Abstract
Hepatocellular carcinoma (HCC) is the fifth most common cancer worldwide, and it is responsible for up to one million deaths annually. Although multiple risk factors for HCC have been identified, and despite preventive measures, the incidence of HCC continues to rise to epidemiologic proportions in the United States. In general, tumor resection and orthotopic liver transplantation are the treatment with the best outcome; however, HCC is generally diagnosed late in its course when patients are not eligible for curative treatment options. HCC is a relatively Chemo-refractory tumor secondary to heterogeneity of the tumor and the high rate of multidrug resistant gene expression. There are no standard treatments for HCC, multiple palliative treatment modalities have been used for patients with unresectable disease. None of these modalities have shown any superiority; and the retrospective nature of these available data has confounded any reasonable conclusions. Different institutions use different treatment schema dependent on the center expertise. Sorafenib, a tyrosine kinase inhibitor, has recently demonstrated a survival advantage in metastatic HCC, and if approved by the FDA, might become the standard of care. In this article we will review the rationale behind the currently available treatment options for HCC.
Introduction
Hepatocellular carcinoma (HCC) is an aggressive tumor that often occurs in the setting of chronic liver disease and cirrhosis. The incidence of HCC has increased in the last 10 years in many countries, and is responsible for up to one million deaths annually worldwide. The incidence of HCC varies widely according to the geographic location, with the highest incidence in sub-Saharan Africa, China, Hong Kong and Taiwan. It also differs among ethnic groups, and between regions within the same country. (Munoz et al. 1989)
The incidence of HCC in the United States is increasing with an expected incidence of nearly 19,000 new cases in the year 2007. The rising incidence of Hepatitis B and C infection is believed to be largely contributing to this increase (Davila et al. 2004). Multiple risk factors for the development of HCC have been identified, including hepatitis B, hepatitis C, environmental toxins, hemochromatosis, and cirrhosis. A smaller number of cases have been reported in non-cirrhotic patients (Bralet et al. 2000).
Staging systems for HCC have been in a constant mode of evolution. This largely reflects our continued improvements in the understanding of the pathogenesis, prognosis and refinements in the therapeutic options. Multiple staging systems for HCC are recognized; however none of them are universally accepted. Staging systems can be divided into two groups: clinical staging, and pathological staging. Clinical staging incorporates clinical and radiological parameters into the scoring scheme, and is more appropriate for patients who are not surgical candidates (Okuda, Barcelona Clinic Liver Cancer, and CLIP). The CLIP staging system is considered the system of choice because it includes easily collected data, and has been externally and prospectively validated. Pathological staging incorporates pathological findings including presence of vascular invasion within the tumor, and fibrosis in the underlying liver and is more useful in patients who are surgical candidates. (AJCC version of the TNM system/UICC).
HCC is typically diagnosed late in its course with an overall median survival of 6–20 months after diagnosis (The Cancer of the Liver Italian Program (Clip) Investigators 1998). Treatment of advanced HCC had remained stagnant following the initial pioneering work which led to the understanding of the link between cirrhosis and HCC. In general, surgical resection, or orthotopic liver transplantation (OLT) has the best long-term outcome; however, only a small percentage of patients will be eligible for these procedures at diagnosis because of the underlying liver dysfunction or tumor extent. There is a world wide shortage of donor organs, and the waiting time for liver ranges from 12–24 months in the United States, with up to 40% of patients dropping out of the transplant list because of disease progression. This has led to the development of procedures such as radiofrequency ablation which have produced successful outcomes in smaller tumors.
Other treatment options which are available for patients who do not fit the criteria for tumor resection or liver transplant or for patients on the transplant waiting list, include radiofrequency ablation (RFA), percutaneous ethanol injection (PEI), transarterial chemoembolization (TACE), systemic and targeted therapy. However no large randomized controlled trials have directly compared OLT to other forms of therapy for early HCC, or compared the different treatment modalities. Most of the data are obtained from retrospective series, and most institutions will use one or more of the modalities, depending on the institution expertise, for patients who are not a candidate for transplantation, resection, or on the waiting list. Table 1 summarizes the treatment options.
Treatment options for Hepatocellular carcinoma.
