and autopsy series estimated that FNH occurs in 4% to 8% of individuals, with a strong predilection for

women aged 20 to 50 years.43 Unlike the origin of hemangioma or hepatocellular adenoma, the origin

of FNH is thought to be due to hyperplastic growth of normal hepatocytes with a malformed biliary

draining system or a hyperplastic response to a pre-existing arteriovenous malformation. FNH does not

contain a portal venous supply.

Figure 60-6. Nodular enhancement typically observed in hemangioma in dynamic CT. The mass is gradually enhanced from the

peripheral part to the central part of the lesion.

Figure 60-7. Typical dynamic CT image of focal nodular hyperplasia. The lesion is homogeneously enhanced in the arterial phase

with central hypodensity area suggestive of a central scar (left). The mass becomes isodense in the portal phase (right).

In plain CT, the lesion is isoattenuated with normal liver and may exhibit the characteristic

hypoattenuated central scar. With contrast in the arterial phase, FNHs enhance brightly and become

again isoattenuated to adjacent parenchyma in the portal phase (Fig. 60-7). The central scar may

enhance on delayed phases. MRI reveals a similar dynamic enhancement pattern: FNHs appear

isointense on T1-weighted images and hyperintense on T2-weighted sequences. With gadolinium

enhancement, the lesion is homogeneously enhanced in the arterial phase and the central scar becomes

hypointense in the delayed phase.

The majority of FNHs is asymptomatic and found incidentally on abdominal imaging for other

indications. Because of the benign nature of FNHs and the very low risk of spontaneous rupture,

treatment should be reserved for patients with persistent symptoms not explained by another problem.

1552

Hepatic Adenoma

Hepatic adenoma (HA) is a rare benign tumor with an estimated prevalence of approximately 1% on

postmortem examinations.52 Typically, HAs occur in women of child-bearing age who often have a

history of long-term use of oral contraceptives. In the majority of the cases, HA lesions are solitary and

asymptomatic. However, the presence of multiple nodules (typically >10 nodules) indicates a distinct

entity called hepatic adenomatosis, which has a different implication for treatment.53

Adenomas are well-circumscribed lesions that contain sheets of hepatocytes without intervening

biliary ductules or portal tracts. Conventionally, HA was considered to have variable imaging

characteristics and was described as a heterogeneous hypervascular mass with areas of fat, hemorrhage,

and necrosis on multiphase CT or MRI.54 With the introduction of the new phenotype–genotype

classification of HA, however, correlation between the imaging characteristics and different subtypes

has been reported.55,56 HAs are grouped into 1 of 4 subtypes. HA-H (HNF1A-mutated hepatic adenoma)

is the most frequently occurring one (35% to 40%), presenting steatosis and absence of inflammatory

infiltrate or cytologic atypia. This subtype is frequently associated with somatic mutation of the TCF1

(HNF1A) gene or heterozygous germline mutations of the CYP1B1 gene. HA-B (β-catenin–mutated

hepatic adenoma, 15% to 19%) is morphologically characterized by pseudoacinar formation and mild

cytologic atypia. Steatosis is rare and inflammation is absent in this subtype. β-Catenin gene-activating

mutations are frequently observed in this subtype. HA-I (inflammatory hepatic adenoma, 30% to 35%)

exhibits pseudoportal tracts that lack veins and bile ducts and contain large, thick-walled arteries

surrounded by fibroconnective tissue with variable ductular reaction and inflammation. The last

subtype, HA-U (10%), is associated with gain-of-function mutations of the IL6ST gene. These HAs lack

distinctive morphologic and molecular features.

Patients with HAs appear to have a risk of either rupture or malignant transformation (estimated risk

of 4.2% to 4.5%57,58), which is different from the other benign lesions. The HA-B subtype has been

found to be more frequently associated with the development of HCC.59,60 For small lesions (<5 cm),

some experts have recommended withdrawal of oral contraceptives in the expectation of tumor

regression.61,62 However, regression of HA upon cessation of oral contraceptives does not remove the

risk of malignant transformation in women.63 High-risk groups for malignant transformation include

male patients, patients with a history of androgenic or anabolic steroid intake, and patients with

glycogen storage disease.

The potential for malignant transformation and the risk of hemorrhage drive active surgical

interventions for large HAs. Surgical resection is recommended for HAs >5 cm, those with intratumoral

hemorrhage, and those that increase in size.64–67 HAs are hypervascular tumors and are sometimes

difficult to distinguish from HCC. Typically, HAs are represented by bright enhancement in the arterial

phase and iso- to hypodensity in the portal phase (Fig. 60-8). With MRI, HAs are visualized as welldefined lesions with isointense or hyperintense features on T1- and T2-weighted images, depending on

the fat content. Similar to dynamic CT, gadolinium enhancement shows marked enhancement in the

arterial phase and isointensity in the portal phase.

