整合CEUS特征与结节位置的改良ACR TI-RADS用于TR4及以上甲状腺结节的诊断与活检决策
改良系统在训练集和验证集中AUC、敏感度均显著高于ACR TI-RADS,并提高恶性结节活检准确率、降低不必要活检率和良性结节活检率,且无安全降级。
要点
- 对ACR TR4及以上结节,CEUS增强特征与结节位置是独立预测因子,可纳入风险分层。
- 改良系统较ACR TI-RADS提高诊断敏感度和AUC,并减少良性结节的不必要活检。
- 验证集显示改良系统AUC达0.95,但尚需外部验证后再用于临床决策。
核心结论
- 整合CEUS特征和结节位置的改良ACR TI-RADS可提高ACR TR4及以上甲状腺结节的诊断性能和FNA推荐效率,待外部验证后可能用于细化风险分层和减少不必要活检。
研究设计与样本
- 采用多变量logistic回归分析CEUS特征和位置特征构建改良系统,并与ACR TI-RADS比较诊断性能和活检推荐效率。
- 训练集334个结节,时间验证集108个结节。
关键影像表现/征象
- 独立预测因子包括增强强度、增强方向、环状增强、增强面积和结节位置。
诊断标准与关键数值
- 训练集改良系统AUC为0.90(95%CI 0.86-0.93),ACR TI-RADS为0.75(95%CI 0.71-0.79)。
- 训练集改良系统敏感度88.5%,ACR TI-RADS为60.3%,P<.01。
- 训练集恶性结节活检准确率(AMB)改良系统67.1%,ACR TI-RADS为49.4%;不必要活检率(UBR)分别为32.9%和50.6%;良性结节活检率(BRB)分别为20.8%和31.2%。
- 验证集改良系统AUC为0.95(95%CI 0.90-0.99),ACR TI-RADS为0.73(95%CI 0.66-0.79);敏感度分别为87.0%和49.3%;AMB分别为83.3%和53.3%;UBR分别为16.7%和46.7%;BRB分别为5.1%和17.9%。
临床应用与阅片要点
- 对ACR TR4及以上结节,除常规灰阶特征外,应关注CEUS增强强度、增强方向、环状增强和增强面积,并结合结节位置进行综合风险评估。
- 改良系统可辅助减少良性结节的不必要穿刺,但原文强调尚无安全降级,应用时需谨慎。
局限性
- 原文指出该改良系统尚待外部验证。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVES: To develop a modified American College of Radiology (ACR) thyroid imaging, reporting and data system (TI-RADS) incorporating contrast-enhanced ultrasound (CEUS) features and nodule location to improve diagnostic performance and fine-needle aspiration (FNA) recommendations for ACR TR4 or higher thyroid nodules.
METHODS: This prospective study enrolled patients who underwent ultrasound and CEUS at 3 hospitals between June 2024 and January 2026. CEUS and location features were analyzed using multivariable logistic regression to construct the modified system. Diagnostic performance and biopsy recommendation efficiency were compared with ACR TI-RADS.
RESULTS: Four hundred and eight patients with 442 nodules were included. Independent predictors included enhancement intensity, enhancement direction, ring enhancement, enhancement area and nodule location. In the training set (n = 334), the modified system achieved higher AUC and sensitivity than ACR TI-RADS (AUC: 0.90 [95% CI: 0.86-0.93] versus 0.75 [95% CI: 0.71-0.79]; sensitivity: 88.5% versus 60.3%; p < .01), higher accuracy of malignant nodule biopsies (AMB: 67.1% versus 49.4%), lower unnecessary biopsy rate (UBR: 32.9% versus 50.6%), and lower biopsy rate of benign nodules (BRB: 20.8% versus 31.2%), with no unsafe downgrading. Similar results were observed in the temporal validation set (n = 108): AUC 0.95 (95% CI: 0.90-0.99) versus 0.73 (95% CI: 0.66-0.79); sensitivity 87.0% versus 49.3%; AMB 83.3% versus 53.3%; UBR 16.7% versus 46.7%, and BRB 5.1% versus 17.9%.
CONCLUSIONS: A modified ACR TI-RADS integrating CEUS features and nodule location improved diagnostic performance and FNA recommendations for ACR TR4 or higher nodules. Pending external validation, this system may provide a practical approach for refining risk stratification and reducing unnecessary biopsies.
引用
- Wang S, Shi X, Ren H, et al. ACR TI-RADS Modified by CEUS Features and Nodule Location for Diagnosis and Biopsy Decision in Thyroid Nodules of TR4 or Higher. J Ultrasound Med. 2026. PMID: 42866996. doi:10.1002/jum.70491
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。