A critical need remains to identify biomarkers that may aid in hepatocellular carcinoma (HCC) surveillance, treatment stratification, and patient management.
1,2High mortality rates for patients with HCC are attributed in part to the frequency of late-stage presentation and lack of therapeutic options. Although many targeted therapies have been evaluated, sorafenib, the multikinase inhibitor, is the only systemic therapy that has shown survival benefits for advanced cases to date.3As such, a critical need remains to identify biomarkers that may aid in HCC surveillance, treatment stratification, and patient management.
Early detection is strongly linked to better clinical outcomes. The 5-year overall survival (OS) rate for patients with late-stage HCC is less than 10%. In contrast, early detection, especially if the tumor is small (<2 cm) increases the 5-year survival rate to over 50% for many patients.4Although noninvasive imaging can diagnose HCC with a certain degree of confidence, tumors often need to be at least 1 cm in size, and scanning alone is not able to predict disease progression or treatment efficacy.5Thus, the identification and development of noninvasive biomarkers with high sensitivity and specificity may offer the greatest benefit to patients because of their potential as diagnostic, prognostic, and therapeutic markers for HCC.
Several recent studies have examined potential noninvasive biomarkers in HCC and the clinical utility of liquid biopsy.
Biomarkers as Prognostic Indicators
Data from a new study show that plasma levels of VEGF and soluble VEGF receptor 1 (sVEGFR1) were associated with poor outcomes in patients with HCC. The results were first reported on April 27, 2016, in theJournal of Hepatologyby Andrew Zhu, MD, PhD, director of hepatobiliary cancer research at Massachusetts General Hospital Cancer Center in Boston, and colleagues.6
Noninvasive approaches have been under intense investigation in recent years. Liquid biopsy, which can detect blood-carrying biomarkers, has shown promise in many tumor types. This approach is attractive for multiple reasons, including the ability to rapidly analyze biomarkers and to evaluate their expression or production at multiple time points, with the ultimate goals of aiding prognosis and guiding treatment strategies.
In the current study, investigators utilized data from the phase III EVOLVE-1 trial, which examined the effects of everolimus in 546 adults with HCC whose disease had progressed during or after sorafenib or who were intolerant of sorafenib. Although efficacy was not demonstrated with everolimus, this exploratory analysis aimed to identify biomarkers associated with prognosis, etiology, or ethnicity in the EVOLVE-1 cohort. Using multiple approaches, investigators sought to identify plasma biomarkers, as well as receptor tyrosine kinase activation and genetic alterations.6
Prior to the first treatment of everolimus, plasma samples were collected and biomarker concentrations analyzed using commercially available ELISA kits. A panel of key angiogenesis regulators were measured, which included basic fibroblast growth factor (bFGF), placenta growth factor (PlGF), vascular endothelial growth factor (VEGF), VEGF-D, c-Kit, collagen type IV (Col-IV), sVEGFR1, and soluble VEGFR2 (sVEGFR2).6
Two regulators of angiogenesis were associated with a poor HCC prognosis: VEGF and sVEGFR1. Compared with patients who had low VEGF levels, patients with high levels (above the cohort median) of VEGF had significantly shorter median OS (6.24 mo vs 9.53 mo; HR, 1.59; 95% CI, 1.31-1.93; P = 3.1 x 10-6). Likewise, high levels of sVEGFR1 were also associated with shorter median OS (5.82 mo vs 11.17 mo; HR, 2.13; 95% CI, 1.75-2.59; P = 6.9 x 10-14). Using a multivariate analysis, these markers retained their prognostic significance independent of other factors, including region (Asia vs non-Asia), presence of macroscopic vascular invasion, and alpha-fetoprotein (AFP) level.6
The authors noted that this study, as well as others, support the use of serum concentrations of VEGF as a prognostic biomarker.7“Our study demonstrates that circulating angiogenesis biomarkers can predict the survival outcome in patients with advanced hepatocellular carcinoma independent of the clinical variables,” the authors concluded. Furthermore, they added that future clinical trials in HCC “…could consider sub-group analysis of HCC patients based their baseline levels of soluble angiogenic biomarkers, due to their prognostic significance.”
