Worldwide, hepatocellular carcinoma is the fifth most common malignancy and the third most common cause of cancer mortality. Most patients with hepatocellular carcinoma suffer from cirrhosis, primarily caused by alcoholism or chronic infection with hepatitis B virus (HBV) or hepatitis C virus (HCV); decades may pass between infection with viral hepatitis and development of this cancer. The approximately equal annual incidence vis-à-vis mortality of 1 million persons reported around the world stands as evidence of its lethality. Taken together, in the US approximately 46,570 cases of liver, gallbladder, and bile duct cancers were diagnosed in 2015, with an estimated 28,250 deaths.
Hepatocellular carcinoma is the most common tumor in males worldwide, with a male to female ratio of 5:1 in Asia and 2:1 in the United States.
Eighty percent of new hepatocellular cancer cases occur in developing countries, but the incidence of this cancer is also rising in developed countries. Modeling of the spread of HCV infection suggests that this number may continue to increase dramatically. In the United States, the incidence of hepatocellular carcinoma is 5 per 100,000, whereas in East Asia and sub-Saharan Africa, this neoplasm occurs at an incidence of 150 per 100,000 population and accounts for almost 50% of all diagnosed tumors.
The incidence of hepatocellular cancer increases with age. The mean age at diagnosis is 53 years in Asia and 67 years in the United States.
From 2003 to 2005, the incidence of hepatocellular tumors in the United States was higher among Asian immigrants (12 per 100,000) and black individuals (7 per 100,000) than among white individuals (4 per 100,000). However, a study analyzing the Surveillance, Epidemiology, and End Results (SEER) data has shown that the incidence of hepatocellular carcinoma is rising in both white and black populations in the United States.
In patients who meet stringent criteria and undergo transplant, the 5-year survival is approximately 65% to 70%; the 5-year recurrence rate is less than 20%. The 5-year survival rate following liver resection (governed by a different set of operative criteria) is about equal to that following transplant. This result is in spite of a majority of patients experiencing metastases or a recurrence in the remaining part of the liver. Fewer than 20% of resected patients who have tumor recurrence are able to undergo transplant. Seventy-five percent of patients have unresectable disease that is diagnosed at an advanced stage. Median survival in these patients can range from months to years, depending on the stage of disease, the biologic aggressiveness of the tumor, vascular invasion, and “background” liver function, among other important factors.
Etiology and Risk Factors
The annual incidence of hepatocellular carcinoma in HBV carriers is 0.5%, and in patients with known cirrhosis, it is 2.5%. In one study, the relative risk of hepatocellular carcinoma in HBV carriers was 100 in Asian patients, who likely acquired the virus at birth. The epidemiology of HBV infection is different in the Western world, where it is acquired later in life, and risk in white individuals appears to be related to inflammatory activity and cirrhosis.
In areas in which hepatitis B is endemic, approximately 90% of all patients with hepatocellular carcinoma are positive for hepatitis B surface antigen (HBsAg). The presence of the hepatitis B “e” antigen has been found to increase risk nine-fold. An important reason to provide antiretroviral agents to patients with chronic HBV infection is the relationship between the risk of hepatocellular carcinoma development and absolute serum levels of HBV DNA. The most compelling epidemiologic evidence of a causal relationship between HBV infection and hepatocellular carcinoma is the observation of a significant decline in the incidence of childhood hepatocellular carcinoma after the introduction of a national immunization program in Taiwan. The hepatitis B “X” gene, which can interact with p53, has been a focus of study on the pathogenesis of hepatocellular carcinoma.
In a study following patients with chronic liver disease for development of hepatocellular carcinoma, 18,000 Chinese patients with chronic HBV infection were screened with α-fetoprotein determinations and liver ultrasonography every 6 months or received no screening for 5 years. Zhang et al found that the hepatocellular carcinoma mortality rate was 37% lower in the screened group than in the controls, and this outcome can be attributed to early-stage tumor detection and resection.
The risk of hepatocellular carcinoma in patients with chronic HCV infection and established cirrhosis is 2% to 8% per year. The molecular mechanisms of HCV infection and carcinogenesis are poorly understood, yet the disease is believed to occur in the context of chronic inflammation, which leads to fibrosis and cirrhosis. Unlike patients with HBV infection, patients with hepatocellular carcinoma infected with HCV usually have cirrhotic livers at diagnosis; this finding suggests an extended period of infection (or hepatic damage) before malignancy develops.
