摘要
目的基于剪切波弹性成像技术(Shear Wave Elastography,SWE)构建甲状腺乳头状癌(Papillary Thyroid Carcinoma,PTC)术后淋巴结颈部转移的预测模型,并探讨其预测价值。方法选取2018年12月至2021年12月于我院行甲状腺根治术治疗的132例PTC患者作为研究对象,根据颈部淋巴结是否出现转移,分为转移组(47例)和非转移组(85例)。采用常规超声的SWE模式杨氏模量测量数据[最大值(E_(max))、最小值(E_(min))、平均值(E_(mean))],记录患者的临床资料。采用多因素Logistic回归分析,以受试者工作特征(Receiver Operating Characteristic,ROC)曲线分析超声测量对患者发生淋巴结颈部转移的预测价值,以R软件构建贝叶斯网络模型并评价其预测效能。结果患者淋巴结转移共发生47例(35.60%)。中央区方面,纵横区>1/2、无淋巴门、血流分布为混乱型/边缘型、E_(mean)增大、E_(max)增大、E_(min)增大为PTC患者术后淋巴结颈部转移的危险因素(P<0.05);颈侧区方面,无淋巴门、血流分布为混乱型/边缘型、E_(mean)增大、E_(max)增大、E_(min)增大为PTC患者术后淋巴结颈部转移的危险因素(P<0.05)。E_(max)、E_(mean)、E_(min)及三者联合的ROC曲线下面积分别为0.885、0.876、0.843、0.896;敏感度分别为86.16%、85.39%、83.34%、87.96%;特异性分别为88.52%、87.36%、86.91%、89.64%;约登指数分别为0.747、0.728、0.703、0.776。贝叶斯网络预测模型显示中央区和颈侧区的纵横区>1/2、无淋巴门、血流分布为混乱型/边缘型与PTC患者术后颈部淋巴结转移存在直接相关性,E_(max)增大、E_(min)增大通过E_(mean)间接影响PTC患者术后颈部淋巴结转移,且贝叶斯网络模型预测的区分度和准确度均较高。结论模型预测基于SWE诊断PTC颈部淋巴结转移具有较高的价值,临床研究可推广使用。
Objective To establish a prediction model of postoperative lymph node neck metastasis in papillary thyroid carcinoma(PTC)based on shear wave elastography(SWE),and to explore its predictive value.Methods A total of 132 PTC patients treated with radical thyroidectomy in our hospital from December 2018 to December 2021 were selected and divided into metastatic group(47 cases)and non-metastatic group(85 cases)according to whether lymph nodes neck showed metastasis.The SWE model measurement data(E_(max),E_(min),E_(mean))of conventional ultrasound was used to record the clinical data of the patients measured by Young’s modulus.Multivariate Logistic regression analysis was used to analyze the predictive value of ultrasound measurements by the receiver operating characteristic(ROC)curve,and the Bayesian network model was constructed with R software and the predictive efficacy was evaluated.Results Total lymph node metastasis was 47 cases(35.60%).In the central area,the risk factors of lymph node neck metastasis in PTC patients were vertical and horizontal area>1/2,no lymphatic gate,blood flow distribution aschaotic/marginal,E_(mean) increasement,E_(max) increasement and E_(min) increasement(P<0.05).In the lateral cervical region,the risk factors of lymph node neck metastasis in PTC patients were no lymphatic gate,blood flow distribution as chaotic/marginal,E_(mean) increasement,E_(max) increasement and E_(min) increasement(P<0.05).The area under ROC curve of E_(max),E_(mean),E_(min) and their combination were 0.885,0.876,0.843,0.896,respectively.The sensitivity was 86.16%,85.39%,83.34%and 87.96%,respectively.The specificity was 88.52%,87.36%,86.91%and 89.64%.Yoden index was 0.747,0.728,0.703,0.776,respectively.Bayesian network prediction model showed the vertical and horizontal area of central area and lateral cervical region>1/2,no lymphatic gate,blood flow distribution as chaotic/marginal were directly correlated with postoperative lymph node neck metastasis in patients with PTC.The E_(max) increasement and E_(min) increasement indirectly affected postoperative lymph node neck metastasis in PTC patients through E_(mean).The Bayesian network model had high differentiation and accuracy.Conclusion The model prediction for the diagnosis of PTC lymph node neck metastasis based on SWE has high value and can be popularized in clinical studies.
作者
葛文华
吕风荣
白雪
GE Wenhua;LV Fengrong;BAI Xue(Department of Ultrasound,General Hospital of Nuclear Industry(The Second Affiliated Hospital of Soochow University),Suzhou Jiangsu 215000,China)
出处
《中国医疗设备》
2023年第3期39-44,共6页
China Medical Devices
作者简介
葛文华,邮箱:18862112696@163.com。