Multiple treatment algorithms have been suggested in managing HCC patients, none of them are universally used. These schemes incorporate different staging systems, and include different treatment strategies. However these strategies cannot be applied to all patients, and these algorithms cannot be generalized because of the ever evolving new treatment options, the lack of large randomized clinical trials, and the variation in available expertise and resources in a specific institution (Figure 1).
Suggested algorithm to be used in selecting the appropriate treatment for HCC.
In this article we will review the most commonly used treatment approaches including surgical resection and liver transplantation, radiofrequency ablation (RFA), percutaneous Ethanol Injection (PEI), transarterial chemoembolization (TACE), systemic and targeted therapy.
Surgical resection
Surgical resection and liver transplantation are the only curative modalities. Surveillance for high-risk patients has made it possible to diagnose HCC at an earlier stage, however only 25%-30% of patients are surgical candidates depending on their underlying liver disease and hepatic function (Shimada et al. 1996). Surgical resection is the treatment of choice for non-cirrhotic patients. The preoperative evaluation should focus on confirmation of the diagnosis, determination of the local extent of the tumor, and assessment of the underlying liver function. The optimal candidates for resection are those with solitary non-metastatic lesions, with no radiographic evidence of invasion of the hepatic or portal veins, and no evidence of portal hypertension with preserved hepatic functions (Child Pugh Class A).
There are no guidelines regarding tumor size for resection. Although patients with smaller tumors (<5 cm) tend to have a better outcome, larger tumors (even >10 cm) may also benefit from surgical resection, in the absence of vascular invasion. (Vauthey et al. 1995). In selected cases, five-year relapse-free survival, and overall survival rates after surgery is about 40 percent and 87 percent respectively. (Arii et al. 2000) Perioperative mortality can range from 1 to 24 percent, with the majority of deaths caused by uncontrolled hemorrhage at the time of surgery, postoperative liver failure and infection. With an experienced surgical team, major surgical complication such as bile leakage and bowel injury can be reduced significantly (Tanaka et al. 2006). The presence of cirrhosis best predicts post-resection liver failure and death; with the 30-day postoperative mortality being twice as high in cirrhotic as in noncirrhotic patients (Bozzetti et al. 1992). CLIP (Cancer of the Liver Italian Program) staging classification, which incorporates tumor stage as well as hepatic function, best predicts survival in HCC.
Recurrence after resection exceeds 70% at 5 years. There is no standard neoadjuvant or adjuvant therapy that can reduce recurrence rate. Repeat resection is feasible for a solitary lesion in non-cirrhotic patients, and patients with recurrence should be considered candidates for salvage transplantation. In conclusion, hepatic resection may be appropriate for small lesions in noncirrhotic patients or possibly in cirrhotic patients with well preserved liver function.
Liver transplantation
Orthotopic liver transplantation (OLT) is a viable and potentially curative option for selected patients with HCC. The rationale behind liver transplantation for patients with HCC is to eradicate the tumor, and replace the cirrhotic liver that remains at risk for the development of new lesions secondary to a field defect. No large randomized controlled trials have directly compared OLT to other forms of therapy for early HCC. However many retrospective series suggested that long-term results after transplantation are as good if not better than those of resection in selected patients, especially in patients with underlying cirrhosis (Iwatsuki et al. 1991; Colella et al. 1998; Bismuth et al. 1993).
Before establishing selection criteria for patients with HCC, earlier experience with transplantation was disappointing with a high perioperative mortality, high recurrence rate, and a 5-year survival rate around 18% (Ringe et al. 1991; Mazzaferro et al. 1996) In 1996, Mazzaferro et al. established criteria for selection of patients with HCC for orthotropic liver transplantation. A single lesion <5 cm, up to three separate lesions, none larger than 3 cm, no evidence of gross vascular invasion, and no regional nodal or distant metastases are now known as the Milan criteria and have been applied in the selection of patients with HCC for liver transplantation. (Mazzaferro et al. 1996) By applying the Milan criteria, five-year survival rate of 75 percent or higher can be achieved which is similar to expected survival rate for patients undergoing transplantation for cirrhosis.