Interestingly, other biomarkers showed variation with disease etiology and patient ethnicity. Most notably, c-Met protein was significantly lower in Asian patients and those with hepatitis B virus (HBV)-associated disease. Genetic analysis indicated that PTEN and TSC2 mutations were highest in Asians. The authors noted that these variations in molecular pathway activation among different cohorts should be weighed carefully in future trial designs of targeted therapy.6
Novel Biomarkers on the Horizon with MicroRNAs
MicroRNAs (miRNAs) are another class of potential biomarkers that are actively being pursued for their clinical utility in HCC. Data from a recent study identified miR-30e and miR-223 as novel noninvasive biomarkers in patients with HCC. Sourav Bhattacharya, PhD, from Saint Louis University in Missouri, and colleagues,8published the study in the February 2016 issue ofThe American Journal of Pathology.
MicroRNAs are a class of small, noncoding RNAs that regulate gene expression at the posttranscriptional level, either through mRNA degradation or translational inhibition. Modulated expression of miRNAs has been associated with the development and progression of HCC as well as other cancers.
Serum was analyzed from 70 patients who were healthy (n=14), had chronic liver disease (n=17), or who had HCC with HBV infection (n=14), hepatitis C virus (HCV) infection (n=14), or nonviral‒associated disease (n=11). Differential analysis of miRNAs was determined using quantitative realtime PCR. Previous array data served as the basis for the selection of target miRNAs.8
In comparison with healthy controls, expression levels of miR-30e and miR-223 were significantly lower in patients with HCC. In stratifying patients with HCC from healthy controls, miR-30e had an area under the curve (AUC) of 0.930 ± 0.067 (95% CI, 0.8640.997; 91.67% sensitivity and 71.43% specificity) and miR-223 had an AUC of 1.00 (100% sensitivity and 100% specificity). Differential expression was present independent of disease etiology. Similarly, expression of miR-30e and miR-223 were significantly lower in patients with HCC than in patients with chronic liver disease.8
Additional analysis of HCC liver biopsy samples also demonstrated downregulation of miR-30e and miR-223, consistent with results from patient sera. Although the role of miR-30e and miR-223 is not fully understood,in silicoanalysis suggested potential roles in autophagy and microtubule regulation, respectively.
The potential advantages of miRNAs as noninvasive markers include their stable expression in serum and urine, and their resistance to RNase activity, extreme pH, and temperature.4Other miRNAs that have been investigated in HCC include miR-122, miR-101, miR-18b, and miR-221.9
“HCC is a silent disease initially, and once diagnosed at a late stage, treatment options are limited,” said Ratna Ray, PhD, senior author of the study. “Identifying a potential biomarker will give more time for treatment options.”
Ongoing Studies with an Established Biomarker, AFP
Ongoing clinical trials are using established biomarkers, such as AFP, to stratify patient response to therapy. One example is the ongoing REACH-2 phase III study that will examine the safety and efficacy of ramucirumab versus placebo and best supportive care as a second-line treatment in patients with advanced liver cancer and elevated baseline AFP (≥400 ng/mL) following first-line therapy with sorafenib (ClinicalTrials.gov Identifier: NCT02435433).10
This proposed study follows the completion of the phase III REACH trial, the results of which were published in July 2015 in The Lancet Oncology by Andrew Zhu and colleagues.11
The randomized, double-blind REACH-2 trial examined over 500 patients who were refractory or not amenable to locoregional therapy and had previously received sorafenib. Although the primary endpoint, OS, was not significantly different with ramucirumab compared with placebo, improvements were noted in progression-free survival, time to progression, and objective response.10
Importantly, additional analysis that stratified patients by AFP levels indicated a significant difference in OS between ramucirumab and placebo in patients with AFP levels greater than 400 ng/mL (7.8 mo vs 4.2 mo; HR, 0.67; 95% CI, 0.51- 0.90; P =.006). Patients who had a baseline AFP 1.5 times the upper limit of normal or greater also had a survival benefit (median OS, 8.6 mo with ramucirumab vs 5.7 mo with placebo: HR, 0.749; 95% CI, 0.603-0.930; P =.0088), suggesting that AFP may be an accurate predictor of response to this anti-VEGFR2 antibody.
