多种CT征象对甲状腺良、恶性结节的预测价值

Predictive value of CT signs in the diagnosis of benign and malignant thyroid nodules

目的 探讨多种CT征象对甲状腺良、恶性结节的预测价值.方法 回顾分析经组织学证实的931例1009枚直径〉1.0 cm甲状腺结节的CT检测资料,包括484例548枚良性结节和447例461枚恶性结节.其中2009—2014年检测的443例患者（485枚结节）为模型组,2015—2016年检测的488例患者（524枚结节）为检测组.观察结节形态不规则、咬饼征、微钙化、增强后范围缩小/模糊、囊变为主和高强化在两组良、恶性结节中的分布差异,并采用单因素和多因素logistic回归分析进行检验,比较两组阳性CT征象对良、恶性结节预测价值的OR值、预测概率曲线下面积.结果模型组和检验组的良、恶性结节分别为252枚（219例）和233枚（224例）、296枚（265例）和228枚（223例）.两组的单因素分析均显示,形态不规则、咬饼征、微钙化、增强后范围缩小/模糊均更常见于恶性结节中,囊变为主和高强化更常见于良性结节中;多因素logistic回归分析显示形态不规则、咬饼征、增强后范围缩小/模糊更常见于恶性结节中（OR值分别为4.172和6.327、3.927和3.493、5.354和6.674）,囊变为主和高强化更常见于良性结节（OR 值分别为11.814和5.082、8.680和14.562）,预测概率曲线下面积值分别为0.946和0.936.微钙化在模型组的多因素logistic回归分析中差异具有统计学意义,在检验组中差异无统计学意义.结论 形态不规则、咬饼征和增强后范围模糊/缩小均是预测恶性结节稳定的CT征象,囊变为主和高强化是预测良性结节稳定的CT征象,多种CT征象联合可提高良、恶性结节的预测价值.

Objective To investigate the predictive value of multiple CT signs in the diagnosis of benign and malignant thyroid nodules.Methods The CT data of 1 009 nodules with diameter 〉1.0 cm confirmed by histology from 931 patients was retrospectively analyzed, including 548 benign nodules from 484 patients and 461 malignant nodules from 447 patients.According to the inspection time, all nodules were divided into model group（2009—2014）and test group（2015—2016）.The distribution of nodules with irregular shape, bite cake syndrome, micro calcification, enhanced range reduction/blur, cystic and high enhancement were evaluated in two groups.Univariate and multivariate logistic regression analysis were performed to evaluate the predicting value of multiple CT sighs for benign and malignant thyroid nodules in two groups.Results The numbers of benign and malignant nodules were 252 and 233 in model group,and were 296 and 228 in test group.The univariate logistic analysis of the model group and the test group showed that irregular shape, bite cake syndrome, micro calcification, enhanced range reduction/blur were more common in malignant nodules.OR values were 4.172 and 6.327,3.927 and 3.493,5.354 and 6.674, 11.814 and 5.082,8.680 and 14.562,respectively.The areas under the predicted probability curve were 0.946 and 0.936, respectively.The micro calcification was an independent predictive factor in the multivariate logistic regression analysis in the model group but not in the test group.Conclusion Irregular shape,bite cake syndrome and enhanced blur/range reduction are the CT signs of the stable malignant nodules,cystic and high enhancement are the CT signs of stable and benign nodules.Combination of multiple CT signs could improve the accuracy of diagnosis for Benign and Malignant Thyroid Nodules.