Hepatic Tumors and Cysts

Hepatic mass lesions include tumors, tumor-like lesions, abscesses, cysts, hematomas, and confluent granulomas. Hepatic tumors may originate in the liver—from hepatocytes, bile duct epithelium, or mesenchymal tissue—or spread to the liver from primary tumors in remote or adjacent organs. In adults in most parts of the world, hepatic metastases are more common than primary malignant tumors of the liver, whereas in children, primary malignant tumors outnumber both metastases and benign tumors of the liver. Except for cavernous hemangiomas, benign hepatic tumors are rare in all geographic regions and in all age groups.

Malignant Tumors

HCC

Epidemiology

HCC is the most common primary malignant tumor of the liver. It is the fifth most common cancer in men and the eighth most common in women, and it ranks fourth in annual cancer mortality rates. ^, Information on incidence is derived from an increasing but still limited number of cancer registries, and it is possible to classify countries into broad risk categories only. Moreover, in low-income (developing) countries, especially in sub-Saharan Africa, HCC is underdiagnosed and underreported, in some cases by as much as 50%. Despite these sources of inaccuracy, HCC clearly has an unusual geographic distribution ( Fig. 96.1 ). The incidence of HCC has increased considerably in Japan since the 1980s, and lesser increases have been recorded in developed Western countries, including North America and Western Europe. Interestingly, a study from Japan has shown that the rate of HCC began to decline in 2000, presumably because of the aging of the cohort of persons infected with HCV. A similar downward trend has been noted in some European countries, including France and Italy. By contrast, in the USA, HCC is the cancer that has been increasing in incidence most rapidly since 2000, at a time when the incidence of other major cancers such as cancers of the lung, breast, prostate, and colon is decreasing. Considerable racial and ethnic variation exists in the incidence of HCC in the USA. The incidence among Asians is highest, almost double that of white Hispanics and more than 4 times higher than that of whites.

Fig. 96.1, Incidence of HCC in different parts of the world. High, age-adjusted rate of more than 15 cases/100,000 population/yr; intermediate, age-adjusted rate of 5-15 cases/100,000/yr; low, age-adjusted rate of fewer than 5 cases/100,000/yr.

Migrants from countries with a low incidence to areas with a high incidence of HCC usually retain the low risk of their country of origin, even after several generations in the new environment. The consequences for migrants from countries with a high incidence to those with a low incidence differ, depending on the major risk factors for the tumor in their country of origin and whether chronic HBV infection, if this is the major risk factor, is acquired predominantly by the perinatal or horizontal route (see later and Chapter 79 ). ^, ^,

Men are generally more susceptible than women to HCC. Male predominance is, however, more obvious in populations at high risk for the tumor (mean male-to-female ratio, 3.7:1) than in those at low or intermediate risk (2.4:1). ^, In industrialized countries, the number of men and number of women with HCC in the absence of cirrhosis is almost equal.

The incidence of HCC increases progressively with advancing age in all populations, although it tends to level off in the oldest age groups. ^, In Chinese and particularly in black African populations, however, the mean age of patients with the tumor is appreciably younger than in other populations. This finding is in sharp contrast to the age distribution in Japan, where the incidence of HCC is highest in the cohort of men 70 to 79 years of age. HCC is rare in children. ^,

Etiology and Pathogenesis

In contrast to many other malignancies, for which risk factors can only sometimes be identified, the immediate cause of HCC can usually be identified and is most commonly chronic viral hepatitis or cirrhosis. HCC is multifactorial in cause and complex in pathogenesis. Four major causative factors have been identified ( Box 96.1 ). The differing blend of risk factors in various parts of the world may explain, in part, the diverse biological characteristics of HCC in various populations.

BOX 96.1

Risk Factors for HCC

Major Risk Factors

Chronic HBV infection
Chronic HCV infection
Cirrhosis
NAFLD

Other Liver Conditions

α ₁ -Antitrypsin deficiency
Hemochromatosis
Membranous obstruction of the inferior vena cava
Type 1 and type 2 glycogen storage disease
Type 1 hereditary tyrosinemia
Wilson disease

Inherited Conditions Not Associated With Liver Disease

Ataxia-telangiectasia
Hypercitrullinemia

Other Factors

Cigarette smoking
Diabetes mellitus
Dietary exposure to aflatoxin B ₁
Oral contraceptive steroid use

HBV

Some 387 million carriers of HBV exist in the world today, and HCC will develop in as many as 25% of them (see Chapter 79 ). HBV infection accounts for up to 80% of HCCs, which occur with high frequency in East Asian and African populations. ^, Persistent HBV infection antedates the development of HCC by several to many years, an interval commensurate with a cause-and-effect relationship between the virus and the tumor. Indeed, in at-risk populations, the HBV carrier state is largely established in early childhood by perinatal or horizontal infection. ^, Approximately 90% of children infected at this stage of life become chronic carriers of the virus, and these early-onset carriers face a lifetime relative risk for developing HCC of more than 100 compared with uninfected controls.

An effective vaccine against HBV has been available since the early 1980s, and in countries where this vaccine has been included in the expanded program of immunization for a sufficient length of time, the HBV carrier rate among children has decreased by 10-fold or more. Studies in Taiwan, where universal immunization was started in 1984 and where the rate of HBV carriage among children has decreased by more than 10-fold, have shown a 70% reduction in the mortality rate from HCC in children in the vaccinated age groups. This finding gives promise for the ultimate eradication of HBV-induced HCC and provides further evidence of the causal role of the virus in the development of this tumor.

