文献速递腹部US

DCE-US应用LI-RADS评估高危肝结节:初步研究与视觉判读比较

Application of DCE-US using the LI-RADS for patients with liver nodules at high risk for hepatocellular carcinoma: A preliminary study and comparison with visual interpretation.
PubMed 文献 · Eur J Radiol · 2022-08-12

本研究纳入220个高危肝癌肝结节,比较视觉判读与基于VueBox软件的DCE-US定量LI-RADS分类的一致性,并以ROC曲线确定的廓清起始截断值作为LR-M的早期廓清时间标准,分析LR-5诊断HCC的性能。结果显示两种判读一致性良好(Kappa=0.730),但DCE-US测得的LR-5廓清起始时间显著短于视觉判读,采用48秒作为LR-M早期廓清标准可提高LR-5诊断HCC的敏感度而不降低特异度。

要点

  • DCE-US定量与视觉LI-RADS判读一致性良好,但廓清起始时间存在显著差异。
  • 以48秒为LR-M早期廓清标准,LR-5诊断HCC敏感度72.8%、特异度95.0%。
  • 48秒标准较60秒标准显著提高LR-5诊断HCC的敏感度。

核心结论

  • DCE-US定量与视觉LI-RADS分类一致性高,但两者评估廓清时间存在差异。
  • 采用DCE-US测得的48秒廓清起始作为LR-M标准,可提高LR-5诊断HCC的敏感度。

研究设计与样本

  • 采用VueBox软件进行DCE-US定量分析,与视觉判读比较LI-RADS分类一致性。
  • 以ROC曲线确定廓清起始截断值作为LR-M早期廓清时间标准,并与60秒标准比较LR-5诊断性能。

关键影像表现/征象

  • 视觉与DCE-US判读的LI-RADS结果一致性良好(Kappa=0.730)。
  • DCE-US测得的LR-5廓清起始时间为82.0±32.8秒,视觉判读为91.2±30.2秒,差异有统计学意义(P=0.020)。
  • 在196个结节中,两种方法检测早期廓清(<60秒)的差异有统计学意义(P=0.047)。

诊断标准与关键数值

  • 鉴别HCC与非HCC恶性肿瘤:AUC 0.85(95%CI 0.8-0.9),Youden指数0.69,截断值48秒,敏感度74.4%,特异度95.0%。
  • 以48秒为LR-M早期廓清标准时,LR-5诊断HCC的敏感度72.8%、特异度95.0%、阳性预测值98.5%。
  • 以60秒为LR-M早期廓清标准时,LR-5诊断HCC的敏感度为62.2%,显著低于48秒标准(P=0.033)。

临床应用与阅片要点

  • 日常阅片可参考DCE-US定量测得的廓清起始时间辅助LI-RADS分类。
  • 对高危肝癌结节,关注48秒内的早期廓清有助于提高LR-5诊断HCC的敏感度。
  • 视觉判读与定量判读在廓清时间上存在差异,报告时应注明所用方法。

局限性

  • 本研究为初步研究,样本来自单中心。
  • 原文未报告其他局限性。

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

OBJECTIVE: To explore the value of dynamic contrast-enhanced ultrasound (DCE-US) for the Liver Imaging Reporting and Data System (LI-RADS).

METHODS: We included 220 liver nodules at high risk for hepatocellular carcinoma (HCC) from January 2019 to October 2021. Visual interpretation and DCE-US-based quantitative categories using VueBox® software were compared for consistency, nodule enhancement intensity, and washout onset following the contrast-enhanced ultrasound LI-RADS. Taking the cut-off value of the ROC curve on washout onset as the time criterion of early washout of LR-M nodules, analyzed the diagnostic performance of LR-5 for HCC and compared it to that of washout onset within 60 s.

RESULTS: The LI-RADS visual and DCE-US interpretation results showed good consistency (Kappa = 0.730), but differed significantly in determining the LR-5 washout onset (91.2 ± 30.2 vs. 82.0 ± 32.8 s, P = 0.020) and detecting early washout (<60 s) in 196 nodules (P = 0.047). Distinguishing HCC from non-HCC malignancies had: area under the ROC curve, 0.85 (95 % confidence interval, 0.8-0.9); Youden index, 0.69; cut-off value, 48 s; sensitivity, 74.4 %; specificity, 95.0 %. When LR-M diagnosis used washout onset within 48 s, LR-5 diagnosis had sensitivity, 72.8 %; specificity, 95.0 %; positive predictive value, 98.5 %; showing a higher sensitivity than with washout onset in 60 s (62.2 %; P = 0.033).

CONCLUSION: DCE-US and visual interpretations showed high consistency in LI-RADS categories but differed in assessing the washout time. According to the DCE-US interpretation, the diagnostic performance of LR-5 could be improved using washout onset of 48 s for LR-M.

来源:Eur J Radiol,查看原文。版权归原作者所有。

引用

  1. Qiu, Ding, Qin, et al. Application of DCE-US using the LI-RADS for patients with liver nodules at high risk for hepatocellular carcinoma: A preliminary study and comparison with visual interpretation. Eur J Radiol. 2022. PMID: 35988395. doi:10.1016/j.ejrad.2022.110473

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。