PRIMARY HEPATIC MALIGNANCIES

Hepatocellular Carcinoma

Primary liver cancer is one of the most common solid tumors and the second-leading cause of cancerrelated deaths worldwide.68 Despite recent developments in preventing and treating viral hepatitis,

nearly 750,000 deaths were reported in 2012 for patients with primary liver cancer. Among these cases,

HCC is the most common type of hepatic malignancy. There is wide geographic variability in incidence,

with the majority of the cases occurring in developing countries compared with developed countries.

China alone accounts for more than half of the new cases of liver cancer; other high-incidence areas are

sub-Saharan Africa, Japan, and Southeast Asia, where viral hepatitis or aflatoxin exposure is more

endemic. In the United States, the incidence of HCC is increasing the second-fastest of all tumor types,

probably due primarily to the HCV epidemic.69 There is also global variation in the risk factors for HCC.

HBV and HCV infections are the leading causes of HCC, accounting for 75% of the cases worldwide.70

HBV infection is more common except in Japan, where HCV infection is the most common cause of

HCC. Alcoholic liver disease accounts for a significant proportion of HCC in the Western countries.

Aflatoxin exposure is also an important risk factor in China and sub-Saharan Africa.

1553

Figure 60-8. A huge hepatic adenoma found in a 31-year-old woman. Dynamic CT revealed early enhancement in the arterial

phase (left) and iso- to hypodensity in the portal phase (right). This patient had no underlying liver disease and both serum alphafetoprotein level and plasma des-gamma-carboxyprothrombin level were within normal limits.

Screening measures for HCC have been assessed in various studies, which primarily focus on the roles

of serum AFP level and ultrasound. The American Association for the Study of Liver Disease (AASLD)

has created guidelines regarding the use of these screening techniques based on the best existing

evidence.20 The AASLD currently recommends serial hepatic ultrasonography and serum AFP

measurement every 6 to 12 months for at-risk populations (e.g., any patient with cirrhosis). Similarly,

the Japanese Society of Hepatology has recommended hepatic ultrasonography and serum AFP or

plasma DCP measurement every 3 to 4 months and dynamic CT or MRI every 6 to 12 months for veryhigh-risk populations (e.g., those with HBV- or HCV-related cirrhosis), and hepatic ultrasonography and

assessment for tumor markers every 6 months for high-risk populations (e.g., HBV- or HCV-related

chronic hepatitis, or cirrhosis with other etiologies) with or without dynamic CT or MRI every 6 to 12

months, as appropriate.71 A randomized controlled trial from China reported that the use of ultrasound

and AFP enables early detection of HCC and lowers HCC-related mortality by 37% in patients with HBV

infection.72 Various nonrandomized studies have confirmed prognostic improvement through regular

surveillance programs among patients at high risk of HCC.73–75 Regarding the interval of screening, a

recent meta-analysis revealed that surveillance every 6 months is significantly better than screening

every 12 months for the early detection of HCC. Santi et al.76 reported that 70% of newly diagnosed

HCCs fell within the Milan criteria (solitary tumor ≤5 cm or ≤3 tumors with each tumor ≤3 cm)77

when screening occurred at least every 6 months, while the proportion of HCCs that fell within the

Milan criteria was 57% and the overall survival rate was inferior with surveillance every 6 to 12

months. When a suspected nodule is detected during regular screening, a contrast-enhanced three-phase

CT or MRI is needed. Because HCCs are fed mainly by arterial flow, early enhancement and washout of

contrast on the delayed phase of the scan are typical findings suggestive of HCC (Fig. 60-9). These

enhancement characteristics increase the specificity of the scan to >95%.78

4 The choice of therapy is individualized based on the tumor burden, degree of underlying liver

disease, patient performance status, and overall possibility of side effects or complications balanced

with acceptable results. The Barcelona Clinic Liver Cancer (BCLC) staging system79 is widely used to

stratify patients for specific therapies (Algorithm 60-1). However, the limitation of the BCLC algorithm

is that it is fairly conservative with regard to the application of surgical therapy. Patients with larger

solitary tumors are not considered surgical candidates despite growing experience with resection with

acceptable outcomes in this group.80,81 In the guidelines on liver cancer treatment proposed by the

Japan Society of Hepatology,22 surgical resection is indicated for CTP class A or class B patients with

HCCs with up to three nodules irrespective of the size of each tumor, while liver transplantation is

limited to CTP class C patients meeting the Milan criteria.