The precise mechanism by which HBV results in HCC is not known; however, the virus appears to be both directly and indirectly carcinogenic. HBV DNA is integrated into cellular DNA in approximately 90% of HBV-related HCCs. The sites of chromosomal insertion appear to be random, and whether viral integration is essential for hepatocarcinogenesis is still uncertain. Possible direct carcinogenic effects include cis -activation of cellular genes as a result of viral integration, changes in the DNA sequences flanking the integrated viral DNA, transcriptional activation of remote cellular genes by HBV-encoded proteins (particularly the X protein), and effects resulting from viral mutations. The transcriptional activity of the HBV X protein may be mediated by interaction with specific transcription factors, activation of the mitogen-activated protein kinase and Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathways, an effect on apoptosis, and modulation of DNA repair. Studies have shown a clear link between the amount of HBV replication (measured as serum level of HBV DNA [viral load]) and subsequent risk of HCC. The long-term risk of HCC increases markedly in patients with serum HBV DNA levels higher than 10 ⁴ copies/mL. A randomized controlled trial of antiviral therapy has also shown a reduction in the incidence of HCC in association with reductions in serum levels of HBV DNA during therapy (see later), although other studies have not been able to confirm this benefit.

Indirect carcinogenic effects are the result of the chronic necroinflammatory hepatic disease, in particular cirrhosis, induced by the virus. The increased hepatocyte turnover rate resulting from continuous or recurring cycles of cell necrosis and regeneration acts as a potent tumor promoter. In addition, the distorted architecture characteristic of cirrhosis contributes to the loss of control of hepatocyte growth, and hepatic inflammation generates mutagenic reactive oxygen species. Data from the REVEAL (Risk Evaluation of Viral Load Elevation and Associated Liver Disease/Cancer)-HBV study in Taiwan have shown that genotype C of HBV and specific alleles of the basal core promoter and precore regions of the HBV genome are associated with a higher risk of HCC, whereas in Alaska, genotype F has been more strongly associated with HCC. The transgenic mouse model of Chisari and coworkers has provided indirect support for the role of prolonged hepatocyte injury in hepatocarcinogenesis. The REACH-B (Risk Estimation for Hepatocellular Carcinoma in Chronic Hepatitis B) score provides a simple-to-use tool for risk estimation for HCC among individuals with chronic HBV infection and includes gender, age, serum ALT level, hepatitis B e antigen status, and serum HBV DNA level.

HCV

Approximately 71 million people in the world today are chronically infected with HCV and are at greatly increased risk for the development of HCC. In Japan, Italy, and Spain, HCV is the single most common etiologic factor for HCC, and in other industrialized countries, HCV infection, often in combination with alcohol abuse, has emerged as a major cause of the malignancy. ^, Patients with HCV-induced HCC generally are older than those with HBV-related tumors, and the HCV infection is likely acquired mainly in adult life.

Almost all HCV-induced HCCs arise in cirrhotic livers, and most of the exceptions are in livers with chronic hepatitis and fibrosis. This observation strongly suggests that chronic hepatic parenchymal disease plays a key role in the genesis of HCV-related tumors. Because the HCV genome does not integrate into host DNA, the virus would have to exert a direct carcinogenic effect by some other means.

Long-term follow-up of a large group of patients with chronic hepatitis C and cirrhosis or bridging fibrosis found a cumulative 5-year frequency of HCC of just over 5%. The rate was higher among those with cirrhosis (7.0%) than those with bridging fibrosis at baseline (4.1%). A multivariate analysis model showed that older age, black race, lower platelet count, presence of esophageal varices, and smoking were additional risk factors.

It has become apparent that successful treatment of chronic HCV infection, with a sustained virologic response (see Chapter 80 ), is associated with regression of hepatic fibrosis and a lower-than-expected rate of HCC. ^, ^, Modern DAAs have increased the rate of viral cure, although HCC may still occur in a cirrhotic patient after treatment has eliminated HCV ^, Long-term maintenance therapy with peginterferon was not successful in preventing HCC in patients with chronic hepatitis C. ^,

Cirrhosis

In all parts of the world, HCC frequently occurs against a background of cirrhosis. All causative forms of cirrhosis may be complicated by tumor formation. A long-term follow-up study of 2126 U.S. military veterans with cirrhosis found that HCC developed in 100 (4.7%) over an average period of 3.6 years. The calculated rate was 1.3/100 patient-years. Risk factors for HCC included obesity, a low platelet count, and the presence of antibody to hepatitis B core antigen. A similar study from Italy found an incidence of HCC of 3.7/100 patient-years among cirrhotic persons with HCV infection and 2.0/100 patient-years among persons with HBV infection. Older age and male gender were confirmed as risk factors among patients with cirrhosis. By contrast, a study from Denmark of more than 8000 patients with alcohol-associated cirrhosis found a 5-year cumulative risk of HCC of 1.0, suggesting that perhaps patients with this form of cirrhosis were at lower risk of HCC than, for example, those with HCV-related cirrhosis.

Aflatoxin B ₁

Dietary exposure to aflatoxin B ₁ , derived from the fungi Aspergillus flavus and Aspergillus parasiticus , is an important risk factor for HCC in parts of Africa and Asia. These molds are ubiquitous in nature and contaminate staple foodstuffs in tropical and subtropical regions (see Chapter 89 ). Epidemiologic studies have shown a strong correlation between the dietary intake of aflatoxin B ₁ and incidence of HCC. Moreover, aflatoxin B ₁ and HBV interact synergistically in the pathogenesis of HCC. Heavy dietary exposure to aflatoxin B ₁ may contribute to hepatocarcinogenesis through an inactivating mutation of the third base of codon 249 of the TP53 tumor suppressor gene. ^,

Other Conditions

HCC develops in as many as 45% of patients with untreated hemochromatosis (see Chapter 75 ). Malignant transformation was previously thought to occur only in the presence of cirrhosis (and is certainly more likely to do so), but this complication also has been reported in patients without cirrhosis. Excessive free iron in tissues may be carcinogenic, perhaps by generating mutagenic reactive oxygen species. Further support for this theory comes from the observations that black Africans with dietary iron overload are at increased risk of HCC and that rats fed a diet high in iron develop iron-free dysplastic foci and HCC in the absence of cirrhosis. HCC occasionally develops in patients with Wilson disease, but only in the presence of cirrhosis (see Chapter 76 ). Malignant transformation has been attributed to the cirrhosis but may also result from oxidant stress secondary to the accumulation of copper in the liver. HCC also may develop in patients with other inherited metabolic disorders that are complicated by cirrhosis, such as α ₁ -antitrypsin deficiency and type 1 hereditary tyrosinemia, and in patients with certain inherited diseases in the absence of cirrhosis—for example, type 1 glycogen storage disease (see Chapter 77 ). HCC develops in approximately 40% of patients with membranous obstruction of the inferior vena cava, a rare congenital or acquired anomaly (see Chapter 85 ).

