Precise anatomical localization and classification of rib fractures on CT using a convolutional neural network

• Published in: Clinical imaging 2022-01, Vol.81, p.24-32
• Main Authors: Zhou, Qing-Qing, Hu, Zhang-Chun, Tang, Wen, Xia, Zi-Yi, Wang, Jiashuo, Zhang, Rongguo, Li, Xinyang, Chen, Chen-Yu, Zhang, Bing, Lu, Lingquan, Zhang, Hong
• Format: Article
• Language: English
• Subjects: Annotations; Artificial intelligence; Artificial neural networks; Classification; Computed tomography; Convolutional neural networks; Datasets; Fractures; Healing; Hospitals; Localization; Medical imaging; Neural networks; Patients; Rib fractures; Training; Trauma; X-ray
• ISSN: 0899-7071; 1873-4499
• DOI: 10.1016/j.clinimag.2021.09.010

To develop a convolutional neural network (CNN) model for the detection, precise anatomical localization (right 1-12th and left 1-12th) and classification (fresh, healing and old fractures) of rib fractures automatically, and to compare the performance with the experienced radiologists. A total of 640 rib fracture patients with 340,501 annotations were retrospectively collected from three hospitals. They consisted of a classification training dataset (n = 482), a localization training dataset (n = 30), an internal testing dataset (n = 90) and an external testing dataset (n = 38). RetinaNet with rib localization postprocessing and the result merging technique were employed to structure the CNN model. ROC curve, free-response ROC curve, AUC, precision, recall, and F1-score were calculated to choose the better option between model I (training classification and localization data together) and model II (adding an additional classification model to model I). The detection and classification performance of rib fractures was better in model II than in model I. The sensitivity of localization reached 97.11% and 94.87% on the right and left ribs, respectively. In the external dataset with different CT scanner and slice thickness, model II showed better diagnostic performance. Moreover, the CNN model was superior in diagnosing fresh and healing fractures to 5 radiologists and consumed shorter diagnosis time. Our CNN model was capable of detection, precise anatomical localization, and classification of rib fractures automatically. •A novel CNN model can automatically detect and output the precise anatomical localization and classification of rib fractures.•The sensitivity of localization reached 94.87% and 97.11% in the right and left ribs, respectively.•The CNN model had a certain robustness verified by the external test, and it matched or surpassed the detection and classification performance of experienced radiologists.