Cheng Meiying, Tan Shifang, Ren Tian, Zhu Zitao, Wang Kaiyu, Zhang Lingjie, Meng Lingsong, Yang Xuhong, Pan Teng, Yang Zhexuan, Zhao Xin
Department of Radiology, Third Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan, China.
Department of Information, Third Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan, China.
Front Oncol. 2023 Jan 13;12:1073983. doi: 10.3389/fonc.2022.1073983. eCollection 2022.
To evaluate the diagnostic ability of magnetic resonance imaging (MRI) based radiomics and traditional characteristics to differentiate between Ovarian sex cord-stromal tumors (SCSTs) and epithelial ovarian cancers (EOCs).
We consecutively included a total of 148 patients with 173 tumors (81 SCSTs in 73 patients and 92 EOCs in 75 patients), who were randomly divided into development and testing cohorts at a ratio of 8:2. Radiomics features were extracted from each tumor, 5-fold cross-validation was conducted for the selection of stable features based on development cohort, and we built radiomics model based on these selected features. Univariate and multivariate analyses were used to identify the independent predictors in clinical features and conventional MR parameters for differentiating SCSTs and EOCs. And nomogram was used to visualized the ultimately predictive models. All models were constructed based on the logistic regression (LR) classifier. The performance of each model was evaluated by the receiver operating characteristic (ROC) curve. Calibration and decision curves analysis (DCA) were used to evaluate the performance of models.
The final radiomics model was constructed by nine radiomics features, which exhibited superior predictive ability with AUCs of 0.915 (95%CI: 0.869-0.962) and 0.867 (95%CI: 0.732-1.000) in the development and testing cohorts, respectively. The mixed model which combining the radiomics signatures and traditional parameters achieved the best performance, with AUCs of 0.934 (95%CI: 0.892-0.976) and 0.875 (95%CI: 0.743-1.000) in the development and testing cohorts, respectively.
We believe that the radiomics approach could be a more objective and accurate way to distinguish between SCSTs and EOCs, and the mixed model developed in our study could provide a comprehensive, effective method for clinicians to develop an appropriate management strategy.
评估基于磁共振成像（MRI）的放射组学及传统特征对卵巢性索间质肿瘤（SCSTs）和上皮性卵巢癌（EOCs）的鉴别诊断能力。
我们连续纳入了148例患者的173个肿瘤（73例患者中的81个SCSTs和75例患者中的92个EOCs），并以8:2的比例随机分为开发队列和测试队列。从每个肿瘤中提取放射组学特征，基于开发队列进行5折交叉验证以选择稳定特征，并基于这些选定特征构建放射组学模型。采用单因素和多因素分析来确定临床特征和传统MR参数中用于区分SCSTs和EOCs的独立预测因素。并使用列线图来可视化最终的预测模型。所有模型均基于逻辑回归（LR）分类器构建。通过受试者工作特征（ROC）曲线评估每个模型的性能。采用校准和决策曲线分析（DCA）来评估模型的性能。
最终的放射组学模型由9个放射组学特征构建而成，在开发队列和测试队列中的AUC分别为0.915（95%CI：0.869 - 0.962）和0.867（95%CI：0.732 - 1.000），显示出卓越的预测能力。结合放射组学特征和传统参数的混合模型表现最佳，在开发队列和测试队列中的AUC分别为0.934（95%CI：0.892 - 0.976）和0.875（95%CI：0.743 - 1.000）。
我们认为放射组学方法可能是区分SCSTs和EOCs的一种更客观、准确的方法，并且我们研究中开发的混合模型可以为临床医生制定合适的管理策略提供一种全面、有效的方法。
