Early Diagnosis of Hepatocellular Carcinoma by Multiple microRNAs: Validity, Efficacy, and Cost-Effectiveness

Hepatocellular carcinoma (HCC) is a common and aggressive cancer. Chronic infection of hepatitis B virus (HBV) and hepatitis C virus are well-documented major etiologic factors for HCC. Both viruses have been classified as human carcinogens by the International Agency for Research on Cancer. There are approximately 520 to 570 million people who are chronically infected with HBV and hepatitis C virus worldwide, and there is an especially high prevalence in the Asia-Pacific region and Sub-Saharan Africa. Malignant transformation that is induced by chronic HBV infection is a multistage pathogenic process and involves multiple risk predictors. Age, gender, family history of HCC, alcohol consumption habits, serostatus of hepatitis B e antigen, HBV genotype and mutant types, as well as serum quantitative levels of ALT and HBV DNA are important long-term risk predictors of HCC. Nomograms and predictive scores for HCC risk in patients with chronic hepatitis B have recently been developed and validated. Liver surveillance is recommended for chronic HBV carriers with high risk scores to detect HCC as early as possible. Small HCCs can be much better managed by embolization or surgical operations, with a good prognosis. Serologic markers, including -fetoprotein (AFP) and aberrantly methylated DNA, as well as imaging technology, including abdominal ultrasonography, computed tomography, magnetic resonance imaging, and angiograms, have been used for the early detection of HCC. The method of choice for the early diagnosis of HCC should have high validity (sensitivity and specificity), efficacy, and cost-effectiveness. HCC detection by serum AFP level is limited by its low sensitivity, whereas imaging methods are limited by their high cost. There is an urgent need to identify and develop new methods with high accuracy and feasibility for the early diagnosis of HCC. MicroRNAs (miRNAs) are endogenous, small, single-stranded, noncoding RNAs that consist of 20 to 25 bases. They control various biologic functions such as cellular proliferation, differentiation, and apoptosis. Approximately half of human miRNAs are located in fragile regions of chromosomes, which are associated with the development of human cancers. MiRNAs have been referred to as so-called oncomirs for their functions as tumor suppressors and oncogenes. Aberrant expression of miRNA has been linked to a variety of cancers, including HCC. Several studies have explored the use of miRNA expression in liver tissues or serum samples to improve diagnosis or prediction of HCC, as shown in Table 1. However, most previous studies were limited by the small sample size or the small number of miRNAs explored. Several miRNAs were found to be potential diagnostic, prognostic, or metastatic markers. The identified numbers of miRNAs in these studies ranged from 10 to 20—too many to be practically applied for routine clinical use. Moreover, the accuracy of the miRNA signatures was not evaluated adequately, which led to difficulties in reaching a conclusion as to whether the aberrant miRNA expression in liver tissues could reliably differentiate patients with HCC from patients without HCC. One study that explored the diagnostic value of miRNAs in serum found excellent performance using two and three signatures, showing an area under the receiver operating characteristic curve (AUROC) of 0.99 for two signatures and 0.93 for three signatures, respectively. Unfortunately, there was no external validation of the diagnostic accuracy of the signatures in this study. On the basis of the screening of 482 miRNAs with a microarray panel, patients whose HCC tissues had low miR-26a expression were found to have a decreased overall survival. In the article that accompanies this editorial, Zhou et al also found a significantly lower expression level of miR-26a in the serum of patients with HCC compared with control patients who did not have HCC. The results suggest that miR-26a may be a potential signature with clinical significance with respect to diagnosis and prognosis. However, Zhou et al found the diagnostic accuracy of this single marker was not satisfactory, with an AUROC of only 0.67. The authors combined seven miRNAs (miR122, miR-192, miR-21, miR-223, miR-26a, miR-27a, and miR-801) in serum as a diagnostic panel, and found a better diagnostic performance to differentiate patients with HCC from control groups of healthy patients, patients with chronic hepatitis B, and patients with liver cirrhosis in an external validation cohort, as shown in Table 1. This study used an adequate sample size to discover a large number of potential HCC-associated miRNAs in blood samples. The selected candidate miRNAs were additionally validated by quantitative reverse-transcriptase polymerase chain reaction. Not only were a cohort of cases and three groups of controls enrolled to derive the prediction panel of miRNAs, an external validation cohort was also enrolled to validate the diagnostic performance. The miRNA panel had sufficient accuracy to discriminate patients with HCC from healthy controls, patients with chronic hepatitis B, and patients with liver cirrhosis, showing an AUROC (sensitivity and specificity) of 0.94 (83.2% and 93.9%), 0.84 (79.1% and 76.4%), and 0.88 (75.0% and 91.1%), respectively. The best diagnostic performance of the miRNA panel was observed in the comparison between patients with HCC and healthy controls. However, 20% of healthy controls had a serum level of ALT greater than 40 U/L. The diagnostic performance may thus be biased. It could be more accurately estimated if the comparison were made for healthy controls without liver dysfunction. In addition, patients with Editorials