Patients with alcoholic cirrhosis are at risk for hepatocellular carcinoma, and the addition of HCV infection increases that risk dramatically. About 60 to 80 g of alcohol (equivalent to four to six beers) must be consumed daily over 5 to 10 years for a male to develop cirrhosis (odds ratio = 4). The interaction between alcohol consumption and HCV infection approximately doubles the odds ratio.
Non-alcoholic fatty liver disease and non-alcoholic steatohepatitis
Several epidemiologic studies have now established fatty liver disease and its complications as important risk factors for hepatocellular carcinoma. In a recent meta-analysis, cohorts with nonalcoholic steatohepatitis (NASH) and cirrhosis (but not those without cirrhosis) had a higher risk of hepatocellular carcinoma (cumulative incidence ranging from 2.4% over 7 y [in a study that combined clinic-based cohorts from four countries] to 12.8% over 3 y [in a single clinic-based cohort from the US]). However, the risk for HCC was substantially lower in these cohorts than for cohorts with hepatitis C–related cirrhosis.
Other possible etiologies
These include aflatoxin, hemochromatosis, hepatic venous obstruction, Thorotrast (a contrast agent no longer used for radiologic procedures), androgens, estrogens, and α1-antitrypsin deficiency.
A large 5-year randomized study performed in China has confirmed that treatment of chronic HBV infection with the antiviral agent lamivudine (Epivir) not only decreases the risk of progression to cirrhosis of the liver but also decreases the rate of progression to hepatocellular cancer. Treated patients had a 3.9% risk of hepatocellular cancer over 5 years, compared with 7.4% for those in the placebo group (hazard ratio [HR] = 0.49; P = .047). It remains controversial whether antiviral therapy with ribavirin (Copegus, Rebetol) and interferon or newer treatments such as Harvoni, a fixed-dose combination of ledipasvir, a hepatitis C virus (HCV) NS5A (nonstructural protein 5A) inhibitor, and sofosbuvir (Sovaldi), an HCV nucleotide analog NS5B (nonstructural protein 5B) polymerase inhibitor, decreases the risk of hepatocellular carcinoma in cirrhotic patients with HCV infection who achieve a sustained virologic response. If there is any benefit to therapy in such cases, it appears to be small.
Signs and Symptoms
Patients with hepatocellular cancer usually present with abdominal pain and other vague symptoms, including malaise, fever, chills, anorexia, weight loss, and jaundice.
An abdominal mass is noted on physical examination in one-third of patients with hepatocellular cancer. Less common findings include splenomegaly, ascites, abdominal tenderness, muscle wasting, and spider nevi. Up to 10% of patients may present with an acute abdomen due to a ruptured tumor.
Screening and Diagnosis
Presently, no organization recommends routine screening of average-risk, asymptomatic adults for liver, gallbladder, and biliary tract cancers.
This serum marker is produced by 70% of hepatocellular carcinomas. The normal range for this serum marker is 0 to 20 ng/mL, and a level greater than 200 ng/mL is essentially diagnostic for hepatocellular cancer in the absence of chronic, active HBV infection, where higher values can be seen in the absence of cancer. In the setting of HCV infection, the cutoff for diagnosis of hepatocellular carcinoma is controversial. Values greater than 200 ng/mL appear to be highly predictive of hepatocellular carcinoma (although with a high false-negative rate compared with lower cutoffs). False-positive results may be due to acute or chronic hepatitis, germ-cell tumors, or pregnancy.
Hepatitis B and C
Given the association between hepatitis B and C and hepatocellular cancer at the time of diagnosis of a hepatic mass, blood should be sent for hepatitis B and C antigen and antibody determinations. Hepatitis B should be treated with antiviral medications prior to chemoembolization, to avoid a disease flare.
The initial diagnostic test in the symptomatic patient may be ultrasonography, because it is noninvasive and can detect lesions as small as 1 cm. Negative results on ultrasonographic examination should not be accepted as diagnostic. Ultrasonographic findings should be followed up with more specific imaging.
