Abstract 273: Commonly measured serum liver enzymes as prospective predictors of hepatocellular carcinoma

Abstract

Abstract Introduction: Hepatocellular carcinoma (HCC) is a highly malignant tumor, usually diagnosed at late stages and often has very poor prognosis. Serum liver enzymes (gamma-glutamyltransferase, GGT; alanine aminotransferase, ALT; aspartate aminotransferase, AST; alkaline phosphatase, ALP) are generally used as clinical indicators of liver disease, but their application for HCC is not common. We aimed to prospectively evaluate the association between liver enzyme levels and their combinations with HCC risk in a Western European setting with relatively low prevalence of hepatitis B and/or C (HBV/HCV) and to identify the best discriminatory model for HCC risk prediction. Methods: A nested-case control study was conducted within the European Prospective Investigation into Cancer and Nutrition (EPIC), a large cohort of >520,000 participants from 10 European countries. During mean 7.5 years of follow up, 121 HCC cases were identified each matched to 2 controls by incidence density sampling. Serum biomarkers’ levels of liver enzymes and HBV/HCV status were measured. Crude and multivariable conditional logistic regression models adjusted for relevant confounders were used to calculate odds ratio and 95% confidence intervals (OR, 95%CI) for serum standardised biomarker levels in relation to HCC. Receiver operating characteristics (ROC) curves were used to compare discriminatory accuracy of models with combination of 4 most discriminatory biomarkers (score=GGT+ ALT+AST+ALP), and additionally for alpha-fetoprotein (score+AFP) or HBV/HCV (score+ HBV/HCV) or both (Score+AFP+HBV/HCV), and for score only by time before diagnosis and hepatitis status. Results: In multivariable models, higher levels of enzymes and total bilirubin were associated with a high risk of HCC. The strongest effect was observed for GGT (OR=3.76, 95%CI:2.31-6.13) and ALP (OR=3.11, 95%CI:1.34-2.85). Weaker effect estimates were observed for the score: OR=1.61, 95%CI:1.39-1.88 and OR=1.43, 95%CI:1.23-1.68 for all and hepatitis-free subjects, respectively. The discriminatory accuracy of individual liver enzymes was the highest for GGT (AUC=81%), only slightly increased with number of biomarkers used ranging from 83.4% (score) -84.6% (Score+AFP+HBV/HCV). For the score, the highest accuracy 93%was observed for subjects within the 1st year before cancer diagnosis vs. 82% for longer follow up and in hepatitis positive subjects (AUC=94%). In hepatitis-free subset the model accuracy was still satisfactory (AUC=73%). Conclusion: Our preliminary results showed that prospectively collected high levels of liver enzymes are associated with increased HCC risk. GGT alone or in combination with other serum biomarkers could be employed for HCC risk prediction or early assessment, also in population with low prevalence of HBC/HCV and over extended time before the onset of cancer. Citation Format: Magdalena Stepien, Veronika Fedirko, Talita Duarte-Salles, Antonia Trichopoulou, Pagona Lagiou, Isabelle Romieu, Mazda Jenab. Commonly measured serum liver enzymes as prospective predictors of hepatocellular carcinoma. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 273. doi:10.1158/1538-7445.AM2014-273