Serum biomarker signature is predictive of the risk of hepatocellular cancer in patients with cirrhosis

Abstract

Background: Inflammatory and metabolic biomarkers have been associated with hepatocellular cancer (HCC) risk in phases I and II biomarker studies. We developed and internally validated a robust metabolic biomarker panel predictive of HCC in a longitudinal phase III study.

Methods: We used data and banked serum from a prospective cohort of 2266 adult patients with cirrhosis who were followed until the development of HCC (n=126). We custom designed a FirePlex immunoassay to measure baseline serum levels of 39 biomarkers and established a set of biomarkers with the highest discriminatory ability for HCC. We performed bootstrapping to evaluate the predictive performance using C-index and time-dependent area under the receiver operating characteristic curve (AUROC). We quantified the incremental predictive value of the biomarker panel when added to previously validated clinical models.

Results: We identified a nine-biomarker panel (P9) with a C-index of 0.67 (95% CI 0.66 to 0.67), including insulin growth factor-1, interleukin-10, transforming growth factor β1, adipsin, fetuin-A, interleukin-1 β, macrophage stimulating protein α chain, serum amyloid A and TNF-α. Adding P9 to our clinical model with 10 factors including AFP improved AUROC at 1 and 2 years by 4.8% and 2.7%, respectively. Adding P9 to aMAP score improved AUROC at 1 and 2 years by 14.2% and 7.6%, respectively. Adding AFP L-3 or DCP did not change the predictive ability of the P9 model.

Conclusions: We identified a panel of nine serum biomarkers that is independently associated with developing HCC in cirrhosis and that improved the predictive ability of risk stratification models containing clinical factors.

Keywords: ALCOHOL; EPIDEMIOLOGY; HEPATITIS C.

Figure 1.

Area under receiver operating curve for HCC at follow-up of 1 year (A) and 2 years (B) shown for base HCC predictive model, base model plus P3, and base model plus P9.

Figure 2.

Calibration plots for HCC predictive models (predicted vs actual HCC) at 2 years of follow-up (A) base model with P3 and (B) base model with P9.

Figure 3.

The cumulative incidence function of HCC stratified by the baseline average risk scores derived from a predictive model that combines clinical and demographic variables with the P9 biomarker panel. The groups are stratified as high (above the median score) and low risk (below median score) for HCC.