The roles of obesity, diabetes mellitus, and NAFLD have come to be recognized in the causation of HCC, although the mechanisms whereby these overlapping conditions contribute to the development of HCC are unknown. Cirrhosis caused by NASH appears to give rise to HCC less frequently than cirrhosis caused by HCV but nevertheless appears to carry significant risk. Diabetes mellitus is also a risk factor for HCC, although it is not clear if the risk is independent of NAFLD or not.

A statistically significant correlation between the use of oral contraceptive steroids and the occurrence of HCC has been demonstrated in countries in which the incidence of HCC is low and no overriding risk factor for development of the tumor is present. Epidemiologic evidence of a link between cigarette smoking and the occurrence of HCC is conflicting, although most of the evidence suggests that smoking is a minor risk factor ; heavy smokers have an approximately 50% higher risk than nonsmokers. The incidence of HCC is increased in patients with HIV infection compared with controls in the general population, presumably because of the increased rate of chronic viral hepatitis in the HIV-positive population.

Although the aforementioned risk factors have been identified, the precise mechanisms whereby they lead to HCC still need to be elucidated. Multiple cellular pathways are involved in causing unconstrained proliferation of hepatocytes and increased angiogenesis against a background of chronic liver disease. These pathways have become the targets for new molecular therapies against HCC ( Box 96.2 ) (see later).

BOX 96.2

Key Molecular Pathways Involved in Hepatocarcinogenesis

Adapted from Roberts L. Emerging experimental therapies for hepatocellular carcinoma: what if you can’t cure? In: McCullough A, editor. AASLD Postgraduate Course, 2007. Boston: AASLD; 2007. p 185.

JAK/STAT, janus kinase/signal transducer and activator of transcription; mTOR, mechanistic (or mammalian) target of rapamycin.

Angiogenic signaling
Epigenetic promoter methylation and histone acetylation
Growth factor-stimulated receptor tyrosine kinase
JAK/STAT signaling
PI3-kinase/AKT/mTOR
p53 and cell cycle regulation
Ubiquitin-proteasome
Wnt/β-catenin

Clinical Features

Although the typical clinical features of HCC are well recognized (including abdominal pain and weight loss in patients with cirrhosis), many patients are now diagnosed at an early stage when they have no specific symptoms or signs. This trend toward earlier diagnosis is probably the result of surveillance programs in patients with chronic liver disease (see later). In advanced disease, patients with HCC usually present with typical symptoms and signs, and diagnosis is straight forward. In addition, HCC often coexists with cirrhosis, and the onset of HCC is marked by a sudden unexplained change in the patient’s condition.

Patients with HCC often are unaware of its presence until the tumor has reached an advanced stage. The most common (and frequently first) symptom is right hypochondrial or epigastric pain. Other symptoms are listed in Table 96.1 .

TABLE 96.1

Symptoms and Signs of HCC

Symptom	Frequency (%)
Abdominal pain	59-95
Weight loss	34-71
Weakness	22-53
Abdominal swelling	28-43
Nonspecific GI symptoms	25-28
Jaundice	5-26
Sign
Hepatomegaly	54-98
Ascites	35-61
Fever	11-54
Splenomegaly	27-42
Wasting	25-41
Jaundice	4-35
Hepatic bruit	6-25

Physical findings vary with the stage of disease (see Table 96.1 ). Early in the course, evidence of cirrhosis alone may be present, or abnormal findings may be absent. When the tumor is advanced at the time of the patient’s first medical visit, the liver is almost always enlarged, sometimes massively. Hepatic tenderness is common and may be profound, especially in the later stages. The surface of the enlarged liver is smooth, irregular, or frankly nodular. An arterial bruit may be heard over the tumor ; the bruit is heard in systole, rough in character, and not affected by changing the position of the patient. Although not pathognomonic, a bruit is a useful clue to the diagnosis of HCC. Less often, a friction rub may be heard over the tumor, but this sign is more characteristic of hepatic metastases or abscesses.

Ascites may be present when the patient is first seen or may appear with progression of the tumor. In most patients, ascites is the result of long-standing cirrhosis and portal hypertension (see Chapter 93 ), but in some cases it is caused by invasion of the peritoneum by the primary tumor or metastases or obstruction of the hepatic veins or superior vena cava. The ascitic fluid may be blood stained. Splenomegaly, if present, reflects coexisting cirrhosis and portal hypertension.

Physical evidence of cirrhosis may also be noted. Severe pitting edema of the lower extremities extending up to the groins occurs when HCC has invaded the hepatic veins and propagates into and obstructs the inferior vena cava. A Virchow-Trosier (supraclavicular) node, Sister Mary Joseph’s (periumbilical) nodule, or enlarged axillary lymph node is rarely present.

Paraneoplastic Manifestations

Some of the deleterious effects of HCC are not caused by local effects of the tumor or metastases ( Box 96.3 ). Each of the paraneoplastic syndromes in HCC is rare or uncommon. One of the more important is type B hypoglycemia, which occurs in less than 5% of patients, manifests as severe hypoglycemia early in the course of the disease, and is believed to result from the defective processing by malignant hepatocytes of the precursor to insulin-like growth factor II (pre-IGF II). By contrast, type A hypoglycemia is a milder form of glycopenia that occurs in the terminal stages of HCC (and other malignant tumors of the liver). It results from the inability of a liver extensively infiltrated by tumor, and often cirrhotic, to satisfy the demands for glucose by a large, often rapidly growing tumor and by the other tissues of the body.

