Romain Leenhardt1, Pauline Vasseur2, Cynthia Li3, Jean Christophe Saurin4, Gabriel Rahmi5, Franck Cholet6, Aymeric Becq1, Philippe Marteau1, Aymeric Histace2, Xavier Dray7. 1. Sorbonne University, Department of Hepato-Gastroenterology, APHP, Saint Antoine Hospital, Paris, France. 2. ETIS, Université de Cergy-Pontoise, ENSEA, CNRS, Cergy-Pontoise Cedex, France. 3. Sorbonne University, Department of Hepato-Gastroenterology, APHP, Saint Antoine Hospital, Paris, France; Drexel University, College of Arts & Sciences, Philadelphia, Pennsylvania, USA. 4. Department of Endoscopy and Gastroenterology, Pavillon L, Hôpital Edouard Herriot, Lyon, France. 5. Georges Pompidou European Hospital, APHP, Department of Gastroenterology and Endoscopy, Paris, France. 6. Digestive Endoscopy Unit, University Hospital, Brest, France. 7. Sorbonne University, Department of Hepato-Gastroenterology, APHP, Saint Antoine Hospital, Paris, France; ETIS, Université de Cergy-Pontoise, ENSEA, CNRS, Cergy-Pontoise Cedex, France.
Abstract
BACKGROUND AND AIMS: GI angiectasia (GIA) is the most common small-bowel (SB) vascular lesion, with an inherent risk of bleeding. SB capsule endoscopy (SB-CE) is the currently accepted diagnostic procedure. The aim of this study was to develop a computer-assisted diagnosis tool for the detection of GIA. METHODS: Deidentified SB-CE still frames featuring annotated typical GIA and normal control still frames were selected from a database. A semantic segmentation images approach associated with a convolutional neural network (CNN) was used for deep-feature extractions and classification. Two datasets of still frames were created and used for machine learning and for algorithm testing. RESULTS: The GIA detection algorithm yielded a sensitivity of 100%, a specificity of 96%, a positive predictive value of 96%, and a negative predictive value of 100%. Reproducibility was optimal. The reading process for an entire SB-CE video would take 39 minutes. CONCLUSIONS: The developed CNN-based algorithm had high diagnostic performances, allowing detection of GIA in SB-CE still frames. This study paves the way for future automated CNN-based SB-CE reading softwares.
BACKGROUND AND AIMS: GI angiectasia (GIA) is the most common small-bowel (SB) vascular lesion, with an inherent risk of bleeding. SB capsule endoscopy (SB-CE) is the currently accepted diagnostic procedure. The aim of this study was to develop a computer-assisted diagnosis tool for the detection of GIA. METHODS: Deidentified SB-CE still frames featuring annotated typical GIA and normal control still frames were selected from a database. A semantic segmentation images approach associated with a convolutional neural network (CNN) was used for deep-feature extractions and classification. Two datasets of still frames were created and used for machine learning and for algorithm testing. RESULTS: The GIA detection algorithm yielded a sensitivity of 100%, a specificity of 96%, a positive predictive value of 96%, and a negative predictive value of 100%. Reproducibility was optimal. The reading process for an entire SB-CE video would take 39 minutes. CONCLUSIONS: The developed CNN-based algorithm had high diagnostic performances, allowing detection of GIA in SB-CE still frames. This study paves the way for future automated CNN-based SB-CE reading softwares.
Authors: Romain Leenhardt; Anthony Buisson; Arnaud Bourreille; Philippe Marteau; Anastasios Koulaouzidis; Cynthia Li; Martin Keuchel; Emmanuele Rondonotti; Ervin Toth; John N Plevris; Rami Eliakim; Bruno Rosa; Konstantinos Triantafyllou; Luca Elli; Gabriele Wurm Johansson; Simon Panter; Pierre Ellul; Enrique Pérez-Cuadrado Robles; Deirdre McNamara; Hanneke Beaumont; Cristiano Spada; Flaminia Cavallaro; Franck Cholet; Ignacio Fernandez-Urien Sainz; Uri Kopylov; Mark E McAlindon; Artur Németh; Gian Eugenio Tontini; Diana E Yung; Yaron Niv; Gabriel Rahmi; Jean-Christophe Saurin; Xavier Dray Journal: United European Gastroenterol J Date: 2019-12-23 Impact factor: 4.623
Authors: Astrid de Maissin; Remi Vallée; Mathurin Flamant; Marie Fondain-Bossiere; Catherine Le Berre; Antoine Coutrot; Nicolas Normand; Harold Mouchère; Sandrine Coudol; Caroline Trang; Arnaud Bourreille Journal: Endosc Int Open Date: 2021-06-21