Triple-phase, high-resolution computed tomography (CT) and contrast-enhanced magnetic resonance imaging (MRI) are the primary imaging modalities used to diagnose and stage hepatocellular carcinoma. Typically, hepatocellular carcinoma is a hypervascular tumor in the arterial phase with washout in the portal venous phase. Reports have documented a high number of false-positive results with CT angioportography and CT hepatic angiography. Lesions that meet the arterial “blush” criteria with washout in the venous imaging phase in a patient at high risk for hepatocellular carcinoma may not require biopsy for confirmation of cancer. This is particularly true for patients with elevated α-fetoprotein levels. CT predicts resectability in only 40% to 50% of cases and does not accurately determine the functional extent of cirrhosis. Major difficulties arise when the liver parenchyma is not homogeneous and the lesions are smaller than 1 cm. Use of positron emission tomography scanning is not defined in this disease, and the uptake of 18F-fluorodeoxyglucose by hepatocellular tumors is relatively low.
This procedure is useful for the evaluation of small tumors, the extent of cirrhosis, peritoneal seeding, guidance during biopsy and ablation of tumors, and the assessment of the volume of noninvolved liver. Although CT and MRI provide excellent preoperative staging, they may be used before open laparotomy. Laparoscopic or intraoperative ultrasonography should be used to confirm preoperative imaging tests. The laparoscopic ultrasound results may change surgical management in up to one-third of select patients.
Individuals at high risk should be screened for hepatocellular carcinoma using ultrasonography and serum α-fetoprotein levels. When ultrasonograms are difficult to interpret, contrast-enhanced CT or MRI may be considered. Screening increases the proportion of cancers that are resectable. However, a study comparing 6-month and 12-month survival intervals in a cohort of HCV-infected patients with hemophilia showed no substantial benefit of more frequent screening.
Three morphologic patterns of hepatocellular carcinoma have been described: nodular, diffuse, and massive. Diffuse and massive types account for more than 90% of cases. The nodular type usually has multiple lesions in both lobes.
Several histologic arrangements have been identified: trabecular; compact; pseudoglandular or acinar; clear cell; and a fibrolamellar variant, which is associated with a relatively favorable prognosis and a younger age at diagnosis. The fibrolamellar variant is more commonly resectable and is not usually associated with infection and cirrhosis.
Staging and Prognosis
The staging system for hepatocellular cancer is based on the number and size of lesions and the presence or absence of vascular tumor invasion (Tables 1 and 2). The Okuda staging system accounts for the degree of liver dysfunction and may better predict prognosis than the TNM staging system. However, the Okuda staging system does not adequately predict resectability and primarily predicts end-stage disease. The Child-Pugh system and MELD (Model for End-Stage Liver Disease) scores measure liver function and are not cancer staging systems. Because of the limited value of standard staging, the most important factors that determine survival are technical resectability of lesions and the degree of dysfunction of the normal liver. Groups in Spain, Italy, and China have created prognostic indices that may prove useful for making treatment decisions. The Barcelona Clinic Liver Cancer staging system was designed to be a diagnostic and treatment strategy to compare tumor stage, liver function status, and performance status in its schema. Surgical resection is considered to be the best treatment for patients with solitary tumors and normal bilirubin levels who do not have portal hypertension. According to the Barcelona schema, patients may be considered for liver transplant if the above criteria are not met or for ablation if the disease is at an early stage (solitary tumors < 5 cm or up to three nodules with no single nodule > 3 cm). Such patients will have a 5-year survival of 50% to 75%. Chemoembolization is appropriate for patients with intermediate-stage disease who are asymptomatic and have preserved liver function, with a bilirubin level of less than 3 mg/dL; their 3-year survival will be 50% or better. For patients with advanced disease, there are fewer established therapeutic options; enrollment in a research study may be the best of these alternatives.
Of the roughly 20% of patients who can undergo resection, factors associated with improved survival include curative resection, small tumor size, well-differentiated tumors, and normal performance status. Cirrhosis, nodal metastases, and an elevated prothrombin time are indicative of a poor prognosis, as are male sex, age over 50 years, poor performance status, duration of symptoms < 3 months, tumor rupture, aneuploidy, high DNA synthesis rate, hypocalcemia, vascular invasion, and a high serum α-fetoprotein level.