BOX 96.3

Paraneoplastic Manifestations Associated with HCC

Carcinoid syndrome
Hypercalcemia
Hypertension
Hypertrophic osteoarthropathy
Hypoglycemia
Neuropathy
Osteoporosis
Polycythemia (erythrocytosis)
Polymyositis
Porphyria
Sexual changes—isosexual precocity, gynecomastia, feminization
Thyrotoxicosis
Thrombophlebitis migrans
Watery diarrhea syndrome

Another important paraneoplastic syndrome is polycythemia (erythrocytosis), which occurs in less than 10% of patients with HCC. This syndrome appears to be caused by the synthesis of erythropoietin or an erythropoietin-like substance by malignant hepatocytes.

Patients with HCC, especially the sclerosing variety, may present with hypercalcemia in the absence of osteolytic metastases. When hypercalcemia is severe, it may result in the typical complications of hypercalcemia, including drowsiness and lethargy. The probable cause is secretion of parathyroid hormone–related protein by the tumor.

Cutaneous paraneoplastic manifestations of HCC are rare except for pityriasis rotunda (circumscripta), which may be a useful marker of the tumor in black Africans. The rash consists of single or multiple, round or oval, hyperpigmented, scaly lesions on the trunk and thighs that range in diameter from 0.5 to 25 cm.

Diagnosis

The gold standard for the diagnosis of HCC is pathology. For practical purposes (i.e., to apply treatment), HCC can be diagnosed in the presence of an abnormality on imaging of the liver. Dysplastic nodules and even regenerative cirrhotic nodules can be seen on imaging studies and are potentially confused with HCC. Although enhancement patterns with dynamic imaging of dysplastic nodules and HCC are fairly specific (see later), some overlap occurs. ^, Nevertheless, there is a growing consensus that, based on guidelines from the major European and American hepatology societies and now backed by published experience, the diagnosis of HCC can be made in the appropriate clinical setting on the basis of specific imaging characteristics, with or without an elevated serum AFP level.

Serum Tumor Markers

Serum tumor markers generally are not diagnostic for HCC by themselves but can be used in conjunction with imaging findings to diagnose HCC. Additionally, they may raise the suspicion of HCC and lead to more sensitive and serial imaging of the liver. Conventional liver biochemical tests do not distinguish HCC from other hepatic mass lesions or cirrhosis.

Many of the substances synthesized and secreted by HCC are not biologically active. Nevertheless, a few are produced by a sufficiently large proportion of tumors to warrant their use as serum markers for HCC. The most helpful of these markers is AFP.

AFP

AFP is an α ₁ -globulin normally present in high concentrations in fetal serum but in only minute amounts thereafter. Reappearance of high serum levels of AFP strongly suggests the presence of HCC (or hepatoblastoma [see later]), especially in populations at risk for HCC.

Measurement of AFP can potentially be used for the diagnosis of HCC, surveillance, and prognostication. With regard to diagnosis, existing guidelines are based on biopsy or liver imaging and do not require use of AFP. Clearly, markedly elevated AFP levels (>10,000 ng/mL to > 1,000,000 ng/mL) can be considered diagnostic for HCC in an appropriate clinical context. Although there is no specific diagnostic cutoff, values above 400 ng/mL in association with a liver mass can be considered diagnostic in most cases.

In the context of surveillance for HCC, the tumor must be detected at an early stage when potentially curative treatment can still be applied. Measurement of AFP has been used for early diagnosis but with sometimes disappointing results. For example, Marrero and colleagues studied a large group of patients with HCC and matched controls and found that the optimal cutoff value of serum AFP level that resulted in the greatest sensitivity was 10.9 ng/mL; still, the sensitivity of the test using this value was only 66%. Therefore, routine use of AFP as part of a surveillance program for HCC remains controversial.

Serum AFP levels appear to have some prognostic utility, particularly with regard to LT, for which levels above 1000 ng/mL have been associated with poorer outcomes and higher rates of tumor recurrence. An AFP level higher than about 500 ng/mL predicts worse outcomes with LT compared with lower levels. Attempts to correlate the degree of differentiation of HCC with production of AFP have produced conflicting results.

False-positive AFP results (for HCC) also may occur in patients with tumors of endodermal origin, non-seminomatous germ cell tumors, pregnancy, and regenerating livers in the setting of ALF. A progressively rising serum AFP concentration is highly suggestive of HCC. Because both false-positive and false-negative results are obtained when AFP is used as a serum marker for HCC, the search for an ideal marker continues; however, alternative markers have not proved to be more useful than AFP.

Fucosylated AFP

AFP is heterogeneous in structure. Its microheterogeneity results from differences in the oligosaccharide side chain and accounts for the differential affinity of the glycoprotein for lectins. AFP secreted by malignant hepatocytes contains unusual and complex sugar chains that are not found in AFP produced by non-transformed hepatocytes. One variant, Lens culinaris agglutinin reactive fraction (AFP-L3), has been suggested to improve the specificity of AFP, particularly AFP serum levels from 10 to 200 ng/mL. ^, The recommended cutoff value for AFP-L3 to diagnose HCC is higher than 10%, although the specificity varies depending on the absolute level of AFP. Studies have not confirmed that AFP-L3 has greater sensitivity or specificity than AFP alone for the diagnosis of early HCC. ^, Therefore, AFP-L3 is not sufficiently validated to confirm the diagnosis of HCC without other supporting findings, such as suggestive imaging.

Des-γ-Carboxy Prothrombin

Serum concentrations of des-γ-carboxy prothrombin (DCP) (also known as prothrombin produced by vitamin K absence or antagonist II) are raised in most patients with HCC. DCP is an abnormal prothrombin that is thought to result from a defect in the post-translational carboxylation of the prothrombin precursor in malignant cells. DCP has been suggested to be a better marker than, or at least complementary to, AFP. A large study in Western patients with HCV–related cirrhosis, however, did not confirm this finding. Therefore, because appropriate diagnostic cutoff values are not well established, the precise role of DCP in the diagnosis of HCC still requires validation.

Other Markers

Multiple other potential serum markers for HCC have been identified, although none of them has an established high–through-put method of measurement, as required for a clinical test. The roles in the diagnosis of HCC for markers such as glypican-3 (GPC3), Golgi protein 73, hepatocyte growth factor, IGF 1, and transforming growth factor-β1 await further study.

