目的： 通过回顾性研究分析,比较2017年美国放射学会提出的甲状腺成像报告和数据系统（the Thyroid Imaging Report and Data System recommended by American College of Radiology, ACR-TIRADS）与弹性成像单独或联合诊断在甲状腺结节良恶性鉴别中的临床价值。方法： 选取2012年1月至2017年10月在南京中医药大学附属中西医结合医院就诊的503例甲状腺结节患者,共653个结节。收集患者的一般临床资料、甲状腺超声和弹性成像报告及术后病理,对所有结节采用ACR-TITADS、弹性成像的分级法和评分法进行分级,并总结二者联合的4种方法。方法1,弹性成像与ACR-TIRADS中任一诊断为恶性结节,则判定联合诊断为恶性;方法2,弹性成像与ACR-TIRADS共同诊断为恶性结节,则判定联合诊断为恶性;方法3,弹性成像评分分值与ACR-TIRADS分级结果相加;方法4,弹性成像评分分值与ACR-TIRADS计分分值相加。以术后病理为金标准,构建受试者工作特征曲线,计算曲线下面积（area under the curve, AUC）,并根据约登指数选取最佳诊断临界值,比较各方法单独和联合诊断的效能。结果： 随着各诊断方法类别1~4的递增,检出结节的恶性率逐渐增加。ACR-TIRADS诊断甲状腺结节良恶性的AUC有高于弹性超声的趋势,但差异无统计学意义（0.853比0.848,P=0.745）。方法1诊断甲状腺结节良恶性的灵敏度、阴性预测值较高（96.94%、84.85%）;方法2的诊断特异度、阳性预测值较高（88.96%、95.53%）;方法3的灵敏度高于方法4（88.78%比85.92%,P<0.001）;方法4的AUC与方法3之间差异无统计学意义（0.913比0.901,P=0.088）。结论： ACR-TIRADS在甲状腺结节性质诊断中的效能优于弹性成像,而弹性成像评分分值与ACR-TIRADS分级结果相加、弹性成像评分分值与ACR-TIRADS计分分值相加都优于ACR-TIRADS或弹性成像单独诊断,其中方法4在鉴别甲状腺结节良恶性中具有更高的价值。

Objective: To analysis retrospectively the value of The Thyroid Imaging Report and Data System recommended by American College of Radiology (ACR-TIRADS) 2017 and ultrasound elastographyalone or in combination for the differential diagnosis of benign and malignant thyroid nodules. Methods: From January 2012 to October 2017, 503 patients with 653 thyroid nodules underwent preoperative thyroid ultrasonography and thyroidectomy at the Integrated Chinese Traditional Medicine and Western Medicine Hospital were enrolled. General clinical information, thyroid ultrasound report, ultrasound elastography and results of postoperative pathology were collected. All the nodules were graded by ACR-TIRADS, elastography grading and scoring system. The four differential diagnostic categories were as follows:category 1, either elastography or ACR-TIRADS considered as malignant nodules; category 2, both elastography and ACR-TIRADS considered as malignant nodules; category 3, elastography score and grade of ACR-TIRADS added; category 4, elastography score and ACR-TIRADS score added. The postoperative pathological results were regarded as a gold standard to construct the receiver operating-characteristic curve (ROC), calculate the area under the curve (AUC), and the best cutoff value was defined according to the Yoden index. Results: The AUC of ACR-TIRADS was as high as that of ultrasound elastography (0.853 vs. 0.848, P=0.745). Category 1 had a differential diagnosis value with higher sensitivity (96.94%) and the negative predictive value (84.85%). The category 2 had a specificity of 88.96% and a positive predictive value of (95.53%). The sensitivity of category 3 was higher than that of category 4 (88.78% vs. 85.92%, P<0.001). There was no statistical difference between AUC of category 3 of category 4.(0.913 vs. 0.901, P=0.088). Conclusions: ACR-TIRADS is superior to ultrasound elastography in the differential diagnosis of benign and malignant thyroid nodules. Both category 3 and category 4 are superior to ACR-TIRADS or ultrasound elastography alone. The category 4 has a higher discriminatory value in the differential diagnosis of benign and malignant thyroid nodules.