Clinical and molecular classification of hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is a major health problem worldwide,1 accounting for more than 626,000 new cases per year.2 The incidence of HCC is increasing in the United States2 and Europe, and it is currently the third cause of cancer-related death globally, behind only lung and colon cancers.3 Chronic hepatitis B virus infection is the main risk factor in Asia and Africa, and hepatitis C virus infection is the main risk factor in western countries and Japan. Cancer classification seeks to establish prognosis and to select the adequate treatment for the best candidates. In addition, it aids researchers in the exchange of information and the design of clinical trials, to permit criteria to be compared. Clinical classifications have been proposed for most cancers. However, few include molecular data. Such is the case for breast cancer, where Her2/nu status discriminates subgroups of patients with different outcome and treatment response.4 Similarly, EGFR mutation status in non–small cell lung cancer identifies a subgroup of patients whose disease will respond to tyrosine kinase inhibitors.5 Clinical classification of HCC is currently available, but no molecular data have been incorporated so far. Specific knowledge of the molecular classification of HCC is preliminary, and further studies are required. Molecular data can be useful to further discriminate the best candidates for liver transplantation. The knowledge of the prognosis factors of HCC may help predict patient outcomes, may help in the design of research investigations, and may provide the basis for a classification of the disease. Prognosis assessment is particularly complex in HCC considering that variables of 2 diseases—cirrhosis and cancer—are involved in up to 80% of cases. The key prognostic predictors are only partially known, and they vary at different stages of the disease. The main prognostic factors are related to tumor status (defined by number and size of nodules, presence of vascular invasion, and extrahepatic spread), liver function (defined by Child-Pugh class, bilirubin, albumin, and portal hypertension), and general health status (defined by Eastern Cooperative Oncology Group [ECOG] classification and presence of symptoms). The cause of HCC has not been identified as an independent prognostic factor. These factors should be incorporated into staging systems in HCC. The Barcelona Clinic Liver Cancer (BCLC) classification has emerged as the standard classification for trial design and clinical management of HCC.1, 6 This classification is endorsed by the European Association for the Study of the Liver (EASL) and the American Association for the Study of the Liver (AASLD) guidelines,6, 7 and it has been externally validated in European and American patient cohorts.8, 9 It links stage stratification with a recommended treatment strategy. Treatment is based on criteria that follow the AASLD guidelines,6 which defines standard of care for each tumor stage. In brief, patients at stage 0 with very early HCC (tumors <2 cm in diameter, Child-Pugh class A) and patients at stage A with early HCC (single or 3 nodules <3 cm, Child-Pugh class A–B) are candidates for radical therapies (resection, liver transplantation, or percutaneous treatments) according to a well-established schedule.1, 10 Untreated patients at an intermediate stage (BCLC stage B; multinodular asymptomatic tumors without an invasive pattern) have a median survival of approximately 16 months. Chemoembolization is associated with a median survival of 19–20 months in randomized, controlled trials11, 12 and is considered the standard of care.6 Untreated patients with advanced stage disease (BCLC class C; i.e., ECOG performance status 1–2, and/or vascular invasion or extrahepatic spread) have a median survival of approximately 6–7 months. Recently, sorafenib, a multi–tyrosine kinase inhibitor, has demonstrated survival benefits for patients with advanced HCC stage disease (median overall survival 10.7 months for sorafenib vs. 7.9 months for placebo, P = 0.0006),13 and it is likely to become the new systemic standard therapy for HCC patients. Finally, patients with stage D lesions and with end-stage disease will receive treatment for their symptoms. There is no doubt that the limitations of the classical staging systems have already been overcome. The Okuda staging and the Child-Pugh classification might be used as a part of any new clinical staging system, but they should no longer be used alone. Among the new classifications, however, the heterogeneous survival figures described for the best stages (3-year survival of 25–80%) reflect that some studies include mostly advanced cases with a small number of effectively treated patients. The Chinese University Prognostic Index (CUPI), Cancer of the Liver Italian Program Investigators (CLIP), and French staging systems14-16 have been constructed mostly with patients with advanced-stage disease, and they represent score systems that predict outcomes for these patients. They use rough descriptions of tumor stage that are not in accordance with the predictive value