Alpha-fetoprotein is a glycoprotein that is expressed in the yolk sac and fetal liver during gestational development, but decreases dramatically into adulthood. Alpha-fetoprotein is a tumor marker for a number of tumor types including HCC, gastric carcinoma, lung cancer, and testicular carcinoma. For diagnostic and prognostic purposes, an AFP level of 400 ng/ mL is the frequently used threshold, with a reported sensitivity level above 90%. For HCC staging, AFP is used as a part of the Cancer of the Liver Italian Program (CLIP) score.9
“Advanced liver cancer carries a poor prognosis with limited treatment options. Several phase III studies to date have not been able to demonstrate improved survival in the second-line setting following sorafenib failure,” said Dr. Zhu, principal investigator of the REACH trial. “Further analyses from the REACH study have identified AFP as a potential marker for selecting patients with advanced hepatocellular carcinoma who may benefit from ramucirumab treatment.”12
Applications of CTCs as a Part of Liquid Biopsy
A new study has shown that circulating tumor cells (CTCs) were detected in 65% of patients with HCC but were absent from healthy controls, using a novel imaging methodology for flow cytometry. The presence of CTCs also was found to be an independent predictor of poorer survival. Laura Ogle, from the Northern Institute for Cancer Research in Newcastle-upon- Tyne, United Kingdom, and colleagues reported the results in theJournal of Hepatology.13
Circulating tumor cells, or those that have detached from a primary site and have traveled into the vascular or lymphatic circulation, were first described in the 1800s. Although the importance of CTCs has been recognized for many years, limitations in detection and cancer-specific biomarkers have left them underutilized. However, recent advances in flow cytometry methods and a greater understanding of tumor-specific markers have advanced the clinical applications of CTCs.
The overall advancement of targeted therapies in HCC has been restricted by the inability to identify or effectively target key drivers of tumorigenesis. This limitation partially can be attributed to the avoidance of tissue biopsy as the standard diagnostic practice for those with HCC because of associated risks of hemorrhage and tumor seeding. Detection of CTCs from liquid biopsy circumvents this issue, potentially assisting in prognosis and treatment stratification.
In the new study, CTCs were detected in blood samples of 69 patients with HCC using ImageStream flow cytometry. Samples from healthy volunteers and patients with cirrhosis without cancer served as controls. For analysis, samples were depleted of red and white blood cells, and stained with a series of immunofluorescent antibodies against epithelial markers (pan- CK and EpCAM), an HCC-specific biomarker (AFP), candidate biomarkers (GPC3 and DNA-PK), and CD45 for the exclusion of white blood cells. Circulating tumor cells were identified using brightfield morphology, size, antigen expression, nuclear signal, and the absence of CD45 expression. Biomarker staining profiles also were analyzed in established HCC cell lines.13
CTCs were detected in 45/69 (65%) of patients with HCC and 0/31 control samples. Cell counts were variable and ranged from 1 to 1642 CTCs per 4 mL of blood. Biomarker analysis demonstrated both inter- and intrapatient heterogeneity of antigen expression levels: CK (29%), DNA-PK (24%), AFP (20%), EpCAM (18%), and GPC3 (12.5%). In 37% of cases, CTCs were positive for 1 or more biomarkers, and in 28% of cases, CTCs were negative for all biomarkers. A significant correlation was observed between CTC number and size of the largest tumor (0.291; P =.015).13
Overall survival was significantly associated with number of CTCs per 4 mL of blood (P =.043), tumor size (P <.0001), and the presence of either portal vein thrombosis (P =.002) or extrahepatic disease (P <.0001). The median survival of patients was 34 months for those with 0 CTCs per 4-mL blood sample compared with just 7.5 months for those with more than 1 CTC (P <.0001). Time to progression was significantly different between those with or without more than 1 CTC per 4 mL of blood (P = .006).13
The authors concluded that: “This paper strongly supports CTC enumeration as a credible liquid biopsy tool. In combination with the development of methods to detect and quantify potential stratification biomarkers, such as c-MET, or pharmacodynamics markers for monitoring treatments, these tools are likely to have a major impact in the future management of our patients with HCC.”
“Taking a biopsy in HCC is associated with risk of tumor seeding. The general potential advantage of using CTCs as a liquid biopsy is that taking a blood sample is a lot less invasive than a biopsy, and it allows multiple sampling at different time points,” said David Jamieson, PhD, coauthor of the study. “We were and are still astonished at the degree of heterogeneity we see. We are in the process of looking for additional biomarkers, but we also want to take a non-biased approach to finding new potential markers, using transcriptome analysis on recovered CTCs,” he added.
Taken together, these studies indicate the significance of noninvasive approaches in the clinical management of HCC. The multipronged approach using liquid biopsy biomarkers in addition to current noninvasive imaging techniques to detect HCC at an early stage will offer the best chance of survival for patients. The ability to use biomarkers not only as prognostic indicators but also for treatment stratification will be a large step forward in optimizing treatment for patients with HCC.