超声定量参数的预测模型在乳腺病变中的应用  

Application of ultrasound quantitative parameter prediction model in breast lesions

作　　者：许仁豪 隋秀芳 裴蓓 张行 张杰[1] Xu Renhao;Sui Xiufang;Pei Bei(Dept of Ultrasound,Affiliated Provincial of Anhui Medical University,Hefei 230001)

机构地区：[1]安徽医科大学附属省立医院超声科,合肥230001

出　　处：《安徽医科大学学报》2021年第5期832-835,共4页Acta Universitatis Medicinalis Anhui

基　　金：安徽省科技厅基金资助项目(编号:1401045018);安徽省高等学校省级自然科学研究项目(编号:KJ2013Z151)。

摘　　要：将241例患者纳入研究,每例患者均有1份乳腺影像学报告及数据系统(BI-RADS)4级的病变经过常规超声、彩色多普勒超声和声脉冲辐射力成像(ARFI)的检查,记录病变的血流阻力指数(RI)、肿块与腺体的剪切波速度(SWV病变、SWV腺体)及其比值(SWV病变/SWV腺体),利用ROC曲线评价各参数诊断性能,再使用分类算法进行分类分析,构建预测模型。经病理证实,241例乳腺病变中140例为恶性、101例为良性。分类算法纳入了RI及SWV病变,深度包括2个分支(当SWV病变≤3.795 m/s,考虑RI≤0.620或0.620<RI≤0.790或RI>0.790;当SWV病变>3.795 m/s,考虑RI≤0.710或>0.710)。分类算法的曲线下面积(AUC)为0.938,灵敏度为98.6%,特异度为57.4%。超声定量参数的预测模型可以明显提高诊断乳腺BI-RADS 4级病变良恶性的准确性,能避免24%乳腺病变的穿刺活检。241 patients were included in this study. Each patient had a BI-RADS 4 breast lesion, and each breast lesion was examined by conventional ultrasound, color doppler ultrasound and ARFI. The resistance index, shear wave velocity(SWV, SWVlesions and SWVglands) and their ratios(SWVlesions/SWVglands) were recorded. The diagnostic performance of each quantitative parameter was evaluated by the subject operation characteristic curve, and then the classification algorithm was used for classification analysis to construct the prediction model. Among the 241 breast masses, 140 cases were malignant and 101 cases were benign. In terms of quantitative indicators RI, SWVlesions and SWVlesions/SWVglands, RI and SWVlesions were included in the classification algorithm, and the depth of the prediction model included two branches(SWVlesions> or ≤3.795 m/s;if SWVlesions≤3.795 m/s, then RI≤0.620 or 0.6200.790, and if SWVlesions>3.795 m/s then RI≤0.710 or > 0.710). The classification algorithm led to an AUC of 0.938, a sensitivity of 98.6%, and a specificity of 57.4%. The prediction model of ultrasonic quantitative parameters can significantly improve the accuracy in the diagnosis of benign and malignant BI-RADS 4 breast lesions, and can avoid the puncture biopsy of 24% breast lesions.

关 键 词：乳腺癌 彩色多普勒超声 弹性成像 穿刺活检 预测模型 

分 类 号：R445.1[医药卫生—影像医学与核医学]