Imaging

The diagnosis of HCC generally requires imaging evidence of a focal lesion in the liver, although large infiltrating lesions can also be diagnostic. Arterial hyperenhancement, particularly seen on dynamic contrast imaging of the liver, is observed because the blood supply of HCC comes from newly formed abnormal arteries (neoangiogenesis). ^, ^, As a nodule transforms from low- to high-grade dysplasia and then to HCC, the primary blood supply shifts from portal to arterial; new abnormal arterial branches produce characteristic findings on dynamic contrast imaging of the liver and subsequent hypoenhancement in the portal venous and delayed phases (“washout”). ^, European and American liver societies recommend that a noninvasive diagnosis of HCC can be made in a nodule greater than 1 cm in diameter that demonstrates arterial hyperenhancement and portal venous or delayed washout. ^,

The American College of Radiology created and updated the Liver Imaging Reporting and Data System (LI-RADS), which attempts to classify liver nodules based on size and imaging characteristics on CT or MRI and has been adapted as the terminology to be used for patients on the UNOS transplant list. The LI-RAD categories assist the clinician in assessing the risk that a nodule is HCC, with LI-RAD 3 being intermediate risk, LI-RAD 4 probable HCC, and LI-RAD 5 definite HCC. The individual criteria for LI-RADS have been validated in prospective and retrospective cohort studies, but the system as a whole has not been fully validated and some studies show little difference between LI-RAD 4 and LI-RAD 5 nodules less than 2 cm in diameter if identified initially on US.

US

US detects most HCCs but may not distinguish this tumor from other solid lesions in the liver. Therefore, US is a more effective as a tool for screening than for diagnosis. As with all imaging methods, the sensitivity increases with increasing size of the lesion. A systematic review of 8 studies using histologic reviews of liver explants has shown that US has fair sensitivity (pooled estimate, 48%; 95% confidence interval [CI], 34% to 62%) with good specificity (97%; 95% confidence interval [CI], 95% to 98%). Advantages of US include safety, availability, and cost-effectiveness. Drawbacks include lack of standardization, examiner dependence, and limited sensitivity with certain body habituses, particularly obesity, and with fatty infiltration of the liver.

The US appearance of HCC is variable because it is influenced by the presence of fat, hemorrhage, and necrosis. Smaller tumors (<5 cm) are most often hypoechoic and may demonstrate a thin peripheral fibrous capsule. Small HCCs can also be uniformly hyperechoic and therefore indistinguishable from focal fat or a hemangioma. With increased size there is generally increased complexity of the nodule. Tumors located immediately under the right hemidiaphragm may be difficult to detect. US with Doppler technology is useful for assessing the patency of the inferior vena cava, portal vein and its larger branches, hepatic veins, and biliary tract.

Dynamic contrast-enhanced Doppler US with IV infusion of CO ₂ microbubbles viewed with grayscale imaging and color Doppler US are refinements that, by characterizing hepatic arterial and portal venous flow in tumorous nodules, facilitate the diagnosis of malignant and benign hepatic nodules. These techniques are generally not performed in the USA, owing to lack of approval by the FDA for noncardiac studies.

CT

Multiphase (also called dynamic) multidetector CT is the most popular imaging technique for the diagnosis of HCC. ^, ^, In order to rely on CT or MRI for the diagnosis of HCC, certain technical specifications for imaging equipment, image acquisition, and dynamic contrast timing are necessary. ^, Dynamic contrast-enhanced CT can include noncontrast, arterial, portal venous, and delayed phases. The classic and most diagnostic pattern for HCC is a combination of hyperenhancement in the arterial phase (with the uninvolved liver lacking enhancement), loss of central nodule enhancement compared with the enhancing uninvolved liver (washout), and capsular enhancement in the portal-venous and delayed phases ( Fig. 96.2 ). ^, When the lesion is larger than 2 cm in diameter, this pattern has almost 100% specificity for HCC. When the nodule is 1 to 2 cm, a diagnosis of HCC or high-grade dysplastic nodule can be made with a specificity greater than 95%. ^, CT often finds so-called hypervascular-only lesions, which enhance in the arterial phase and become isodense to the surrounding liver in the portal-venous and delayed phases. These lesions may be dysplastic nodules, arterial-portal shunts, atypical hemangiomas, HCC, intrahepatic cholangiocarcinoma, confluent fibrosis, or aberrant venous drainage. Only about 30% of nodules less than 2 cm in diameter are HCCs. Both HCCs and cholangiocarcinomas grow over time, whereas other nodules disappear or remain stable on follow-up studies. HCC may also have other patterns on CT, such as washout only on delayed imaging, a hypovascular nodule, or a fat-containing nodule. ^, Guidelines recommend biopsy of lesions larger than 1 cm and serial imaging for lesions smaller than 1 cm that do not have characteristic arterial enhancement and washout. Overall, the pooled estimates of sensitivity and specificity of CT for detecting HCC are 67.5% (95% CI, 55% to 80%) and 92.5% (95% CI, 89% to 96%), respectively. Dynamic CT is also useful for detecting invasion into the portal or hepatic veins and identifying the location and number of tumors; these findings are critical for planning treatment (see later).

Fig. 96.2, Dynamic CT of a patient with HCC showing no lesion in the noncontrast phase, an enhancing lesion in the right lobe of the liver in the arterial phase of contrast administration (arrow) , and a faint lesion in the portal venous phase, seen better in the delayed phase.

MRI

Dynamic MRI using gadolinium contrast agents (extracellular) provides another way of distinguishing HCC from normal liver tissue. The performance of MRI and the findings on multiphase contrast enhancement are similar or perhaps slightly superior to those described for CT ( Fig. 96.3 ). Hyperintensity of a nodule on T2-weighted images is specific for HCC. ^, The pooled estimates of sensitivity and specificity of MRI for detecting HCC are 80.6% (95% CI, 70% to 91%) and 84.8% (95% CI, 77% to 93%), respectively. Although MRI may be slightly superior overall to CT, especially with regard to sensitivity, local expertise and patient factors (ability to hold breath in a confined space, presence of large amount of ascites and renal function) should dictate the choice of imaging technique. Close attention to technical specifications is essential. Findings using newer techniques that may improve the specificity of MRI for HCC, particularly those with atypical vascular-enhancement patterns, include hyperintensity on diffusion-weighted images and lack of enhancement on late images using a hepatobiliary-specific contrast agent (gadoxetic acid). LI-RADS is widely used as a way of categorizing nodules recognized on CT or MRI, in patients at high risk of HCC, as definitely benign, probably benign, having an intermediate probability of being HCC, probably HCC, and definitely HCC (corresponding to LI-RADS categories 1 to 5, respectively) (see earlier).

