Background: Thyroid nodule image reporting and data systems (TIRADS) provide the indications for fine-needle aspiration (FNA) based on a combination of nodule sonographic features and size. We compared the TIRADS-based recommendations for FNA with those based on the personal expertise of qualified US investigators in the diagnosis of thyroid malignancy. Methods: Seven highly experienced ultrasound (US) investigators from 4 countries evaluated, online, the US video recordings of 123 histologically verified thyroid nodules. Technical resources provided the operators with a diagnostic approach close to the real-world practice. Altogether, 4,305 TIRADS scores were computed. The combined diagnostic potential of TIRADS (TIRSYS) and the personal recommendations of the investigators (PERS) were compared against 3 possible goals: to recognize all malignant lesions (allCA), nonpapillary plus non-pT1 papillary cancers (nPnT1PCA), or stage II-IV cancers (st2-4CA). Results: For allCA and nPnT1PCA, TIRSYS had lower sensitivity than PERS (69.8 vs. 87.2 and 83.5 vs. 92.6%, respectively, p #x3c;0.01), while in st2-4CA the sensitivities were the same (99.1 vs. 98.6% and TIRSYS vs. PERS, respectively). TIRSYS had a higher specificity than PERS in all 3 types of cancers (p #x3c; 0.001). PERS recommended FNA in a similar proportion of lesions smaller or larger than 1 cm (76.9 vs. 82.7%; ns). Conclusions: Recommendations for FNA based on the investigators’ US expertise demonstrated a better sensitivity for thyroid cancer in the 2 best prognostic groups, while TIRADS methodology showed superior specificity over the full prognostic range of cancers. Thus, personal experience provided more accurate diagnoses of malignancy, missing a lower number of small thyroid cancers, but the TIRADS approach resulted in a similar accuracy for the diagnosis of potentially aggressive lesions while sparing a relevant number of FNAs. Until it is not clearly stated what the goal of the US evaluation is, that is to diagnose all or only clinically relevant thyroid cancers, it cannot be determined whether one diagnostic approach is superior to the other for recommending FNA.
Eur Thyroid J

中文翻译：

---

基于超声的甲状腺细针穿刺适应症：基于 TIRADS 的方法与操作者专业知识的结果

背景：甲状腺结节图像报告和数据系统 (TIRADS) 根据结节超声特征和大小的组合提供了细针穿刺 (FNA) 的适应症。我们将基于 TIRADS 的 FNA 建议与基于美国合格研究人员诊断甲状腺恶性肿瘤的个人专业知识进行比较。方法：来自 4 个国家的 7 名经验丰富的超声 (US) 调查人员在线评估了 123 个经组织学验证的甲状腺结节的美国视频记录。技术资源为运营商提供了一种接近现实世界实践的诊断方法。总共计算了 4,305 个 TIRADS 分数。将 TIRADS (TIRSYS) 的综合诊断潜力和研究人员的个人建议 (PERS) 与 3 个可能的目标进行了比较：识别所有恶性病变 (allCA)、非乳头状加非 pT1 乳头状癌 (nPnT1PCA) 或 II-期IV 癌症 (st2-4CA)。结果：对于 allCA 和 nPnT1PCA，TIRSYS 的敏感性低于 PERS（分别为 69.8 对 87.2 和 83.5 对 92.6%，p #x3c；0.01），而在 st2-4CA 中，灵敏度相同（分别为 99.1 对 98.6% 和 TIRSYS 对 PERS）。在所有 3 种癌症中，TIRSYS 的特异性均高于 PERS（p #x3c；0.001）。PERS 建议在类似比例的小于或大于 1 cm 的病灶中进行 FNA（76.9 对 82.7%；ns）。结论：基于研究人员的美国专业知识对 FNA 的建议表明，在 2 个最佳预后组中对甲状腺癌的敏感性更高，而 TIRADS 方法在整个癌症预后范围内显示出更高的特异性。因此，个人经验提供了更准确的恶性肿瘤诊断，遗漏了较少数量的小型甲状腺癌，但 TIRADS 方法在诊断潜在侵袭性病变时具有相似的准确性，同时保留了相关数量的 FNA。除非明确说明美国评估的目标是什么，即诊断所有或仅诊断临床相关的甲状腺癌，否则无法确定一种诊断方法是否优于另一种推荐 FNA 。
欧洲甲状腺杂志