From Large-Scale Characterization to Subgroup-Specific Predictive Modeling: A Study on the Diagnostic Value of Liver Stiffness Measurements in Focal Liver Lesions

从大规模表征到亚组特异性预测建模:肝脏硬度测量在局灶性肝脏病变诊断价值方面的研究

阅读:1

Abstract

Background/Objectives: As a noninvasive indicator of liver fibrosis and stiffness, liver stiffness measurement (LSM) has also shown significant value in differentiating focal liver lesions (FLLs). This study aimed to assess the characteristics of LSM values across different liver lesions and explore their value in differential diagnosis. Methods: A total of 8817 individuals with FLLs were assessed using liver stiffness measurements (LSMs). We evaluated the LSM characteristics across different FLL categories and further compared these values within subgroups based on their alpha-fetoprotein (AFP) and hepatitis B surface antigen (HBsAg). The LSM was visualized graphically. We compared two logistic regression models (with the LSM and without the LSM) in a cohort of 2271 patients who were both AFP-normal (<20 ng/mL) and HBsAg-negative. The differentiation value of the LSM was quantified by comparing the models' area under the curves (AUCs) and through decision curve analysis (DCA). Results: The LSM showed significant differences (p < 0.001) among malignant lesions, benign lesions, and cirrhotic nodules (CN). Among benign lesions, only focal nodular hyperplasia (FNH) and simple hepatic cysts (SHC) showed a significant difference (p < 0.05). Among malignant lesions, significant differences in the LSM were observed between all pairs (p < 0.001) except between hepatocellular carcinoma (HCC) and combined hepatocellular-cholangiocarcinoma (cHCC-CC). Patients with elevated AFP levels exhibited significantly higher LSM across most lesion types. HBsAg-positive patients also showed significantly increased LSM in all five lesion types, except for CN and cHCC-CC. The full model (with LSM) for differentiating primary malignant lesions from benign ones was built using six variables. The AUCs of the full model were 0.897 and 0.896 in the training and validation sets, significantly outperforming the comparison model (AUC: 0.882, p = 0.0002; 0.879, p = 0.017). Conclusions: The LSM can provide additional information on focal liver lesions.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。