luis A de Souza2,7, João P Papa7, Pierre H Deprez8, Tsuneo Oyama9, Akiko Takahashi9, Stefan Seewald10, Prateek Sharma11, Michael F Byrne12, Christoph Palm2,3,6, Helmut Messmann1. 1. III Medizinische Klinik, Universitätsklinikum Augsburg, Augsburg Germany. 2. Regensburg Medical Image Computing (ReMIC), Ostbayerische Technische Hochschule Regensburg (OTH Regensburg), Regensburg Germany. 3. Regensburg Center of Health Sciences and Technology (RCHST), OTH Regensburg, Regensburg, Germany. 4. Gastroenterology, Sana Klinikum Lichtenberg, Berlin, Germany. 5. Department of Gastroenterology, Hepatology and Interventional Endoscopy, Asklepios Klinik Barmbek, Hamburg, Germany. 6. Regensburg Center of Biomedical Engineering (RCBE), OTH Regensburg and Regensburg University, Regensburg, Germany. 7. Department of Computing, São Paulo State University, São Paulo, Brazil. 8. Cliniques Universitaires St-Luc, Université Catholique de Louvain, Brussels, Belgium. 9. Saku Central Hospital Advanced Care Center, Nagano, Japan. 10. GastroZentrum, Klinik Hirslanden, Zurich, Switzerland. 11. Department of Gastroenterology and Hepatology, Veterans Affairs Medical Center and University of Kansas School of Medicine, Kansas City, Missouri, United States. 12. Division of Gastroenterology, Vancouver General Hospital, University of British Columbia, Vancouver, British Columbia, Canada.
Abstract
BACKGROUND: The accurate differentiation between T1a and T1b Barrett's-related cancer has both therapeutic and prognostic implications but is challenging even for experienced physicians. We trained an artificial intelligence (AI) system on the basis of deep artificial neural networks (deep learning) to differentiate between T1a and T1b Barrett's cancer on white-light images. METHODS: Endoscopic images from three tertiary care centers in Germany were collected retrospectively. A deep learning system was trained and tested using the principles of cross validation. A total of 230 white-light endoscopic images (108 T1a and 122 T1b) were evaluated using the AI system. For comparison, the images were also classified by experts specialized in endoscopic diagnosis and treatment of Barrett's cancer. RESULTS: The sensitivity, specificity, F1 score, and accuracy of the AI system in the differentiation between T1a and T1b cancer lesions was 0.77, 0.64, 0.74, and 0.71, respectively. There was no statistically significant difference between the performance of the AI system and that of experts, who showed sensitivity, specificity, F1, and accuracy of 0.63, 0.78, 0.67, and 0.70, respectively. CONCLUSION: This pilot study demonstrates the first multicenter application of an AI-based system in the prediction of submucosal invasion in endoscopic images of Barrett's cancer. AI scored equally to international experts in the field, but more work is necessary to improve the system and apply it to video sequences and real-life settings. Nevertheless, the correct prediction of submucosal invasion in Barrett's cancer remains challenging for both experts and AI. Thieme. All rights reserved.
BACKGROUND: The accurate differentiation between T1a and T1b Barrett's-related cancer has both therapeutic and prognostic implications but is challenging even for experienced physicians. We trained an artificial intelligence (AI) system on the basis of deep artificial neural networks (deep learning) to differentiate between T1a and T1b Barrett's cancer on white-light images. METHODS: Endoscopic images from three tertiary care centers in Germany were collected retrospectively. A deep learning system was trained and tested using the principles of cross validation. A total of 230 white-light endoscopic images (108 T1a and 122 T1b) were evaluated using the AI system. For comparison, the images were also classified by experts specialized in endoscopic diagnosis and treatment of Barrett's cancer. RESULTS: The sensitivity, specificity, F1 score, and accuracy of the AI system in the differentiation between T1a and T1b cancer lesions was 0.77, 0.64, 0.74, and 0.71, respectively. There was no statistically significant difference between the performance of the AI system and that of experts, who showed sensitivity, specificity, F1, and accuracy of 0.63, 0.78, 0.67, and 0.70, respectively. CONCLUSION: This pilot study demonstrates the first multicenter application of an AI-based system in the prediction of submucosal invasion in endoscopic images of Barrett's cancer. AI scored equally to international experts in the field, but more work is necessary to improve the system and apply it to video sequences and real-life settings. Nevertheless, the correct prediction of submucosal invasion in Barrett's cancer remains challenging for both experts and AI. Thieme. All rights reserved.