Fig. 96.3, Multiphase MRI of the liver showing HCC with characteristic features, including hyperintensity ( arrow ) on a T2 -weighted image ( top left panel ) but not on a T1 -weighted image ( top right panel ), enhancement during the arterial phase of contrast administration ( bottom left panel ), with central washout of contrast and capsular enhancement during the venous and delayed phases ( bottom middle and right panels ).

PET

Whole-body fluorine-18-fluorodeoxyglucose PET combined with CT (PET-CT) may have a role in the evaluation of some patients with HCC. The sensitivities of dynamic CT and MRI are superior to that of PET-CT. Several retrospective case series have shown that high avidity in the primary hepatic lesion predicts an increased risk of recurrence after potentially curative treatment. Once a diagnosis of HCC is made, staging involves imaging of the chest, usually with noncontrast CT, and imaging of other areas of the body based on clinical symptoms. Particularly if the tumor within the liver is beyond the Milan criteria (see later), either a bone scan or PET-CT can sometimes identify an unrecognized extrahepatic metastasis that would change the treatment plan. The use of PET-CT in HCC needs further study.

Hepatic Angiography

Since the advent of CT and MRI, the role of diagnostic hepatic angiography has been limited. Digital subtraction angiography is helpful for recognizing small hypervascular HCCs but may miss hypovascular tumors. Findings in HCC include arteries that are irregular in caliber and do not taper in the usual way, with smaller branches showing a bizarre pattern and delay in capillary emptying, which is seen as a blush. Angiography is essential for delineating the hepatic arterial anatomy in planning bland embolization, chemoembolization, and radioembolization of the tumor or infusion of cytotoxic drugs directly into the hepatic artery or its branches (see later).

Laparoscopy

Laparoscopy is now rarely performed for this purpose but can be used to detect peritoneal and other extrahepatic spread, ascertain whether the nontumorous part of the liver is cirrhotic, and obtain biopsies under direct vision.

Pathology

Definitive diagnosis of HCC depends on demonstrating the typical histologic features. Suitable samples generally can be obtained by percutaneous biopsy or FNA. The yield and safety of the procedure can be increased by directing the needle under US or CT guidance. Laparoscopically directed biopsy is an alternative approach. Needle biopsy of the tumor carries a small but definite risk of spread along the needle track. Pathologic diagnosis of HCC is based on the recommendations of the International Consensus Panel. Immunostaining for GPC3, heat shock protein HSP70, and glutamine synthetase or gene expression profiling ( GPC3 , LYVE1 [encoding lymphatic vessel endothelial hyaluronan receptor-1], BIRC5 [encoding baculoviral inhibitor of apoptosis repeat-containing-5, or survivin]) or both, is recommended to differentiate high-grade dysplastic nodules from early HCC.

Gross Appearance

HCC may take 1 of 3 forms: nodular, massive, or diffusely infiltrating. The nodular variety of HCC is most common and usually coexists with cirrhosis. It is characterized by numerous round or irregular nodules of various sizes scattered throughout the liver; some of the nodules are confluent. The massive type is characterized by a large circumscribed mass, often with small satellite nodules. This type of tumor is most prone to rupture and is more common in younger patients with a noncirrhotic liver. In the rare diffusely infiltrating variety, a large part of the liver is infiltrated homogeneously by indistinct minute tumor nodules, which may be difficult to distinguish from the regenerative nodules of cirrhosis that are almost invariably present. The portal vein and its branches are infiltrated by tumor in up to 70% of cases seen at autopsy; the hepatic veins and bile ducts are invaded less often.

Microscopic Appearance

HCC is classified histologically into well-differentiated, moderately-differentiated and undifferentiated (pleomorphic) forms, as well as progenitor cell HCC and fibrolamellar HCC (see later).

Well-Differentiated

Despite the aggressive nature and poor prognosis of HCC, most tumors are well-differentiated. Trabecular and acinar (pseudoglandular) varieties occur, sometimes in a single tumor. In the trabecular variety, the malignant hepatocytes grow in irregular anastomosing plates separated by often inconspicuous sinusoids lined by flat cells resembling Kupffer cells. The trabeculae resemble those of normal adult liver but often are thicker and may be composed of several layers of cells. Scanty collagen fibers may be seen adjacent to the sinusoid walls. The malignant hepatocytes are polygonal, with abundant, slightly granular cytoplasm that is less eosinophilic than that of normal hepatocytes. The nuclei are large and hyperchromatic with prominent nucleoli. Bile production is the hallmark of HCC, regardless of the pattern. Gland-like structures are present in the acinar variety. The structures are composed of layers of malignant hepatocytes surrounding the lumen of a bile canaliculus, which may contain inspissated bile. A tubular or pseudopapillary appearance may be produced by degeneration and loss of cells, or cystic spaces may form in otherwise solid trabeculae. The individual cells may be more elongated and cylindrical than in the trabecular variety.

Moderately-Differentiated

Solid, sarcomatous, scirrhous, and clear cell varieties of HCC are described, as well as HCC with lymphoid stroma. In the solid variety, the cells usually are small, although they vary considerably in shape. Pleomorphic multinucleated giant cells are occasionally present. The tumor grows in solid masses or cell nests. Evidence of bile secretion is rare, and connective tissue is inconspicuous. Central ischemic necrosis is common in larger tumors. In the scirrhous variety, the malignant hepatocytes grow in narrow bundles separated by abundant fibrous stroma. Duct-like structures are occasionally present. In most tumors, the cells resemble hepatocytes. In an occasional tumor, the malignant hepatocytes are predominantly or exclusively clear cells. More often, tumors contain areas of clear cells. The appearance of these cells results from a high glycogen or, in some cases, fat content.

