Deep learning-based classification of primary bone tumors on radiographs: A preliminary study
Yu He,
Ian Pan,
Bingting Bao,
Kasey Halsey,
Marcello Chang,
Hui Liu,
Shuping Peng,
Ronnie A. Sebro,
Jing Guan,
Thomas Yi,
Andrew T. Delworth,
Feyisope Eweje,
Lisa J. States,
Paul J. Zhang,
Zishu Zhang,
Jing Wu,
Xianjing Peng,
Harrison X. Bai
Affiliations
Yu He
Department of Radiology, The Second Xiangya Hospital of Central South University, No.139 Middle Renmin Road, Changsha, Hunan 410011, PR China
Ian Pan
Department of Diagnostic Imaging, Warren Alpert Medical School of Brown University, Providence 02912, USA
Bingting Bao
Department of Radiology, The Second Xiangya Hospital of Central South University, No.139 Middle Renmin Road, Changsha, Hunan 410011, PR China
Kasey Halsey
Department of Diagnostic Imaging, Warren Alpert Medical School of Brown University, Providence 02912, USA
Marcello Chang
Stanford School of Medicine, Palo Alto 94305, USA
Hui Liu
Department of Radiology, The Second Xiangya Hospital of Central South University, No.139 Middle Renmin Road, Changsha, Hunan 410011, PR China
Shuping Peng
Department of Radiology, The Second Xiangya Hospital of Central South University, No.139 Middle Renmin Road, Changsha, Hunan 410011, PR China
Ronnie A. Sebro
Musculoskeletal Imaging, Department of Radiology, University of Pennsylvania, Philadelphia 19104, USA
Jing Guan
Department of Radiology, The Second Xiangya Hospital of Central South University, No.139 Middle Renmin Road, Changsha, Hunan 410011, PR China
Thomas Yi
Warren Alpert Medical School of Brown University, Providence 02903, USA
Andrew T. Delworth
Brown University, Providence 02912, USA
Feyisope Eweje
Perelman School of Medicine at the University of Pennsylvania, Philadelphia 19104, USA
Lisa J. States
Department of Radiology, Children's Hospital of Philadelphia, 19104, USA
Paul J. Zhang
Department of Pathology and Laboratory Medicine, Hospital of the University of Pennsylvania, Philadelphia 19104, USA
Zishu Zhang
Department of Radiology, The Second Xiangya Hospital of Central South University, No.139 Middle Renmin Road, Changsha, Hunan 410011, PR China
Jing Wu
Department of Radiology, The Second Xiangya Hospital of Central South University, No.139 Middle Renmin Road, Changsha, Hunan 410011, PR China; Corresponding authors.
Xianjing Peng
Department of Radiology, Xiangya Hospital, Central South University, No.87 Xiangya Road, Changsha, Hunan 410008, PR China; Corresponding authors.
Harrison X. Bai
Department of Diagnostic Imaging, Warren Alpert Medical School of Brown University, Providence 02912, USA; Corresponding authors.
Background: To develop a deep learning model to classify primary bone tumors from preoperative radiographs and compare performance with radiologists. Methods: A total of 1356 patients (2899 images) with histologically confirmed primary bone tumors and pre-operative radiographs were identified from five institutions’ pathology databases. Manual cropping was performed by radiologists to label the lesions. Binary discriminatory capacity (benign versus not-benign and malignant versus not-malignant) and three-way classification (benign versus intermediate versus malignant) performance of our model were evaluated. The generalizability of our model was investigated on data from external test set. Final model performance was compared with interpretation from five radiologists of varying level of experience using the Permutations tests. Findings: For benign vs. not benign, model achieved area under curve (AUC) of 0•894 and 0•877 on cross-validation and external testing, respectively. For malignant vs. not malignant, model achieved AUC of 0•907 and 0•916 on cross-validation and external testing, respectively. For three-way classification, model achieved 72•1% accuracy vs. 74•6% and 72•1% for the two subspecialists on cross-validation (p = 0•03 and p = 0•52, respectively). On external testing, model achieved 73•4% accuracy vs. 69•3%, 73•4%, 73•1%, 67•9%, and 63•4% for the two subspecialists and three junior radiologists (p = 0•14, p = 0•89, p = 0•93, p = 0•02, p < 0•01 for radiologists 1–5, respectively). Interpretation: Deep learning can classify primary bone tumors using conventional radiographs in a multi-institutional dataset with similar accuracy compared to subspecialists, and better performance than junior radiologists. Funding: The project described was supported by RSNA Research & Education Foundation, through grant number RSCH2004 to Harrison X. Bai.
