Endoscopic prediction of submucosal invasion in Barrett's cancer with the use of artificial intelligence: a pilot study

Autor: Prateek Sharma, David Rauber, Pierre Henri Deprez, Christoph Palm, Tobias Rückert, Helmut Messmann, Akiko Takahashi, Tsuneo Oyama, Stefan Seewald, Ingo Steinbrück, João Paulo Papa, Luis Antonio De Souza, Johannes Manzeneder, Laurin Schuster, Siegbert Faiss, Robert Mendel, Andreas Probst, Alanna Ebigbo, Michael F. Byrne, Friederike Prinz, Matthias Mende
Přispěvatelé: Univ Klinikum Augsburg, Ostbayer TH Regensburg OTH Regensburg, OTH Regensburg, Sana Klinikum Lichtenberg, Asklepios Klin Barmbek, Regensburg Univ, Universidade Estadual Paulista (Unesp), Catholic Univ Louvain, Saku Cent Hosp Adv Care Ctr, Klin Hirslanden, Vet Affairs Med Ctr, Univ Kansas, Univ British Columbia, UCL - SSS/IREC/GAEN - Pôle d'Hépato-gastro-entérologie, UCL - (SLuc) Centre du cancer, UCL - (SLuc) Service de gastro-entérologie
Jazyk: angličtina
Rok vydání: 2020
Předmět:
Zdroj: Web of Science
Repositório Institucional da UNESP
Universidade Estadual Paulista (UNESP)
instacron:UNESP
Endoscopy, Vol. 53, no.9, p. 878-883 (2021)
Popis: Made available in DSpace on 2021-06-26T02:53:52Z (GMT). No. of bitstreams: 0 Previous issue date: 2020-11-16 Bavarian Academic Forum (BayWISS) 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. Univ Klinikum Augsburg, Med Klin 3, Stenglinstr 2, D-86156 Augsburg, Germany Ostbayer TH Regensburg OTH Regensburg, Regensburg Med Image Comp ReMIC, Regensburg, Germany OTH Regensburg, Regensburg Ctr Hlth Sci & Technol RCHST, Regensburg, Germany Sana Klinikum Lichtenberg, Gastroenterol, Berlin, Germany Asklepios Klin Barmbek, Dept Gastroenterol Hepatol & Intervent Endoscopy, Hamburg, Germany OTH Regensburg, Regensburg Ctr Biomed Engn RCBE, Regensburg, Germany Regensburg Univ, Regensburg, Germany Sao Paulo State Univ, Dept Comp, Sao Paulo, Brazil Catholic Univ Louvain, Clin Univ St Luc, Brussels, Belgium Saku Cent Hosp Adv Care Ctr, Nagano, Japan Klin Hirslanden, GastroZentrum, Zurich, Switzerland Vet Affairs Med Ctr, Dept Gastroenterol & Hepatol, Kansas City, MO USA Univ Kansas, Sch Med, Kansas City, MO USA Univ British Columbia, Vancouver Gen Hosp, Div Gastroenterol, Vancouver, BC, Canada Sao Paulo State Univ, Dept Comp, Sao Paulo, Brazil
Databáze: OpenAIRE
Externí odkaz: