Serum miRNAs as predictive and preventive biomarker for pre-clinical hepatocellular carcinoma

Abstract

The extremely poor prognosis of patients with symptomatic hepatocellular carcinoma (HCC) diagnosed clinically at advanced stages suggests an urgent need for biomarkers that can be used for prospective surveillance and pre-clinical screening for early presence of pre-malignant lesions and tumors. In a retrospective longitudinal phase 3 biomarker study in seven medical centers of China, time-series and 6 months interval-serum samples were collected from chronic hepatitis B virus infected (CHB) patient cohorts at the pre-malignant or pre-clinical stages (average 6 months prior to clinical diagnosis) and CHB patients that did not develop cancer, and circulating miRNAs measured. A set of serum miRNAs including miR-193a-3p, miR-369-5p, miR-672, miR-429 and let-7i* were identified in pre-clinical HCC patients and have the potential to screen for CHB patients at high risk to develop HCC 6-12 months after miRNAs measurement. These circulating miRNAs combined with the conventional screening tools using α-fetoprotein and ultrasound, may have great promise for the prediction and prevention of HCC in high-risk populations.

Keywords: Biomarker; Circulating miRNAs; Hepatocellular carcinoma; Prediction; Prevention.

Fig. 1.

Retrospectively search for circulating miRNAs candidates as prospective surveillance biomarkers for pre-clinical HCC in high-risk populations. (A) Time series serum samples were recruited for monitor of HCC occurrence. The expression profiles of circulating miRNAs were analyzed in the G1 and G2 groups of patients as described. (B) The geographical distribution of all the patients and medical centers. (C) Design of the experimental procedure, number of patients and follow up intervals included.

Fig. 2.

A set including five serum miRNAs act as an effective screening biomarker for pre-clinical HCC. (A) In the discovery cohort, unique miRNAs that were either up-regulated or down-regulated in G2 relative to G1 were identified, based on calculation with permutation t test (n = 26). (B) ROC curve analysis was performed for the G2 (n = 65) and the G1 (n = 105) patients in the training cohort. (C) The logistic regression equation and parameters created in the training cohort was used to construct the ROC curve in the validation cohort. n = 104 for G2, n = 304 for G1. (D) The proportion of positive results of AFP (cut off = 17.850), miRNAs set (cut off=−0.272), miRNAs set plus AFP (cut off=−0.424) and ultrasound test in all 104 pre-clinical patients (G2) in the validation cohort.

Fig. 3.

The prediagnostic trajectory of serum miRNAs biomarker. (A and B) The rate of positive values for serum miRNAs set (cut off =−0.272) in all 104 pre-clinical patients (G2) in the validation cohort by liver cirrhosis and AFP status (cut off = 17.850). Sixteen-six of 104 pre-clinical HCC patients in the validation cohort were lost the cirrhosis follow up messages in the study interval. (C) Twenty-seven of 35 pre-clinical HCC patients from the Qidong center (center a) in the validation cohort were estimated for serum miRNAs set positives at the cut off value −0.272, in which eighteen patients have spendable multiple sequential samples. Color line: patients with three time points sequential samples, n = 9. Black line: patients with two time points sequential samples, n = 9.

Fig. 4.

A model for circulating miRNAs set action in prospective surveillance and pre-clinical screening of HCC occurrence at the stage when the AFP levels was low and no abnormal ultrasound findings.