文献速递腹部US

超声联合CEUS列线图鉴别黑色素瘤肝转移与其他常见肝转移

Development and Internal Validation of a Nomogram Model Discriminating Liver Metastases of Melanoma From Other Common Liver Metastases Based on the Combination of Conventional Ultrasonography and Contrast-enhanced Ultrasonography.
PubMed 文献 · Acad Radiol · 2026-07-11

本研究为单中心回顾性病例对照研究,纳入108例黑色素瘤肝转移(109个病灶)与95例非黑色素瘤肝转移(109个病灶)。基于常规超声和CEUS特征建立列线图模型,AUC为0.892(95%CI:0.849-0.935),经Bootstrap校正后AUC为0.873(95%CI:0.805-0.933)。列线图显示出良好的校准度和临床净获益。

要点

  • 黑色素瘤肝转移与其他肝转移的鉴别可参考形态、后方回声增强、数目、增强模式及廓清时间。
  • 列线图模型AUC为0.892,提示在已知肝转移患者中具有较好的鉴别效能。
  • 该模型可为后续治疗评估与随访提供影像学参考,但需结合病理确诊。

核心结论

  • 基于常规超声及CEUS特征构建的黑色素瘤肝转移鉴别列线图有助于临床筛查补充、精准诊断以及后续治疗评估与随访。

研究设计与样本

  • 单中心、回顾性、病例对照研究,纳入2017年1月至2025年4月共108例黑色素瘤肝转移患者(109个病灶)和95例非黑色素瘤肝转移患者(109个病灶)。
  • 所有病例均经病理诊断证实,将临床特征及常规超声/CEUS特征进行组间单因素分析。

关键影像表现/征象

  • 进入列线图的独立影响因素包括形态、后方回声增强、数目、增强模式及廓清时间。

诊断标准与关键数值

  • P<0.05的变量进入基于AIC的逐步双向回归模型,采用方差膨胀因子评估共线性。
  • 列线图经logistic回归与广义估计方程构建,AUC为0.892(95%CI:0.849-0.935)。
  • 经Bootstrap(1000次重抽样)校正后AUC为0.873(95%CI:0.805-0.933)。
  • Hosmer-Lemeshow检验P>0.05,提示校准度良好;决策曲线分析显示临床净获益。

临床应用与阅片要点

  • 在已知肝转移患者中,可重点关注病灶形态、后方回声、数目、CEUS增强模式及廓清时间以辅助鉴别黑色素瘤肝转移。
  • 不应单独依据超声/CEUS特征确诊黑色素瘤肝转移,仍需结合临床与病理。

局限性

  • 研究为单中心回顾性病例对照设计,纳入病例来自单一机构。

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

RATIONALE AND OBJECTIVES: By comparing conventional ultrasonography (US) and contrast-enhanced ultrasonography (CEUS) features of liver metastases of melanoma (LMM) with those of other common liver metastases from non-melanoma malignancies (non-LMM), a differential diagnosis nomogram was developed and internally validated to explore its clinical application value.

MATERIALS AND METHODS: This single-center, retrospective, case-control study enrolled 108 patients with LMM (109 lesions) and 95 patients with non-LMM (109 lesions) at our institution from January 2017 to April 2025. All cases were confirmed by pathological diagnosis. Univariate analysis was performed to compare clinical characteristics and US/CEUS features between groups. Variables with P < 0.05 were entered into a stepwise bidirectional regression model (based on the Akaike information criterion (AIC)) to identify independent risk factors and assess collinearity (using variance inflation factor (VIF)). A nomogram was developed using logistic regression and generalized estimating equations (GEE) and evaluated by the area under the receiver operating characteristic curve (AUC-ROC), bootstrapping (1000 resamples), calibration curves, and decision curve analysis (DCA).

RESULTS: The independent risk factors included in the nomogram model were morphology, posterior acoustic enhancement, number, enhancement pattern and wash-out time. The model demonstrated an AUC of 0.892 (95% CI: 0.849-0.935) and a bias-corrected AUC of 0.873 (95% CI: 0.805-0.933) by bootstrapping. Both calibration curves and DCA demonstrated the nomogram's favorable calibration and clinical utility (Hosmer-Lemeshow test, P > 0.05).

CONCLUSION: The LMM differential diagnosis nomogram based on US/CEUS features complements clinical screening and facilitates precise diagnosis while aiding in subsequent treatment evaluation and follow-up.

来源:Acad Radiol,查看原文。版权归原作者所有。

引用

  1. Xu, Wu, Tian, et al. Development and Internal Validation of a Nomogram Model Discriminating Liver Metastases of Melanoma From Other Common Liver Metastases Based on the Combination of Conventional Ultrasonography and Contrast-enhanced Ultrasonography. Acad Radiol. 2026. PMID: 42436059. doi:10.1016/j.acra.2026.06.027

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。