Undifferentiated

In undifferentiated HCC, cells are pleomorphic and vary greatly in size and shape. The nuclei are also extremely variable. Large numbers of bizarre-looking giant cells are present and may be spindle shaped, resembling those of sarcomas. Globular hyaline structures may be seen in all types of HCC. These structures reflect the presence of AFP, α1-antitrypsin, or other proteins. Mallory’s hyaline is occasionally present.

Progenitor Cell HCC

A class of HCC appears to have its origins in progenitor cells, the stem cells of the liver, located in association with the canals of Hering (see Chapter 71 ). Progenitor cell activation is seen in association with chronic viral hepatitis and cirrhosis, presumably related to senescence of hepatocytes. These tumors may appear morphologically like typical HCC or mixed cholangiohepatocellular carcinoma. Tumor cells stain positively for cytokeratin (CK) 19, and the tumor appears to have a more aggressive course than typical HCC.

Metastases

Extrahepatic metastases are present at autopsy in 40% to 57% of patients with HCCs. ^, The most common sites are the lungs (up to 50% in some reports) and regional lymph nodes (≈20%). The adrenal glands are also frequently involved.

Fibrolamellar HCC

Fibrolamellar HCC is a distinct variant of HCC that typically occurs in young patients, has an approximately equal gender distribution, does not secrete AFP, is not caused by chronic hepatitis B or C, and almost always arises in a noncirrhotic liver. The hepatocytes are characteristically plump, deeply eosinophilic, and encompassed by abundant fibrous stroma composed of thin, parallel fibrous bands that separate the cells into trabeculae or nodules. The cytoplasm is packed with swollen mitochondria and, in approximately half of the tumors, contains pale or hyaline bodies. Nuclei are prominent, and mitoses are rare. Fibrolamellar HCC has different immunohistochemical characteristics than usual HCC, occurring either with or without cirrhosis; therefore, fibrolamellar HCC is much less likely to stain positively for GPC3, although expression of CK7 is more abundant. Fibrolamellar HCC is more often amenable to surgical treatment and therefore generally carries a better prognosis than conventional HCC. It does not, however, respond to chemotherapy any better than other forms of HCC. A mixed form of fibrolamellar HCC has been described, in which some areas of the tumor have the histologic appearance of usual HCC and others resemble fibrolamellar HCC. This mixed type seems to behave more like usual HCC and has a poorer prognosis than typical fibrolamellar HCC.

Staging

Accurate staging of HCC is necessary for prognostication and for selection of therapy. Determining the optimal staging system for HCC has been controversial, in part because staging has to take into account both the severity of the underlying liver disease and the size and degree of spread of the tumor. As with all cancers, the TNM system can be used to stage HCC, but this system does not account for the underlying liver disease. A study comparing the usefulness of 7 staging systems, including the Okuda, TNM, Cancer of the Liver Italian Program, Barcelona Clinic Liver Cancer (BCLC), Chinese University Prognostic Index, Japanese Integrated Staging, and Group d’Etude et Traitement du Carcinome Hépatocellulaire systems in a cohort of patients from the USA, found that the BCLC staging system ( Fig. 96.4 ) had the best independent predictive power for survival. The BCLC system has been adopted by the AASLD for use in its practice guidelines on management of HCC. This staging classification also includes a treatment schedule based on stage. In addition to staging of the cancer, several systems are in use to stage the degree of liver damage, which is often a limiting factor in applying potentially curative treatments. The Child-Pugh classification has been widely used (see Chapter 92 ), and the albumin-bilirubin (ALBI) grade is an objective score that can also assist in treatment planning.

Natural History and Prognosis

Symptomatic HCC carries a grave prognosis; in fact, the annual incidence and mortality rates for the tumor are almost identical. The main reasons for the poor outcome are the extent of tumor burden when the patient is first seen and the frequent presence of coexisting cirrhosis and hepatic dysfunction. The natural history of HCC in its florid form is one of rapid progression, with increasing hepatomegaly, abdominal pain, wasting, and deepening jaundice, and with death ensuing in 2 to 4 months. In industrialized countries, however, the tumor appears to run a more indolent course with longer survival times. Rare cases of spontaneous tumor regression have been reported (see later). When HCC is detected at an early stage, several options for treatment are available and often lead to prolonged survival.

Treatment

Important advances in the treatment of HCC have occurred since the 1980s and have resulted in improvement of the USA population-based 5-year survival rate to 24.5% ; these advances include randomized controlled trials that support the benefits of certain treatments such as chemoembolization and the advent of the multikinase inhibitor sorafenib. Additional treatments for advanced HCC include newer multikinase inhibitors and immune checkpoint inhibitors, including regorafenib, lenvatinib, cabozantenib, ramucirimab, nivolumab, and pembrolizumab. Overwhelming evidence supports the superiority of LT over other therapies for patients with portal hypertension and cirrhosis (see Chapter 97 ). Because HCC is usually a combination of 2 diseases—the underlying liver disease (usually cirrhosis with varying degrees of decompensation) and the cancer itself—both factors must be considered when selecting treatment. When presented with a patient with HCC, the clinician should decide which is the best initial therapy: surgical resection or LT, if the patient is a candidate for either; ethanol or radiofrequency ablation (RFA), if possible, based on the size of the tumor; chemoembolization; and, if the tumor is too advanced, targeted chemotherapy. Table 96.2 describes the treatment options for HCC. The BCLC staging classification and treatment schedule can help guide the clinician in choosing the most appropriate treatment (see Fig. 96.4 ).

