超声影像组学在鉴别腮腺多形性腺瘤与基底细胞腺瘤中的临床价值  

Clinical Value of Ultrasound-Based Radiomics in the Differential Diagnosis of Pleomorphic Adenoma and Basal Cell Adenoma of the Parotid Gland

作　　者：洪龙城 苏唤忠 吴宇卉 张晓东[1] 黄美[1] 张佐炳[1] 余坤 Hong Longcheng;Su Huanzhong;Wu Yuhui;Zhang Xiaodong;Huang Mei;Zhang Zuobing;Yu Kun(Department of Ultrasound,The First Affiliated Hospital of Xiamen University,Xiamen,Fujian 361003,China;Department of Pathology,The First Affiliated Hospital of Xiamen University,Xiamen,Fujian 361003,China)

机构地区：[1]厦门大学附属第一医院超声影像科,福建省厦门市361003 [2]厦门大学附属第一医院病理科,福建省厦门市361003

出　　处：《中国超声医学杂志》2024年第2期136-140,共5页Chinese Journal of Ultrasound in Medicine

基　　金：福建省自然科学基金(No.2023J011617)。

摘　　要：目的探讨超声影像组学在腮腺多形性腺瘤(PA)与基底细胞腺瘤(BCA)鉴别诊断中的价值。方法纳入222例病理诊断为腮腺PA或BCA患者,并分为训练集(130例PA和28例BCA)与验证集(51例PA和13例BCA)。从每个病例超声图像中提取了1316个影像组学特征,经降维筛选后构建超声影像组学评分。同时构建临床+超声模型、超声影像组学模型及超声影像组学评分模型。通过受试者工作特征(ROC)曲线、校准曲线评价模型的预测效能,使用决策曲线分析(DCA)评估临床应用价值。结果选择8个特征构建超声影像组学评分。年龄、形态、超声影像组学评分是鉴别PA与BCA的独立预测变量,基于上述变量构建超声影像组学模型。校准曲线显示超声影像组学模型的一致性较好,训练集与验证集的曲线下面积(AUC)分别为0.892(95%CI:0.834~0.951)、0.878(95%CI:0.793~0.963)。DCA结果表明其临床价值优于其他模型。结论超声影像组学模型可较准确地鉴别PA和BCA,具有优化临床决策的潜力。Objective To explore the value of ultrasound-based radiomics in the differential diagnosis of pleomorphic adenoma(PA)and basal cell adenoma(BCA)of the parotid gland.Methods A total of 222 patients with pathological diagnosis of PA or BCA of the parotid gland were enrolled in training cohort(130 PA cases and 28 BCA cases)and validation cohort(51 PA cases and 13 BCA cases).Overall,1316 radiomics features were extracted from ultrasound images of each case.Then a ultrasound-based radiomics score was constructed after feature dimensionality reduction and screening.Next,a clinic-ultrasound model,an ultrasound-based radiomics model,and an ultrasound-based radiomics score model were constructed.The predictive efficacy of the models were assessed by the calibration curve,receiver operating characteristic(ROC)curve,and the clinical usefulness was judged by decision curve analysis(DCA).Results Eight features from ultrasound images were used to build the ultrasound-based radiomics score.Age,shape and ultrasound-based radiomics score were independent predictors for distinguishing PA from BCA,and ultrasound-based radiomics model were constructed based on the above predictors.The calibration curves showed that the model had good calibration.The area under the curve(AUC)of the ultrasound-based radiomics model was 0.892(95%CI:0.834-0.951)and 0.878(95%CI:0.793-0.963)in the training cohort and validation cohort,respectively.DCA demonstrated that the model was clinically useful.Conclusions Ultrasound-based radiomics model differentiated PA from BCA reasonably accurately.It holds potential for optimizing the clinical decision-making process.

关 键 词：超声影像组学 腮腺 多形性腺瘤 基底细胞腺瘤 鉴别诊断 

分 类 号：R445.1[医药卫生—影像医学与核医学] R739.87[医药卫生—诊断学]