of tumor size and multicentricity. For instance, the CLIP score classifies the tumor burden as above or below 50% of liver involvement, thus by definition making it impossible to identify patients at early stages. The new tumor, node, metastasis (TNM) staging system according with the American Joint Committee on Cancer (AJCC) has only internal validation; it is based on series of patients undergoing resection.17 Pathological information is needed in all cases, which is a limitation for its wide clinical use. Finally, the Japan Integrated Staging (JIS)18 is a new scoring system that includes 2 previous classifications: the TNM endorsed by the International Union Against Cancer (UICC), mostly applied in Japan; and Child-Pugh classification. A recent validation of the score in >4,500 patients has shown it to be better than the CLIP score. HCC, hepatocellular carcinoma; BCLC, Barcelona Clinic Liver Cancer; EASL, European Association for the Study of the Liver; AASLD, American Association for the Study of the Liver; CUPI, Chinese University Prognostic Index; CLIP, Cancer of the Liver Italian Program Investigators; TNM, tumor, node, metastasis system; AJCC, American Joint Committee on Cancer; JIS, Japan Integrated Staging; UICC, International Union Against Cancer. There is no molecular classification of HCC.19 Global gene expression profiling may be the most appropriate technology to unravel the complex pathogenesis of HCC and explore its heterogeneous origin. In fact, application of gene expression profiling of HCC has identified subgroups of patients according to etiological factors, different stages of the disease, disease recurrence, and survival20-25 (Table 1). Other molecular classifications based on activation of specific signal transduction pathways have been recently reported.26 These investigations represent promising progress in the use of gene expression profiling in elucidating the molecular pathogenesis of HCC and in improving the prognostic prediction for HCC patients. Prediction of survival by gene signatures is limited because the patients die not only from tumor progression, but also from liver failure. Thus, accurate selection of the cohort and/or use of cancer-related death as an end point to minimize this bias should be a priority. A gene signature able to discriminate 2 populations of good and poor survival was reported in hepatitis B virus–infected patients.20 Application of a knowledge-based annotation of the 406 genes revealed molecular pathways responsible for the biological differences observed in the 2 subclasses of HCC. Measurement of cell proliferation, apoptosis, ubiquitination, and histone modification provided the best quantitative separation of the 2 survival subclasses. More recently, the same group reported one additional HCC subgroup of patients with a grim prognosis presenting an hepatoblastoma-like signature assumed to be of progenitor cell origin.21 Another study analyzed the expression profiles of 67 primary and metastatic HCC samples from 40 patients. By means of a supervised machine-learning algorithm, the authors generated a 153-gene molecular signature that permitted classification of metastatic HCC patients and identified genes that were relevant to patient survival.22 Even with optimal selection of candidates for resection or local ablation, tumor recurrence complicates 50% of cases at 3 years, comprising previously undetected metastases and de novo tumors.10 Pathological variables, including vascular invasion, poor histological differentiation, and satellites, predict metastases.10 Biological markers that predict aggressive behavior are not well defined. Two studies have used gene expression profiling to address the issue of HCC recurrence after resection (Table 1). A Japanese study reported a 12-gene signature identified through high-density oligonucleotide microarrays (∼6,000 genes) that reported an accuracy in the training and validation set >90% (overall sample size, 60 patients).23 Another study analyzed gene expression by using a polymerase chain reaction–based array platform of 3,072 genes in 100 HCC patients. The authors identified a 20-gene signature that was an independent predictor for recurrence.24 Staging systems and clinical classifications are needed to predict survival outcomes and properly assign treatment options. The BCLC classification has been validated in American and European cohorts and is endorsed by AASLD and EASL. This evolving classification integrates new advancements in the knowledge of prognosis and treatment, as has recently occurred with sorafenib, a multi–kinase inhibitor that showed survival advantages in patients with advanced tumors.13 There is no molecular classification of HCC. Preliminary studies suggest that tumors can be classified according to molecular biology. Signatures coming from the tumor and from the microenvironment have shown the capacity to discriminate subgroups of cancers with different survival outcomes, but these data require external validation. The incorporation of the knowledge of molecular biology in the clinical practice might aid in the selection of patients for resection and liver transplantation.