Overall survival after liver transplant is affected by histological grade, tumor size >5 cm, and the presence of positive nodes. Tumor size >5 cm, positive lymph nodes, bilobar spread, and vascular invasion are associated with a lower recurrence-free survival (Molmenti EP et al. 2002). Despite the benefit of transplantation, the group of patients who actually receive a graft is limited by the availability of donor livers. Waiting time is a major prognostic factor, and up to 40% of patients will drop out of the transplant list because of disease progression during the long waiting time of up to 24 months. The shortage of donors, has led to the novel technique of living donor transplantation which has been performed worldwide with promising results. Long term outcome and factors affecting survival are not yet available for this procedure. MELD score “model for end stage liver disease” is a statistical model that uses serum bilirubin, creatinine, and INR, to predict survival in patient with cirrhosis. The higher the score, the worse the short term prognosis. Most HCC patients have a preserved liver function earlier in their course, and the MELD scoring system cannot predict their disease severity. This led the UNOS (United Network for Organ sharing) to develop a separate point system for HCC beyond the Milan criteria that would provide these patients access to an allograft before their HCC progresses.
The neoadjuvant treatment of patients on the transplant waiting list is controversial. No available randomized trials have compared different “bridging” modalities, or compared “bridging” to no therapy, and all the available data comes from cohort studies. Most centers perform either chemoembolization or radiofrequency ablation upon listing to reduce the tumor burden and to delay progression, based on their institutional expertise. While the impetus to investigate novel neoadjuvant treatments is justified to prevent tumor progression, the absence of randomized controlled trials leaves uncertainty about the utility of these maneuvers in improving outcome.
Percutaneous ethanol injection (PEI)
Before the advent of RFA, percutaneous ethanol injection (PEI) had been the most widely accepted minimally invasive method for treating small HCC tumors. It involves the injection of ethanol inside the liver tumor; and can be performed in multiple sessions as an outpatient, or alternatively in a single inpatient session under general anesthesia. PEI is recommended for patient with tumors less than 3 cm in diameter, whose volume is less than 30 percent of the total liver volume. It is contra-indicated in patients with extrahepatic disease, portal vein thrombosis, Child-Pugh class C cirrhosis with a prothrombin time >40 percent of normal, or a platelet count below 40,000/μl (Livraghi et al. 1995). PEI should not be considered for lesions in the dome of the liver due to the technical difficulties, and complications with this location.
PEI is a safe and well tolerated procedure; localized pain tends to be commonest morbidity associated with this procedure. Serious complications such as bleeding, liver failure, bile duct necrosis or billiary fistula, hepatic infarction, hypotension, and renal failure have been reported in less than 5 percent of patients (Fujimoto, T. 1988), (Livraghi et al. 1998).
PEI has never been compared to other modalities in randomized trials; however, in an uncontrolled series its efficacy was better than no treatment and equivalent to that of surgery (Livraghi et al. 1992).
Although PEI is a cheap, minimally invasive procedure, that requires minimum amount of equipment, it has been replaced by radiofrequency ablation in many areas. Many prospective trials support the superiority of RFA over PEI. In one of these trials 232 patients were treated with either PEI or RFA. RFA patients required fewer treatment sessions and had a 43 percent lower risk of overall recurrence and a better 4-year survival. The toxicity profile in both procedures was similar. (Shiina et al. 2005) Similar results were found in another trial where 157 patients were randomly assigned to RFA or PEI. Compared to PEI, RFA patients required fewer sessions, had a lower three-year local tumor progression rate, and higher cancer-free survival rate (Lin et al. 2004).
In summary, RFA has replaced PEI in many centers as a safe and effective treatment for HCC in cirrhotic patients with small tumors who are not a surgical candidate.
Radiofrequency ablation
Improvement in imaging and active screening has increased the detection of small HCC tumors. Percutaneous local treatment has been adopted alone or in combination with other treatment modalities in the treatment of small HCCs. Surgery or transcatheter arterial embolization might not be a viable option for all patients due to poor liver function, or inadequate angiogenesis in small HCC. There has always been a need for more effective technique to treat small nodules.
Radiofrequency ablation (RFA) involves the local application of radiofrequency thermal energy to the lesion. Ionic movement within the tissue results in heating of the tissue (up to 60 °C), causing an area of necrosis surrounding the electrode. RFA has been used for many years to treat chronic pain, excessive bleeding during surgery, and cardiac arrhythmia. More recently, the technique has been used to treat different types of cancer including hepatocellular carcinoma. RFA has been shown to be superior to Ethanol injection as outlined above.
