虽然大部分甲状腺结节是良性的，但其中10%至15%患有癌症。甲状腺超声检查对结节的风险分层是有用的，美国放射学学会甲状腺成像报告和数据系统（ACR TI-RADS）分类根据这些结节的客观超声特征提供了细针穿刺细胞学（FNAC）的建议。此研究旨在验证ACR TI-RADS分类与Bethesda分类和组织病理学的一致性。该回顾性队列研究于2021年3月至2021年5月在新加坡中央医院门诊耳鼻喉科进行。数据分析于2021年5月完成。当ACR TI-RADS建议与Bethesda分数一致时，结果被视为一致。相反，结果被归类为与Bethesda分数和/或组织病理学结果不一致，当建议进行FNAC的结节给出良性结果或未建议进行FNAC的结节给出恶性结果时。共有446名患者（女性370例[83%]；平均[范围]年龄60 [24-89]岁）进行了甲状腺超声检查和超声引导下的甲状腺FNAC。共有630个结节中的492个（78.1%）在FNAC（Bethesda II）中良性。得分为3的ACR TI-RADS结节具有最高的阴性预测值：与Bethesda评分相比为94.6%（95% CI，92.9%-95.9%；P < .001），与组织病理学相比为100.0%（95% CI，15.8%-100.0%；P = .003）。得分为4或5的ACR TI-RADS结节与Bethesda评分相比，阳性预测值分别为2.8%和16.2%，与组织病理学相比为6.1%和66.7%。ACR TI-RADS得分为4和5的小结节（＜1.5cm），如果未建议进行FNAC，则Bethesda 5和6的恶性风险分别为5.7%和25.0%。在手术切除时，46个ACR TI-RADS 4结节中有5个（10.9%）为恶性，21个ACR TI-RADS 5结节中有15个（71.4%）为恶性。在最初未建议进行FNAC的结节中，分别有13个（30.7%）和6个（50.0%）的结节中发现组织病理学证实的癌症。这些发现表明ACR TI-RADS得分为3的结节患癌风险低，仅当结节为2.5 cm或更大时才应考虑进行FNAC。应适当告知患有小型（＜1.5 cm）ACR TI-RADS 4和5结节的患者进行FNAC以排除癌症的可能性。

Although most thyroid nodules are benign, 10% to 15% of them harbor cancer. Thyroid ultrasonography is useful for risk stratification of nodules, and American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) classification provides recommendations for fine-needle aspiration cytology (FNAC) based on objective ultrasonographic features of these nodules.To validate the concordance of ACR TI-RADS classification with Bethesda classification and histopathology.This retrospective cohort study was performed to evaluate the concordance of ACR TI-RADS classification with Bethesda classification and histopathology and was conducted in Singapore General Hospital Outpatient Otolaryngology clinic in March 2021 to May 2021. Data analysis was performed in May 2021.Results were deemed concordant when ACR TI-RADS recommendations aligned with Bethesda scores. Conversely, results were classified as nonconcordant with Bethesda scores and/or histopathology results when nodules that were recommended for FNAC yielded benign results or nodules that were not recommended for FNAC yielded malignant results.A total of 446 patients (370 women [83%]; mean [range] age, 60 [24-89] years) who underwent ultrasonography of the thyroid and ultrasonography-guided thyroid FNACs were identified. A total of 492 of 630 nodules (78.1%) were benign on FNAC (Bethesda II). Score 3 ACR TI-RADS nodules yielded the highest negative predictive values: 94.6% (95% CI, 92.9%-95.9%; P < .001) compared with Bethesda scoring and 100.0% (95% CI, 15.8%-100.0%; P = .003) compared with histopathology. Score 4 or 5 ACR TI-RADS nodules yielded positive predictive values of 2.8% and 16.2%, respectively, compared with Bethesda scoring and 6.1% and 66.7%, respectively, compared with histopathology. Small (<1.5 cm) ACR TI-RADS nodules of scores of 4 and 5 that were not recommended for FNAC yielded a malignant risk of 5.7% and 25.0% on Bethesda 5 and 6, respectively. On surgical excision, 5 of 46 (10.9%) ACR TI-RADS 4 nodules and 15 of 21 (71.4%) of ACR TI-RADS 5 nodules were confirmed to be malignant. Among nodules initially not recommended for FNAC, histopathology-proven cancer was found in 4 of 13 (30.7%) and 3 of 6 (50.0%) of nodules, respectively.These findings suggest that ACR TI-RADS score 3 nodules have a low risk of cancer and should be considered for FNAC only if nodules are 2.5 cm or larger. Patients with small (<1.5 cm) ACR TI-RADS 4 and 5 nodules should be appropriately counseled for FNAC to exclude cancer.