TABLE 96.2

Treatment Options for HCC

Modality	Comments
Surgical resection	Curative but limited to noncirrhotic patients and cirrhotic patients without portal hypertension May be technically difficult High recurrence rate
LT	Successful in selected patients (Milan criteria; see text and Chapter 97 ) Requires lifelong immunosuppression Expensive and not available worldwide
Radiofrequency ablation or ethanol injection	Potentially curative for small tumors, including multiple tumors High recurrence rate
Transarterial chemoembolization	Prolongs survival in unresectable tumors if hepatic function is preserved; not curative
Chemotherapy	No clear benefit; palliative only Drug toxicity is common
Targeted molecular therapies	Sorafenib is the first such agent shown to improve patient survival Improvement in patient survival with lenvatinib is similar to that with sorafenib Regorafenib, cabozantinib, and ramucirumab (if AFP >400 ng/mL) improve survival after sorafenib failure
Immune checkpoint inhibitors	Nivolumab and pembrolizumab are associated with improved survival after failure of or intolerance to sorafenib

Surgical Resection

Surgical therapy, whether by tumor resection or LT, offers the best chance of cure for HCC. For resection to be considered, the tumor should be confined to one lobe of the liver and favorably located, and ideally, the nontumorous liver tissue should not be cirrhotic. Expert surgical centers can achieve 5- and 10-year survival rates of 40% and 26%, respectively, with a mean tumor diameter of 8.8 cm in noncirrhotic patients. Unfortunately, these patients represent less than 5% of Western cases. ^, Resection is also effective if the tumor is limited to the left lobe or a portion of the right lobe, thereby permitting a segmental resection if the patient has Child-Pugh class A cirrhosis, the serum bilirubin level is normal, and portal hypertension is not present (based on imaging, a normal platelet count, absence of varices on endoscopy, and a directly measured hepatic venous pressure gradient <10 mm Hg). Using these criteria, 5-year survival rates of 50% or better can be achieved. Patients with smaller and solitary tumors have better outcomes. In parts of the world where LT is not available, surgical resection is a viable option, particularly for Child-Pugh class A patients without portal hypertension and with a MELD score of 9 or less (see Chapter 97 ). All the tumor nodules need to be removed, with a negative margin of resection, and the patient needs to be left with enough functional liver volume (usually defined as ≥40% in a patient with cirrhosis) to survive the postoperative period. Overall, resection is feasible in only approximately 15% of patients. Resection performed at expert surgical centers carries an operative mortality rate of less than 5%, but at low-volume centers the mortality rate is almost 3 times greater. Unfortunately, the rate of recurrence after resection is more than 50% in the long term, and salvage LT is rarely possible.

LT

LT is performed in patients in whom the tumor is not resectable but is confined to the liver or in whom advanced cirrhosis and poor liver function preclude resection (see Chapter 97 ). LT is the ideal therapy for HCC because it provides the largest possible resection margin, removes the remaining liver, which is at high risk for de novo tumors, and replaces the dysfunctional liver. LT can fail in patients with extrahepatic tumor, which tends to grow rapidly under the influence of post-transplantation immunosuppression. Because the availability of donor livers is limited, the consensus is that anticipated outcomes of LT for HCC should be similar to those for other indications for LT and superior to those for other treatments for HCC. Several large series have demonstrated that if one selects candidates based on the Milan criteria—a single tumor up to 5 cm in size or 2 to 3 lesions, each up to 3 cm in size, with no large-vessel vascular invasion or metastasis—the 5-year survival rate is 70% to 75%, and the tumor recurrence rate is 10% to 15%. ^, These criteria led to the HCC MELD exception pathway, which was adopted in the USA in 2002. Because of the change, the frequency of HCC as an indication for LT rose from 4.6% to 26% of the total adult liver transplant population. Additionally, progression of the tumor beyond the Milan criteria before a patient undergoes LT has largely been eliminated. ^, If the estimated time to LT is greater than 6 months, bridging therapy with RFA or transarterial chemoembolization (TACE) can often be performed to prevent tumor growth beyond Milan criteria (see later). In other parts of the world, waiting times before transplantation remain critical, and when the waiting time increases to one year, as many as half of patients will not receive a transplant. An analysis of 4-year survival rates for all patients transplanted in the USA has confirmed that overall outcomes for those transplanted with HCC are only minimally worse than for those transplanted for other indications. Certain subgroups of patients do worse, including those with nodules 3 to 5 cm in diameter, a MELD score of 20 or greater, and a serum AFP level of 455 ng/mL or higher.

Some authorities have advocated expansion of the Milan criteria, provided that the tumor shrinks to within Milan criteria and remains stable for 3 months after application of locoregional therapy, based on prospective outcomes from small, single-center series, but these patients generally need a special exception from the regional review board in the USA. ^, Although these criteria are being widely applied, a larger multicenter study is needed to confirm the outcomes and define which patients would benefit.

Local Ablation

Local ablative therapies are potentially curative treatments for patients with small tumors (usually <3 to 5 cm in diameter) that are not amenable to resection or LT because of patient preference, the number and location of lesions, or significant hepatic dysfunction (Child-Pugh class B or C; see Fig. 96.4 ). The first of these techniques available was percutaneous ethanol injection (PEI), a relatively effective and safe method that is still used and is most effective for lesions smaller than 2 cm and effective in those up to 3 cm in diameter. PEI requires multiple sessions and, in patients with small tumors and preserved hepatic function, can lead to survival rates similar to those for surgical resection, although no randomized studies have been performed to demonstrate equivalent outcomes. Complications are rare and include tumor seeding of the needle track. RFA has generally supplanted PEI because it is more effective, particularly with larger tumors (most effective in lesions up to 3 cm and effective in those up to 5 cm), requires fewer sessions, and has similar complication rates. RFA can be performed percutaneously or by a laparoscopic or open surgical approach. Survival rates are similar to those for surgical resection, although recurrence rates are higher and complications are uncommon. ^, PEI is generally favored over RFA for lesions adjacent to a major vessel or large bile duct. Microwave ablation is a more recently developed thermal ablation technique for HCC that has the potential advantage over RFA of less heat sink loss to adjacent vessels and faster treatment times with similar clinical outcomes. Current UNOS rules require a 6-month waiting period to obtain MELD exceptions points for HCC, and, therefore, local ablative therapy with PEI or RFA is commonly applied even though benefit has not been well established by randomized trials.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here