探讨美国甲状腺协会（ATA）2015年关于成人甲状腺结节与分化型甲状腺癌的诊治指南中甲状腺结节超声分类系统的临床应用价值。

选取2008年1月至2010年12月在北京协和医院行超声引导下细针穿刺抽吸活组织检查（US-FNAB）的480例患者，共483个甲状腺结节，按2015年ATA指南中甲状腺结节的超声分类方法，将结节分为高风险、中等风险、低风险及极低风险4类。112个结节有手术病理结果，其余结节的诊断结合US-FNAB的病理及随访结果。采用独立样本t检验比较良恶性结节的年龄、大小等特征，采用独立样本秩和检验比较不同性别、组成、回声、边界、钙化、纵横比等特征的结节恶性风险。采用独立样本秩和检验比较ATA各分类的结节恶性风险有无显著性差异。并将ATA分类的高风险定为阳性，中等风险、低风险及极低风险定为阴性，计算ATA分类的诊断效能，包括敏感度、特异度、准确性、阳性预测值、阴性预测值。

483个甲状腺结节中，经手术及随访证实，381个（78.9%）为良性，102个（21.1%）为恶性。在结节的超声特征方面，实性、低回声、边缘不规则、有微钙化、纵横比>1的结节恶性风险较高，与对照组比较，差异均有统计学意义（z=-6.255、-6.893、-13.000、-11.080、-6.718、P均<0.001）。在结节的ATA指南超声分类方面，高风险、中等风险、低风险与极低风险的结节恶性实际发生率分别为53.3%（90/169）、5.5%（6/109）、3.9%（6/154）及0（0/51），4个分类之间恶性率比较，差异有统计学意义（χ²=161.462，P<0.001）。将ATA指南超声分类的高风险定为阳性，中等风险、低风险及极低风险定为阴性时，其阴性预测值较高（96.2%）。

2015版ATA指南中甲状腺结节的超声分类系统为甲状腺结节提供了有效的恶性风险分层方法，高风险结节是行US-FNAB的较好的适应证，可获得较高阴性预测值。

To discuss the clinical value of sonographic patterns of thyroid nodules in 2015 American Thyroid Association (ATA) management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer.

From January 2008 to December 2010, 483 thyroid nodules in 480 patients who underwent ultrasound-guided fine-needle aspiration biopsy (US-FNAB) in Peking Union Medical College Hospital were included in this study. Ultrasound images of the thyroid nodules were reviewed and their sonographic patterns were determined as high suspicion, intermediate suspicion, low suspicion and very low suspicion according to 2015 ATA guidelines. 112 nodules had surgical pathology. Diagnosis of other nodules depended on both US-FNAB pathology and follow-up of patients. Independent-sample t test was used to compare ages and sizes between benign and malignant nodules. Independent sample rank sum test was used to compare the malignancy risks between nodules of male and female patients, and between nodules that were solid, hypoechoic, with irregular margins, with microcalcifications and with taller than wide shape and each control group. Independent sample rank sum test was also used to compare malignancy risks of nodules with different sonographic patterns in 2015 ATA guidelines. Defining high-suspicion as positive, and intermediate to very low suspicion as negative, the diagnostic performance of sonographic patterns in 2015 ATA guidelines was calculated, including sensitivity, specificity, positive predictive value, negative predictive value, and accuracy.

Of the 483 thyroid nodules, 381 (78.9%) were benign and 102 (21.1%) were malignant proven by operation and follow-up. The malignancy rates were higher in nodules that were solid, hypoechoic, with irregular margins, with microcalcifications and with taller than wide shape than each control group, all of which had statistically significant differences (z=-6.255, -6.893, -13.000, -11.080, -6.718, P<0.001). Actual malignancy rates of nodules determined as high, intermediate, low and very low suspicion according to the ATA guidelines were 53.3% (90/169), 5.5% (6/109), 3.9% (6/154), and 0 (0/51), respectively, with statistically significant differences between the four patterns (χ²=161.462, P<0.001). When defining high suspicion as positive, and intermediate to very low suspicion as negative, the negative predictive value of sonographic patterns in ATA guidelines was relatively high (96.2%).

Sonographic patterns in 2015 ATA guidelines provide effective malignancy risk stratification for thyroid nodules. ^{High suspicion} is a good indication of US-FNAB for ^{thyroid nodules and has relatively high negative predictive value.}