RFA safety and efficacy when used in a neoadjuvant fashion, prior to liver transplant has been evaluated in few studies. The procedure was considered safe. Mortality rate is only 0%-0.3%. Side effects like pleural effusion and peritoneal bleeding were rare (Bruix and Morris, 2005). Most of these studies showed a high rate of complete necrosis in the explants better than those reported with ethanol injection. These studies have been heterogeneous; treating different size of tumors and used different types of RF generators (Table 2).
Summary of the studies that used Radiofrequency ablation for the local treatment of small HCC.
Tumors treated percutaneously/during laparotomy.
Transonography-guided percutaneous microwave coagulation therapy (PMCT) has also been used for the local treatment of small HCC. It was reported to be a safe, effective and minimally invasive local treatment for small tumors <3 cm. However it requires multiple treatments because of the small size of the necrotic area that is achieved with each treatment. When compared to PMCT, RFA requires smaller number of treatments, and the necrotic area was significantly larger. The local recurrence rate was significantly lower after RFA than after PMCT, while the ectopic recurrence rate showed no significant difference between the two procedures. The survival rate was significantly higher after RFA compared with PMCT. Pain, fever, bile duct injury, pleural effusion and ascites after treatment were significantly higher in PMCT patients. (Ohmoto et al. 2006)
American association for the study of liver disease (AASLD) considers local ablation safe and effective therapy for patients who are not a resection candidate, or as a bridge to transplantation (level II evidence). In summary, and despite the lack of controlled clinical trials, RFA is considered a safe, effective and cost effective local treatment option, in patients with preserved hepatic function. The necrotic effect of RFA is more predictable, and its efficacy is superior to that of alcohol injection for small unresectable HCCs. The effects of RFA on long-term, disease-free survival are not clear. Larger studies using intent-to-treat analysis are still needed.
Transarterial chemoembolization (TACE)
One of the cardinal pathological features of HCC is its ability to exhibit neoangiogenises. The majority of tumor blood supply is derived from the hepatic artery. This property is used in radiology to diagnose HCC and provide the basis to support arterial embolization as a treatment modality. Transarterial embolization (TAE) involves hepatic artery obstruction during angiography, while transarterial chemoembolization (TACE) involves the injection of chemotherapeutic agents with lipiodol into the artery. Lipiodol is an oily contrast agent used usually for lymphographic studies; it promotes retention of chemotherapy drugs inside the tumor. TACE has been traditionally used for the treatment of patient with large unresectable HCCs that are not eligible for other treatments such as resection, RFA, or liver transplant.
The most commonly used chemotherapeutic agents are adriamycin, cisplatin. (Bruix J et al. 2004) or a combination of these with mitomycin C (Solomon B et al. 1999). Several TACE protocols with different embolization methods (Gelfoam, starch, glass microspheres, polyvinyl alcohol) have been developed, with no consensus as to the most effective technique (Ramsey et al. 2002). TACE is contraindicated in the absence of portal blood flow, in patient who have advanced liver disease (Child-Pugh B or C), encephalopathic or in patients with billiary obstruction. TACE is relatively contraindicated in patient with serum bilirubin >2 mg/dL, Lactate dehydrogenase >425 U/L, Aspartate aminotransferase > 100 U/L, large tumor burden involving more than half the liver, patient with cardiac or renal disease, ascites, or significant thrombocytopenia. Chemotherapy used with TACE has minimum systemic side effects. TACE can cause fever, abdominal pain, and ileus (also called post embolization syndrome). These symptoms are self-limited and can be treated conservatively. Prophylactic antibiotics are used routinely is some centers, but severe infectious complications are rare.
