Department of Radiology, The Third Affiliated Hospital of Southern Medical University (Academy of Orthopedics, Guangdong Province), 183 Zhongshan Da Dao Xi, Guangzhou, Guangdong, 510630, China.
Guangdong Provincial Key Laboratory of Medical Image Processing, School of Biomedical Engineering, Southern Medical University, 1023 Shatai Road, Baiyun District, Guangzhou, 510515, Guangdong, China.
Eur J Radiol. 2024 Jul;176:111496. doi: 10.1016/j.ejrad.2024.111496. Epub 2024 May 7.
To develop a deep learning (DL) model for classifying histological types of primary bone tumors (PBTs) using radiographs and evaluate its clinical utility in assisting radiologists.
This retrospective study included 878 patients with pathologically confirmed PBTs from two centers (638, 77, 80, and 83 for the training, validation, internal test, and external test sets, respectively). We classified PBTs into five categories by histological types: chondrogenic tumors, osteogenic tumors, osteoclastic giant cell-rich tumors, other mesenchymal tumors of bone, or other histological types of PBTs. A DL model combining radiographs and clinical features based on the EfficientNet-B3 was developed for five-category classification. The area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity were calculated to evaluate model performance. The clinical utility of the model was evaluated in an observer study with four radiologists.
The combined model achieved a macro average AUC of 0.904/0.873, with an accuracy of 67.5 %/68.7 %, a macro average sensitivity of 66.9 %/57.2 %, and a macro average specificity of 92.1 %/91.6 % on the internal/external test set, respectively. Model-assisted analysis improved accuracy, interpretation time, and confidence for junior (50.6 % vs. 72.3 %, 53.07[s] vs. 18.55[s] and 3.10 vs. 3.73 on a 5-point Likert scale [P < 0.05 for each], respectively) and senior radiologists (68.7 % vs. 75.3 %, 32.50[s] vs. 21.42[s] and 4.19 vs. 4.37 [P < 0.05 for each], respectively).
The combined DL model effectively classified histological types of PBTs and assisted radiologists in achieving better classification results than their independent visual assessment.
开发一种深度学习（DL）模型，用于使用 X 光片对原发性骨肿瘤（PBT）的组织学类型进行分类，并评估其在协助放射科医生方面的临床应用价值。
本回顾性研究纳入了来自两个中心的 878 名经病理证实的 PBT 患者（训练集 638 例，验证集 77 例，内部测试集 80 例，外部测试集 83 例）。我们根据组织学类型将 PBT 分为五类：软骨源性肿瘤、成骨性肿瘤、破骨细胞丰富性巨细胞瘤、骨其他间叶性肿瘤或其他 PBT 的组织学类型。基于 EfficientNet-B3 开发了一种结合 X 光片和临床特征的 DL 模型，用于五类分类。计算受试者工作特征曲线下面积（AUC）、准确率、敏感度和特异度来评估模型性能。通过四位放射科医生的观察者研究评估模型的临床应用价值。
联合模型在内部/外部测试集上的宏观平均 AUC 分别为 0.904/0.873，准确率分别为 67.5%/68.7%，宏观平均敏感度分别为 66.9%/57.2%，宏观平均特异度分别为 92.1%/91.6%。与独立视觉评估相比，模型辅助分析提高了初级（50.6%比 72.3%，53.07s 比 18.55s 和 3.10 比 3.73，在 5 分李克特量表上，P<0.05 均有统计学意义）和高级放射科医生（68.7%比 75.3%，32.50s 比 21.42s 和 4.19 比 4.37，P<0.05 均有统计学意义）的准确率、解读时间和信心。
联合 DL 模型可有效对 PBT 的组织学类型进行分类，并协助放射科医生获得比其独立视觉评估更好的分类结果。
