摘要
目的 回顾性分析甲状腺影像报告和数据系统4类5类甲状腺结节超声声像图特征并结合弹性成像技术,旨在探寻区分甲状腺结节良恶性的有效方法,为减少不必要的穿刺或手术提供可靠依据。方法 回顾性分析2019年3—12月在本院诊断为ACR TI-RADS 4类5类甲状腺结节患者的临床资料,依据病理分为良性组和恶性组。所有患者均在门诊或住院行甲状腺常规超声和弹性成像检查,短时间内行甲状腺手术、病理检查,收集相关的临床相关资料(性别、年龄),超声相关资料(结节所在位置、结节内部回声、结节大小、结节形态、结节边缘、结节纵横比、是否存在钙化、结节内部及周边血流情况、弹性评分、是否合并桥本甲状腺炎等)。以病理为金标准对相关资料进行整理及分析,将上述12个因素进行单因素分析,并在单因素分析的基础上进行二元Logistic回归模型,应用基于偏最大似然估计的前进法筛选危险因素。分析比较超声声像图中结节的各种表现、超声弹性评分、ACR积分、ACR积分与弹性评分的联合积分在具有可疑恶性征象良性结节和恶性结节中的独立风险因素,根据具有统计学意义的独立风险因素及各因素合并指标,构建受试者工作特性(R0C)曲线,得出曲线下面积,根据约登指数求出各因素的最佳敏感度、特异度。结果 688例患者临床及超声相关资料,共746个结节,其中男148例(21.5%),女540例(78.5%),良性结节142个(19%),恶性结节604个(81%)。超声TI-RADS 4类5类结节中良性结节与恶性结节患者通过单因素分析显示,患者年龄、性别、结节边缘、结节内部回声、结节形态、结节血流信号、结节有无钙化、结节弹性评分、纵横比及是否合并桥本甲状腺炎具有统计学意义(P<0.05)。通过二元logistic回归模型前进法分析显示,年龄、结节回声、形态、钙化、结节弹性评分及合并桥本甲状腺炎及各因素合并指标具有统计学意义(P<0.05)。ROC曲线显示,年龄、结节回声、钙化及合并桥本甲状腺炎曲线下面积小于0.65准确率低。而结节弹性评分、结节形态、ACR积分、ACR积分与弹性评分的联合积分及各因素合并指标的诊断价值高。结论 结节弹性评分、ACR积分、ACR积分与弹性评分的联合积分、结节形态及各因素合并指标对甲状腺结节良恶性具备很好的预测价值。
Objective A retrospective analysis of thyroid imaging reports and data systems,including ultrasound features of four and five types of thyroid nodules,combined with elastography technology,aims to explore effective methods for distinguishing benign and malignant thyroid nodules and provide reliable evidence for reducing unnecessary punctures or surgeries.Methods Retrospective analysis of clinical data of patients diagnosed with ACR TI-RADS class 4 and 5 thyroid nodules in our hospital from March to December 2019,divided into benign and malignant groups based on pathology.All patients underwent routine thyroid ultrasound and elastography examinations in outpatient or inpatient settings,and underwent thyroid surgery and pathological examination in a short period of time.Relevant clinical data(gender,age)and ultrasound related data(location of nodules,internal echoes of nodules,nodule size,nodule morphology,nodule edges,nodule aspect ratio,presence of calcification,blood flow inside and around nodules,elasticity score,and presence of Hashimoto's thyroiditis)were collected Organize and analyze relevant data using pathology as the gold standard,conduct univariate analysis on the 12 factors mentioned above,and use a binary logistic regression model based on univariate analysis to screen risk factors using the forward method based on partial maximum likelihood estimation Analyze and compare various manifestations of nodules in ultrasound images,ultrasound elasticity score,ACR score,and the combined score of ACR score and elasticity score as independent risk factors in benign and malignant nodules with suspected malignant signs.Based on statistically significant independent risk factors and combined indicators of each factor,construct a receiver operating characteristic(R0C)curve,obtain the area under the curve,and calculate the optimal sensitivity and specificity of each factor based on the Jordan index.Results 688 patients with 746 nodules was conducted,including 148 males(21.5%)and 540 females(78.5%),with 142 benign nodules(19%)and 604 malignant nodules(81%).Univariate analysis showed statistically significant differences in age,gender,nodule edge,internal echo,nodule morphology,blood flow signal,calcification,elasticity score,aspect ratio,and Hashimoto's thyroiditis between benign and malignant nodule patients(P<0.05).Binary logistic regression model forward analysis indicated that age,nodule echo,morphology,calcification,elasticity score,and combined Hashimoto's thyroiditis,along with various factor combinations,were statistically significant(P<0.05).The ROC curve revealed that age,nodule echo,calcification,and combined Hashimoto's thyroiditis had low accuracy with areas under the curve less than 0.65.In contrast,the diagnostic value of nodule elasticity scores,nodule morphology,ACR scores,and the combined scores of ACR and elasticity,along with various factor combinations,was high.Conclusions Nodule elasticity scores,ACR scores,combined ACR and elasticity scores,nodule morphology,and various factor combinations have good predictive value for the malignancy of thyroid nodules.
作者
冯峰
王兴田
Feng Feng;Wang Xingtian(The Affiliated Hospital of Xuzhou Medical University,Xuzhou,Jiangsu 221000,China;The Second Affiliated Hospital of Xuzhou Medical University,Xuzhou,Jiangsu 221000,China)
出处
《齐齐哈尔医学院学报》
2025年第4期369-376,共8页
Journal of Qiqihar Medical University
