Department of Radiology, First Affiliated Hospital of Hebei North University, Zhangjiakou 075000, China; Department of Radiology, National Clinical Research Center for Cancer, Tianjin Key Laboratory of Cancer Prevention and Therapy, Tianjin's Clinical Research Center for Cancer, Tianjin Medical University Cancer Institute and Hospital, Tianjin 300060, China.
Department of Radiology, National Clinical Research Center for Cancer, Tianjin Key Laboratory of Cancer Prevention and Therapy, Tianjin's Clinical Research Center for Cancer, Tianjin Medical University Cancer Institute and Hospital, Tianjin 300060, China.
Clin Radiol. 2023 Sep;78(9):e635-e643. doi: 10.1016/j.crad.2023.05.004. Epub 2023 Jun 1.
To construct and validate a computed tomography (CT)-based radiomics nomogram integrating radiomics signature and clinical factors to distinguish ovarian cystadenomas and endometriotic cysts.
A total of 287 patients with ovarian cystadenomas (n=196) or endometriotic cysts (n=91) were divided randomly into a training cohort (n=200) and a validation cohort (n=87). Radiomics features based on the portal venous phase of CT images were extracted by PyRadiomics. The least absolute shrinkage and selection operation regression was applied to select the significant features and develop the radiomics signature. A radiomics score (rad-score) was calculated. The clinical model was built by the significant clinical factors. Multivariate logistic regression analysis was employed to construct the radiomics nomogram based on significant clinical factors and rad-score. The diagnostic performances of the radiomics nomogram, radiomics signature, and clinical model were evaluated and compared in the training and validation cohorts. Diagnostic confusion matrices of these models were calculated for the validation cohort and compared with those of the radiologists.
Seventeen radiomics features from CT images were used to build the radiomics signature. The radiomics nomogram incorporating cancer antigen 125 (CA-125) level and rad-score showed the best performance in both the training and validation cohorts with AUCs of 0.925 (95% confidence interval [CI]: 0.885-0.965), and 0.942 (95% CI: 0.891-0.993), respectively. The accuracy of radiomics nomogram in the confusion matrix outperformed the radiologists.
The radiomics nomogram performed well for differentiating ovarian cystadenomas and endometriotic cysts, and may help in clinical decision-making process.
构建并验证一个基于计算机断层扫描（CT）的放射组学列线图，该列线图综合了放射组学特征和临床因素，以区分卵巢囊腺瘤和子宫内膜异位囊肿。
总共 287 名卵巢囊腺瘤（n=196）或子宫内膜异位囊肿（n=91）患者被随机分为训练队列（n=200）和验证队列（n=87）。通过 PyRadiomics 从 CT 图像门静脉期提取放射组学特征。应用最小绝对收缩和选择操作回归（LASSO）选择显著特征并建立放射组学特征。计算放射组学评分（rad-score）。通过显著的临床因素建立临床模型。采用多变量逻辑回归分析建立基于显著临床因素和 rad-score 的放射组学列线图。在训练和验证队列中评估和比较放射组学列线图、放射组学特征和临床模型的诊断性能。计算这些模型在验证队列中的诊断混淆矩阵，并与放射科医生的结果进行比较。
从 CT 图像中提取了 17 个放射组学特征来建立放射组学特征。纳入癌抗原 125（CA-125）水平和 rad-score 的放射组学列线图在训练和验证队列中表现最佳，AUC 分别为 0.925（95%置信区间[CI]：0.885-0.965）和 0.942（95%CI：0.891-0.993）。放射组学列线图在混淆矩阵中的准确性优于放射科医生。
放射组学列线图在区分卵巢囊腺瘤和子宫内膜异位囊肿方面表现良好，可能有助于临床决策过程。