Numerous efficacy endpoints have been used to demonstrate the utility of TACE. These include tumor shrinkage after TACE, impact on overall survival, or alternatively decrease the dropout rate of patients on the transplant waiting list in small trials. However large prospective randomized trials have failed to demonstrate a survival advantage (Table 3). In a systematic review of seven randomized trials comparing chemoembolization to best supportive care for unresectable hepatocellular carcinoma, there was a significant benefit of chemoembolization with cisplatin or doxorubicin but none with embolization alone. There was an improved 2-year survival compared with control (odd ration of 0.53), and treatment induced objective responses in 35% of patients (Llovet and Bruix, 2003). Another systematic review failed to show a survival advantage associated with therapeutic embolization versus supportive care alone in patients with unresectable HCC. The odds ratios for 3- and 6-month survival were 1.31 and 0.91 which was not statistically significant (Geschwind et al. 2003). Independent predictors of worse survival were ascites, alpha-fetoprotein (>400 U/L), tumor size (>50%), Child-Pugh grade (Child C), pattern of iodized oil uptake, and portal vein thrombosis. (Marcko et al. 2000). Another popular use of TACE in many transplant centers, despite the lack of randomized clinical trials, is as a “bridging therapy” before transplant, to prevent tumor progression while the patient is on the transplant list.
Summary of randomized clinical trial comparing Chemoembolization to best supportive care.
Control patients received Tamoxifen.
Groupe d'Etude et de Traitement du Carcinome Hépatocellulaire.
In conclusion TACE should be offered to symptomatic patients who are not a candidate for resection, RFA, or liver transplantation. There is no standard chemotherapy agent which has been universally accepted for embolization. Most of the larger centers have accepted adriamycin as the standard of care. Lipiodol enhances embolization of the vessels and might help retain the chemotherapy thus enhancing the efficacy of the procedure.
Systemic therapy
HCC is a relatively chemotherapy refractory tumor secondary to the high rate of multidrug resistant gene expression, and the liver dysfunction associated with the disease which limits the use of many systemic agents. Several agents have been used in the treatment of hepatocellular carcinoma and can be divided into hormonal therapy, systemic chemotherapy, and targeted therapy.
Hormonal therapy
Up to third of HCC tumors have estrogen receptors, and the theoretical benefit of estrogen receptor antagonist, Tamoxifen, was tested in multiple prospective randomized trials in patients with unresectable HCC. Theses trials failed to show any efficacy, survival benefits, or advantage to quality of life in patients with metastatic disease (Castells et al. 1995; CLIP group 1998; Chow et al. 2002). Megesterol has also been used in the same setting in small randomized trials, and failed to show any objective response (villa et al. 2001). Liver tissues from HCC patients also express a variable concentration of somatostatin receptors. Long and short acting octerotids have been used in small randomized trials for the treatment of HCC, and although it showed some survival advantage in one randomized trial (Kouroumalis et al. 1998), the results could not be duplicated in subsequent studies.
Systemic chemotherapy
Alteration of drug resistance and tumor suppressor genes in human HCC (Huang et al. 1992; Caruso et al. 1999) leads to intrinsic resistance to chemotherapy. Multiple chemotherapy agents (doxorubicin, 5FU, Gemcitabine, Cisplatin, Capecitabine, and Mitoxantrone) have been tested in phase II and III trials for the treatment of unresectable HCC, alone or in combination with other chemotherapy agents, immunotherapy (INF-alfa), or hormonal treatment. Most of these trials were underpowered, single institution studies that recruited highly selected patients (Nowak et al. 2004). There is no strong evidence supporting the use of any specific chemotherapeutic regimen in the treatment of HCC, The most active chemotherapy in HCC is adriamycin with a response rate under 20% (Olweny et al. 1975). Despite this modest response rate, this has been associated with only an extremely small survival advantage when compared to best supportive care. Cytotoxic treatment, if used, should be in selective patients in the setting of a clinical trial.
Molecular targeted therapy
The rapid development of targeted therapies and the lack of effective chemotherapy have made the evaluation of novel therapies with signal transduction modulation a natural alternative approach. Innovative modern techniques have provided further insight into the intracellular pathways that result in sensitivity and resistance of the neoplastic cells to drug treatment. This acquisition of new knowledge is occurring at a breakneck pace resulting in a change in the understanding of the biology of the disease, a stage shift in clinical presentation, and the development of highly accurate prognostic models.
