Department of Ultrasound, Affiliated Hospital of Guilin Medical University, Guilin, China.
Department of Ultrasound, Affiliated Hospital of Guilin Medical University, Guilin, China.
Ultrasound Med Biol. 2019 Jul;45(7):1627-1637. doi: 10.1016/j.ultrasmedbio.2019.03.014. Epub 2019 May 4.
The clinical importance of thyroid nodules rests with the need to exclude thyroid cancer. In the present study, we developed a modified Thyroid Imaging Reporting and Data System (TI-RADS) score using gray-scale ultrasound, contrast-enhanced ultrasound (CEUS) and shear-wave elastography (SWE) images to predict malignancy of thyroid nodules and compared this modified score system with the subjective scoring criteria based on the Thyroid Imaging Reporting and Data System (TI-RADS, 2017 edition). The results revealed that by using SWE and CEUS (enhanced pattern) to downgrade TI-RADS category 4 and 5 nodules, the malignancy rate for TI-RADS category 4 and 5 nodules increased from 47.6% with American College of Radiology (ACR) TI-RADS assessment alone to 49.4% with ACR TI-RADS combined with shear wave elastography (SWE) and CEUS (enhanced pattern). Likewise, by using the modified TI-RADS to adjust TI-RADS category 3 nodules, the malignancy rate for TI-RADS category 3 nodules increased from 13.9%-20.0%. The discriminating power for detection of malignancy of the variable score 2 (ACR TI-RADS + SWE + CEUS), with an area under the curve (AUC) of 0.899 (95% confidence interval [CI]: 86.1%-93.6%), was higher than that of score 1 (ACR TI-RADS), with an AUC of 0.862 (95% CI: 81.9%-90.6%; p > 0.05). With a point 4.5 as the optimal cutoff value, a score of 1 predicted malignancy with an accuracy of 75.6%, sensitivity of 85.0% and specificity of 71.6%. However, with a point 5.5 as the optimal cutoff value, a score of 2 predicted malignancy with an accuracy of 84.9%, sensitivity of 81.0% and specificity of 86.6%. The modified TI-RADS based on ACR TI-RADS + SWE + CEUS (enhanced pattern) could contribute to a reduction in the number of biopsies performed on benign nodules and the implementation of consistent follow-up in clinical practice.
甲状腺结节的临床重要性在于需要排除甲状腺癌。在本研究中，我们开发了一种改良的甲状腺影像报告和数据系统（TI-RADS）评分，使用灰阶超声、对比增强超声（CEUS）和剪切波弹性成像（SWE）图像来预测甲状腺结节的恶性程度，并将该改良评分系统与基于甲状腺影像报告和数据系统（TI-RADS，2017 年版）的主观评分标准进行了比较。结果表明，通过使用 SWE 和 CEUS（增强模式）对 TI-RADS 4 类和 5 类结节进行降级，单独使用美国放射学会（ACR）TI-RADS 评估时，TI-RADS 4 类和 5 类结节的恶性率从 47.6%增加到 ACR TI-RADS 联合 SWE 和 CEUS（增强模式）的 49.4%。同样，通过使用改良 TI-RADS 调整 TI-RADS 3 类结节，TI-RADS 3 类结节的恶性率从 13.9%-20.0%增加。变量评分 2（ACR TI-RADS+SWE+CEUS）的鉴别力更高，曲线下面积（AUC）为 0.899（95%置信区间[CI]：86.1%-93.6%），高于评分 1（ACR TI-RADS），AUC 为 0.862（95%CI：81.9%-90.6%；p>0.05）。以 4.5 分为最佳截断值，评分 1 预测恶性的准确率为 75.6%，灵敏度为 85.0%，特异性为 71.6%。然而，以 5.5 分为最佳截断值，评分 2 预测恶性的准确率为 84.9%，灵敏度为 81.0%，特异性为 86.6%。基于 ACR TI-RADS+SWE+CEUS（增强模式）的改良 TI-RADS 有助于减少对良性结节进行活检的数量，并在临床实践中实施一致的随访。
