Engineering (Jul 2023)

Serum N-Glycan Markers for Diagnosing Significant Liver Fibrosis and Cirrhosis in Chronic Hepatitis B Patients with Normal Alanine Aminotransferase Levels

  • Lin Wang,
  • Yiqi Liu,
  • Qixin Gu,
  • Chi Zhang,
  • Lei Xu,
  • Lei Wang,
  • Cuiying Chen,
  • Xueen Liu,
  • Hong Zhao,
  • Hui Zhuang

Journal volume & issue
Vol. 26
pp. 151 – 158

Abstract

Read online

The aim of this study was to explore the role of serum N-glycomic-derived models in diagnosing significant liver fibrosis and cirrhosis in 285 chronic hepatitis B (CHB) patients with normal (< 40 IU·L−1) alanine aminotransferase (ALT) levels. Liver biopsies were performed in all enrolled patients, and the stages of liver fibrosis were assessed using the Ishak scoring system. Serum N-glycan profiles were tested using DNA sequencer-assisted fluorophore-assisted carbohydrate electrophoresis (DSA-FACE). A total of nine N-glycan peaks were identified in serum samples for each subject. A machine learning method—namely, random forest (RF) analysis—was adopted to construct more ideal serum N-glycan models in order to distinguish significant liver fibrosis (≥ F3) and cirrhosis (≥ F5). The diagnostic value of the constructed N-glycan models and other fibrotic markers was evaluated. The liver biopsy results revealed that 63.86% (182/285) and 16.49% (47/285) of patients had significant liver fibrosis and cirrhosis, respectively, and 4.91% (14/285) of patients had significant inflammation. In distinguishing significant liver fibrosis, the diagnostic efficiency of the serum N-glycan RF model constructed for distinguishing significant liver fibrosis (≥ F3; RF-A model) was excellent (area under receiver operating characteristic (AUROC) curve: 0.94), and the coincidence rate of the serum N-glycan RF-A model compared with liver biopsy was 90.45%. In distinguishing liver cirrhosis, the diagnostic AUROC curve of the serum N-glycan RF model constructed for distinguishing liver cirrhosis (≥ F5; RF-B model) was 0.97, and the coincidence rate was 88.94%. The diagnostic efficiency of the constructed serum N-glycan models (RF-A and RF-B) was superior to that of liver stiffness measurement (LSM), the fibrosis index based on the four factors (FIB-4), and the aspartate aminotransferase-to-platelet ratio index (APRI). Serum N-glycan models are promising markers for the differentiation of significant liver fibrosis and cirrhosis in CHB patients with normal ALT levels.

Keywords