超声多因素Logistic回归分析鉴别甲状腺良恶性结节的价值

Clinical study of multivariate Logistic regression analysis of ultrasound features for benign and malignant thyroid nodule

目的探讨多因素Logistic鉴别诊断甲状腺结节的价值。方法经超声引导下穿刺活检或手术后病理证实的甲状腺恶性结节32例,良性结节50例,灰阶超声观察结节的位置、大小、形态、边界、回声及血流分布等超声特征;超声造影观察结节增强模式。单因素分析以上特征,将有差异的指标作多因素logistic回归分析。结果单因素分析显示极低回声、边界不清、边缘毛刺、纵横比、微钙化、内部血流丰富、不均性增强鉴别甲状腺良恶性结节差异有统计学意义(P<0.01)。多因素分析显示极低回声、不均匀增强模式及微钙化与甲状腺癌相关(P=0.001,0.012,0.030)。多因素分析诊断甲状腺癌的准确性、敏感性、特异性、阳性预测值和阴性预测值分别为89.0%(73/82)、84.4%(27/32)、92.0%(46/50)、90.2%(46/51)和87.1%(27/31)。结论多因素logistics回归分析能有效提高甲状腺癌诊断的准确性,可以作为综合判断甲状腺结节良恶性的可靠手段。

Objective To investigate the value of ultrasound multiple factor logistic regression analysis in differential diagnosis of thyroid nodules. Methods Thyroid nodules, including 32 benign and 50 malignant nodules, which were confirmed by pathology, were enrolled in the study. We analysed the ultrasound features of location, size, morphology, boundary, echogenicity and blood distribution using gray-scale. Enhanced pattern was evaluated by contrast-enhanced ultrasound (CEUS). Univariate and multivariate analysis was performed to assess the significance of ultrasound features in the diagnosis of nodules. Results On univariate analysis, marked hypoechoic, blurred boundary, spiculated sign, microcalcificaition, longitudinal/transverse >1, abundant internal blood, heterogeneous enhanced pattern had statistically significant difference between benign and malignant thyroid nodules ( P <0.01). On multivariate analysis, marked hypoechoic, heterogeneous enhanced pattern and microcalcificaition correlated with thyroid malignant nodule ( P =0.001, 0.012, 0.030). Based on multiple factors analysis, the accuracy, sensitivity, specificity, positive predictive value and negative predictive value of ultrasound in diagnosis of thyroid carcinoma were 89.0%(73/82), 84.4%(27/32), 92.0%(46/50), 90.2%(46/51) and 87.1%(27/31), respectively. Conclusion Multiple factors analysis can be an efficient method for improving the accuracy of thyroid malignant nodule from benign nodule.