Open AccessDifferentiating Gastrointestinal Stromal Tumors from Leiomyomas Using a Neural Network Trained on Endoscopic Ultrasonography Images
Author(s) -
Gülseren Seven,
Gökhan Silahtaroğlu,
Ozden Ozluk Seven,
Hakan Şentürk
Publication year - 2021
Publication title -
digestive diseases
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.879
H-Index - 66
eISSN - 1421-9875
pISSN - 0257-2753
DOI - 10.1159/000520032
Subject(s) - medicine , endoscopic ultrasonography , ultrasonography , stromal cell , radiology , leiomyoma , pathology , endoscopy
Endoscopic ultrasonography (EUS) is crucial to diagnose and evaluate gastrointestinal mesenchymal tumors (GIMTs). However, EUS-guided biopsy does not always differentiate gastrointestinal stromal tumors (GISTs) from leiomyomas. We evaluated the ability of a convolutional neural network (CNN) to differentiate GISTs from leiomyomas using EUS images. The conventional EUS features of GISTs were also compared with leiomyomas.