Prediction of liver fibrosis and cirrhosis in chronic hepatitis B infection by serum proteomic fingerprinting

摘要

Aim: Most noninvasive predicti ve models of liver fi brosis are complicated andhave subopti mal sensiti vity. The study described in this paper was designed to identi fy serumproteomic signatures associated with liver fi brosis and to develop a proteome-basedfi ngerprinti ng model for predicti on of liver fi brosis. Methods: Serum proteins from 46 patients with chronic liver hepati ti s B were profi led quanti tati vely on surface-enhanced laserdesorpti on/ionizati on proteinChip arrays. This liver fi brosis-associated proteomic fi ngerprintwas used to construct an arti fi cial neural network (ANN) model that produced a fi brosis indexwith a range of 0-6. Results: Values for 30 serum proteomic features were signifi cantlyassociated with the degree of fi brosis. Conclusions: A unique serum proteomic fi ngerprint ispresent in the sera of pati ents with fi brosis. An ANN fi brosis index derived from this fi ngerprintcould diff erenti ate between diff erent stages of fi brosis and predict fi brosis and cirrhosisin chronic hepati ti s B.