超声征象Logistic回归分析诊断甲状腺恶性结节

发布时间：2018-05-05 19:56

  本文选题：超声检查 + 甲状腺结节　； 参考：《中国介入影像与治疗学》2017年12期

【摘要】：目的探讨Logistic回归分析在常规超声及CEUS诊断甲状腺良恶性结节中的价值。方法选取经超声引导下穿刺活检或术后病理证实的甲状腺结节患者218例,其中恶性结节74例,良性结节144例,观察结节边界、形态、纵横比、微钙化、回声类型、血流分级以及CEUS增强模式等超声特征,并对其行单因素分析,以有统计学意义的指标为因变量,行多因素Logistic回归分析,建立ROC曲线。结果单因素分析显示低回声、形态不规则、边界不清、纵横比≥1、微钙化、血流分级(Ⅰ、Ⅱ级)、不均匀增强及低增强是诊断甲状腺恶性结节的重要指标(P均0.01)。多因素分析显示形态不规则、微钙化、不均匀增强及低增强是甲状腺恶性结节的独立预测指标(P均0.05)。以Logistic回归模型预测甲状腺恶性结节的准确率为82.57%,ROC曲线下面积为0.906。结论根据甲状腺结节边界、形态、纵横比、微钙化、回声类型、血流分级以及CEUS增强特征建立的Logistic回归模型有助于诊断甲状腺恶性结节。
[Abstract]:Objective to evaluate the value of Logistic regression analysis in the diagnosis of benign and malignant thyroid nodules by conventional ultrasound and CEUS. Methods 218 patients with thyroid nodules confirmed by ultrasound guided biopsy or postoperative pathology were selected, including 74 malignant nodules and 144 benign nodules. The boundary, shape, aspect ratio, microcalcification and echo type of thyroid nodules were observed. Ultrasound features such as blood flow grading and CEUS enhancement mode were analyzed by univariate analysis. The multivariate Logistic regression analysis was performed to establish the ROC curve. Results univariate analysis showed that hypoechoic, irregular shape, unclear boundary, aspect ratio 鈮，

本文编号：1849084