文献速递腹部技术与安全USCTMR

基于多参数超声的慢加急性肝衰竭预后模型

An Optimal Prognostic Model Based on Multiparameter Ultrasound for Acute-on-Chronic Liver Failure.
PubMed 文献 · Ultrasound Med Biol · 2023-07-12

该前瞻性研究纳入149例慢加急性肝衰竭患者,联合常规超声、剪切波弹性成像、衰减成像、彩色多普勒、超微血管成像及超声造影,发现超声造影峰值门静脉流速与肝实质增强的时间间隔TI(PV, LP)与预后独立相关,联合INR建立模型,预测预后AUC为0.851,优于MELD等模型(0.785~0.352),独立验证队列AUC为0.920。

要点

  • 超声造影TI(PV, LP)是ACLF预后的独立超声指标。
  • TI(PV, LP)联合INR的模型预测AUC达0.851,优于MELD。
  • 该模型在独立验证队列AUC为0.920,需多中心大样本验证。

核心结论

  • 基于超声造影TI(PV, LP)与INR建立的预后模型对ACLF预后预测性能优于MELD、LP下降时间模型、2-D SWE列线图及单独TI(PV, LP)。
  • 该模型有潜力作为无创预后预测工具,但尚需多中心大样本研究验证。

研究设计与样本

  • 前瞻性研究,纳入149例慢加急性肝衰竭患者,接受常规超声、2-D剪切波弹性成像、衰减成像、彩色多普勒超声、超微血管成像及超声造影检查。
  • 采用单因素和多因素分析筛选独立超声预后指标,并结合临床特征建立预后模型;另设独立验证队列验证模型性能。

关键影像表现/征象

  • 超声造影肝灌注特征在预后不良组与预后良好组间差异显著。
  • 峰值门静脉流速与肝实质增强之间的时间间隔TI(PV, LP)与ACLF预后独立相关。

诊断标准与关键数值

  • 新模型(TI(PV, LP)联合INR)预测ACLF预后的AUC为0.851。
  • MELD模型AUC为0.785,LP下降时间模型AUC为0.754,2-D SWE列线图AUC为0.708,单独TI(PV, LP)的AUC为0.352。
  • 独立验证队列中新模型AUC为0.920。

临床应用与阅片要点

  • 在ACLF患者超声检查中可关注超声造影峰值门静脉流速与肝实质增强的时间间隔,该指标有助于预后分层。

局限性

  • 研究为单中心前瞻性设计,样本量149例,结论需多中心大样本研究进一步验证。

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

OBJECTIVE: Acute-on-chronic liver failure (ACLF) is associated with a considerably high mortality, and accurate prognosis prediction is critical to navigate intervention decisions and improve clinical outcomes. The objective of this study was to establish a better prognostic model for ACLF based on multiparameter ultrasound in combination with clinical features.

METHODS: A total of 149 patients with ACLF were prospectively enrolled and underwent conventional ultrasound, 2-D shear wave elastography (SWE), attenuation imaging, color Doppler sonography, superb microvascular imaging and contrast-enhanced ultrasound (CEUS). Univariate and multivariate analyses were performed to identify independent ultrasound signatures for the prognosis of ACLF, which, when integrated with clinical characteristics, were used to establish a prognostic model.

RESULTS: Hepatic perfusion features of CEUS differed significantly between the poor and good prognosis groups, among which the time interval (TI) between peak portal vein (PV) velocity and liver parenchyma (LP) enhancement, TI(PV, LP), was independently associated with the prognosis of ACLF. A prediction model comprising TI(PV, LP) and the international normalized ratio was established, and the area under the curve (AUC) was 0.851, which is greater than those of the Model for End-stage Liver Disease (0.785), fall time of LP model (0.754), 2-D SWE nomogram (0.708) and TI(PV, LP) (0.352). Furthermore, the performance of the model was verified in an independent validation cohort (AUC = 0.920).

CONCLUSION: The newly developed model performs better than existing tested models; thus, it has potential as a better non-invasive model for predicting the prognosis of patients with ACLF. A future multicenter, large-sample study is required to validate the performance of this model.

来源:Ultrasound Med Biol,查看原文。版权归原作者所有。

引用

  1. Wu L, Wu Y, Jin J, et al. An Optimal Prognostic Model Based on Multiparameter Ultrasound for Acute-on-Chronic Liver Failure. Ultrasound Med Biol. 2023. PMID: 37451951. doi:10.1016/j.ultrasmedbio.2023.06.014

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。