Application of a deep learning algorithm in the detection of hip fractures

• Published in: iScience Vol. 26; no. 8; p. 107350
• Main Authors: Gao, Yan, Soh, Nicholas Yock Teck, Liu, Nan, Lim, Gilbert, Ting, Daniel, Cheng, Lionel Tim-Ee, Wong, Kang Min, Liew, Charlene, Oh, Hong Choon, Tan, Jin Rong, Venkataraman, Narayan, Goh, Siang Hiong, Yan, Yet Yen
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 18.08.2023; Elsevier
• Subjects: Computational bioinformatics; Imaging anatomy; Musculoskeletal anatomy; Musculoskeletal anatomy; Computational bioinformatics; Imaging anatomy
• ISSN: 2589-0042; 2589-0042
• DOI: 10.1016/j.isci.2023.107350

This paper describes the development of a deep learning model for prediction of hip fractures on pelvic radiographs (X-rays). Developed using over 40,000 pelvic radiographs from a single institution, the model demonstrated high sensitivity and specificity when applied to a test set of emergency department radiographs. This study approximates the real-world application of a deep learning fracture detection model by including radiographs with sub-optimal image quality, other non-hip fractures, and metallic implants, which were excluded from prior published work. The study also explores the effect of ethnicity on model performance, as well as the accuracy of visualization algorithm for fracture localization. [Display omitted] •Deep learning model developed with >40k images predicts hip fractures on pelvic X-rays•All x-rays included regardless of technical quality, other pathologies or implants•Differs from previous work which tended to have selective criteria for image inclusion•Model achieved high sensitivity (94.2%) and specificity (96.3%) Computational bioinformatics; Imaging anatomy; Musculoskeletal anatomy