Value of GPR, APPRI and FIB-4 in the early diagnosis of hepatocellular carcinoma: a prospective cohort study.

OBJECTIVE: There is an urgent need for novel biomarkers that are inexpensive, effective and easily accessible to complement the early diagnosis of hepatocellular carcinoma. This study aimed to analyze the relationship between serum gamma-glutamate-transpeptidase to platelet ratio, alkaline phosphatase-to-platelet ratio index, fibrosis index based on four factors and the risk of hepatocellular carcinoma, and to determine the optimal cut-offs for predicting hepatocellular carcinoma. METHODS: Based on a prospective cohort study, 44 215 participants who were cancer-free at baseline (2011-13) were included in the study. Cox proportional hazard models and receiver operating characteristics curves were used to analyze the diagnostic value and optimal cut-off value of gamma-glutamyl-transpeptidase to platelet ratio, alkaline phosphatase-to-platelet ratio index and fibrosis index based on four factors in predicting hepatocellular carcinoma patients. RESULTS: Gamma-glutamyl-transpeptidase to platelet ratio, alkaline phosphatase-to-platelet ratio index and fibrosis index based on four factors can be used as early independent predictors of hepatocellular carcinoma risk. The risk of hepatocellular carcinoma in the fourth quantile of gamma-glutamyl-transpeptidase to platelet ratio and alkaline phosphatase-to-platelet ratio index was 4.04 times (hazard ratio = 4.04, 95% confidence interval: 2.09, 7.80) and 2.59 times (hazard ratio = 2.59, 95% confidence interval: 1.45, 4.61), respectively, compared with the first quantile. With fibrosis index based on four factors first quantile as a reference, fibrosis index based on four factors fourth quantile had the highest risk (hazard ratio = 18.58, 95% confidence interval: 7.55, 45.72). Receiver operating characteristic results showed that fibrosis index based on four factors had a stronger ability to predict the risk of hepatocellular carcinoma (area under curve = 0.81, 95% confidence interval: 0.80, 0.81), and similar results were shown for gender stratification. In the total population, the optimal cut-off values of gamma-glutamyl-transpeptidase to platelet ratio, alkaline phosphatase-to-platelet ratio index and fibrosis index based on four factors were 0.208, 0.629 and 1.942, respectively. CONCLUSIONS: Gamma-glutamyl-transpeptidase to platelet ratio, alkaline phosphatase-to-platelet ratio index and fibrosis index based on four factors were independent predictors of hepatocellular carcinoma risk. Amongst them, fibrosis index based on four factors shows a stronger predictive ability for hepatocellular carcinoma risk, and gamma-glutamyl-transpeptidase to platelet ratio and alkaline phosphatase-to-platelet ratio index can be used as complementary indicators.