Objectives: Liver cirrhosis (LC) is the end-stage of fibrosis in chronic liver diseases, non-invasive early detection of liver fibrosis (LF) is particularly essential for therapeutic decision. Aberrant glycosylation of glycoproteins has been demonstrated to be closely related to liver abnormalities.
Methods: This study was designed to enroll a total of 1,565 participants with LC/LF, chronic hepatitis virus (CHB) and healthy controls. Fibrosis was confirmed by liver biopsy. Using capillary electrophoresis N-glycan fingerprint (NGFP) analysis, we developed a nomogram algorithm (FIB-G) to discriminate LC from non-cirrhotic subjects.
Results: The FIB-G demonstrated good diagnostic performances in identifying LC with the area under the curve (AUC) 0.895 (95%CI: 0.857-0.915). Furthermore, the diagnostic efficiencies of FIB-G were superior to that of log (P2/P8), procollagen III N-terminal (PIIINP), type IV collage (IV-C), laminin (LN), hyaluronic acid (HA), aspartate transaminase to platelets ratio index (APRI), and FIB-4 when detecting significant fibrosis (S0-1 vs. S2-4, AUC: 0.787, 95%CI: 0.701-0.873), severe fibrosis (S0-2 vs. S3-4, AUC: 0.844, 95%CI: 0.763-0.924), and LC (S0-3 vs. S4, AUC: 0.773, 95%CI: 0.667-0.880). Besides, changes of FIB-G were associated well with the regression of fibrosis and liver function Child-Pugh classification.
Conclusions: FIB-G is an accurate multivariant N-glycomic algorithm for LC prediction and fibrosis progression/regression monitoring. The high throughput feasible NGFP using only 2 μL of serum could help physicians make the more precise non-invasive staging of LF or cirrhosis and reduce the need for invasive liver biopsy.
Keywords: N-glycan; liver cirrhosis; liver fibrosis; nomogram; prediction.
© 2020 Walter de Gruyter GmbH, Berlin/Boston.