Erlotinib, an oral tyrosine kinase inhibitor with specificity to epidermal growth factor receptor (EGFR/HER1) was tested in a phase II study in patients with unresectable HCC who have progressed following one prior systemic therapy. 32% of patients were progression-free at 6 months; disease control was seen in 59% of patients with overall survival of 13 months (Philip et al. 2005). Bevacizumab, an anti-vascular endothelial growth factor (VEGF) monoclonal antibody was used in combination with chemotherapy in a small phase II trial (Zhu et al. 2006). The 6 month progression free survival rate was encouraging (48% of patients). These results await confirmation in a phase III trial.
Thalidomide, an antiangiogenic agent with activity in hematological malignancy, has been used in a phase II trial in 37 patients with hepatocellular carcinoma (Chuah B et al. 2007). The median dose was 400 mg/day. There was no complete response. One patient (3%) had a radiological partial response and six (16%) had stable disease for more than 6 months. Side effects were tolerable.
RAF/mitogen-activated protein/extracellular signal-regulated kinase (ERK) kinase (MEK)/ERK cascade have been reported to play an important role in the development of HCC. Sorafenib is a multikinase inhibitor that targets Raf kinase and receptor tyrosine kinases. It demonstrated a modest activity in a phase II trial for patients with HCC where 33% had stable disease for at least 16 weeks, with manageable toxicity. (Abou-Alfa et al. 2006) This led to the phase III randomized placebo-controlled trial (SHARP trial) comparing Sorafenib to placebo (Sor/P) in the treatment of patients with advanced measurable HCC. The results presented in a preliminary presentation (abstract form) at the plenary session of the 2007 ASCO meeting showed a survival advantage for the sorafenib arm. 602 patients (mostly with Child-Pugh A) were randomized. The hazard ratio for overall survival (OS) between Sorafenib/placebo was 0.69 and was statistically significant, representing a 44% improvement in OS. Median OS was 10.7 vs 7.9 months. Median time to progression was longer (5.5 vs. 2.8 mos) and the disease control rate was higher (43% vs 32%) with Sorafenib. Incidence of serious adverse events was similar for both arms (52% vs 54%). The most frequent grade 3/4 events were diarrhea, hand-foot skin reaction, fatigue, and bleeding. (Llovet et al. 2007), so in selected patients with Child-Pugh class A, who are not surgical, or local therapy candidates, Sorafenib might be a viable first line treatment.
Conclusions
Few randomized clinical trials address the different management options of HCC, which makes the establishment of an evidence based treatment difficult. Most of the recommended approaches are based on heterogeneous observational studies. Tumor resection should be considered for single tumors preferably less than 5 cm with good liver function. Percutaneous ablation should be considered in patients with small tumors who are not a surgical candidate. Transplantation should be considered in patients with impaired liver function, who are not eligible for resection. Because of the small number of donor livers available for transplantation, living donor should be considered when available. The significant improvement in overall survival recently found with the use of Sorafenib in the treatment of patients with metastatic disease might provide a new line of treatment for patients with advanced disease.
Many of the trials which are reviewed in this article suffer from similar deficiencies secondary to small numbers, heterogeneous patient populations, and nonstandardized endpoints which limit the conclusions that can be drawn. Analysis of outcome data from contemporary clinical trials is further confounded by the stage migration due to earlier detection of HCC in patients with viral hepatitis. To achieve further progress in this field, randomized trials are desperately needed which will require a large number of patients and the support of multiple institutions across the continents.
Improvements in survival with the addition of currently available approaches have been modest at best. The dawn of an exciting era of drug development is upon us, representing a treatment paradigm shift. An expanding portfolio of newer targeted agents, have lent themselves to testing in this setting. Enhancements in understanding of the molecular pathways over the last few decades have resulted in targeted agents, which can be combined with classic cytotoxic agents. Recognizing the multifactorial nature of drug resistance and the inherent survival mechanisms has made us realize that collateral and downstream pathways can circumvent targeting single mechanisms of drug resistance. Bypassing these adaptive cellular changes would be better served with a quiver full of arrows.
While none of the newer treatment modalities have yet been shown to be more effective than standard treatments, the potential armamentarium is steadily growing, triggering cautious optimism. Combining local ablative approaches with the newer targeted agents, raise the appealing plausibility of parallel, or possibly complementary, therapeutic effects. Efficient trial design, appropriate selection of correlative markers, greater cooperation between hepatologists, surgeons and medical oncologists, with close toxicity monitoring will propel this field further and improve our management of this disease.