For HCC patients undergoing liver resection, strict assessment of the hepatic functional reserve is

important because HCC usually develops in an injured liver and, accordingly, the maximum extent of

resection must be estimated carefully. Functional investigations have been proposed to better evaluate

the severity of chronic liver disease. Measurement of indocyanine green (ICG) retention rate at 15

minutes (ICG-R15) is the most frequently used test. For patients with obstructive jaundice or congenital

intolerance to ICG, 99mTc-galactosyl human serum albumin (GSA) scintigraphy sensitively estimates the

hepatic functional reserve. Algorithm 60-2 shows a decision tree used for surgical planning according to

1554

the individual hepatic functional reserve (Makuuchi criteria).82 By strictly following this algorithm, no

mortality due to hepatic insufficiency was reported in the University of Tokyo series.1

Figure 60-9. Typical enhancement pattern of HCC with early enhancement in the arterial phase (left) and washout in the portal

phase (right).

Because HCC tends to spread via portal veins, anatomic resection of the tumor-bearing portal

territory is a theoretically reasonable approach (Fig. 60-10). Although the efficacy of anatomic resection

remains controversial, various studies have reported prognostic superiority of anatomic resection with

an apparent lower local recurrence rate and a higher recurrence-free survival rate compared to

nonanatomic limited resection of the liver.83–85 Liver transplantation is another surgical approach with a

theoretically higher chance of eradicating the tumor burden, especially for patients with severe hepatic

dysfunction. Clinical outcomes of liver transplantation for HCC were initially poor in the early era.

However, after the landmark study by Mazzaferro et al.,77 good survival outcomes are expected in

select HCC patients with tumors of limited size and a small number of nodules (Milan criteria). The

selection criterion for transplantation used by Mazzaferro et al.77 was a tumor diameter of ≤5 cm in

patients with a single HCC or ≤3 cm in patients with 2 or 3 lesions. This criterion has since been

extended with or without tumor markers or biopsy findings in several high-volume transplant centers

(Table 60-2).86–97

Algorithm 60-1. BCLC algorithm for treatment selection in patients with HCC. (Adapted from Llovet JM, Burroughs A, Bruix J.

Hepatocellular carcinoma. Lancet 2003;362:1907–1917 with permission.)

1555

Algorithm 60-2. Treatment algorithm of patients with hepatocellular carcinoma (HCC) based on serum bilirubin level and

indocyanine green retention rate at 15 minutes. (Adapted from Makuuchi Kosuge T, Takayama T, et al. Surgery for small liver

cancers. Semin Surg Oncol 1993;9:298–304 with permission.)

Intrahepatic Cholangiocarcinoma

Intrahepatic cholangiocarcinoma (ICC) is a relatively uncommon disease, accounting for 5% to 30% of

all primary liver malignancies.98 In the United States, the age-adjusted incidence of ICC increased from

0.32 in 1975 to 0.85 per 100,000 population in 2000 and has yet to plateau.99 The overall 5-year

survival rate for patients with unresectable tumor is dismal: 5% to 10%. Because systemic therapy for

ICC has not yet been established, surgical resection offers the only chance for cure. However, the

overall 5-year survival rate after curative-intent surgery is disappointing at 20% to 35%.100

Figure 60-10. Theoretical concept of anatomic resection for HCC. This schema shows an example of resection for HCC located in

the ventral part of Segment VIII. The dashed line A-A’ indicates nonanatomic limited resection with adequate surgical margin and

the dotted line B-B’ represents anatomic resection of the ventral part of Segment VIII. Because the “tumor-bearing” portal territory is

at high risk of harboring micrometastases scattered via portal veins (arrows), systematic removal of the corresponding portal region

would offer a better chance of eradicating the cancer cells. With systematic removal of the territory of the ventral branch of

Segment VIII, the tumor-bearing portal territory, which is at high risk of micrometastases, can be resected. T, tumor; P8v, ventral

branch of Segment VIII portal pedicle; P8d, dorsal branch of Segment VIII portal pedicle; MHV, middle hepatic vein; V8i,

intermediate vein for Segment VIII.

ICC is thought to derive from a common hepatic progenitor cell that may also give rise to HCC,101

and combined hepatocellular-cholangiocarcinoma having histopathologic features of both HCC and ICC

is sometimes observed. Risk factors for ICC have not yet been established, although chronic liver

disease and cirrhosis are reportedly correlated with ICC. On dynamic imaging studies, ICC is not

enhanced due to the hypovascular nature of this tumor type. However, various degrees of ring

enhancement are observed surrounding the lesion, reflecting the fibrous connective tissue frequently

observed around the tumor. In laboratory tests, elevated carcinoembryonic antigen or CA19-9 in

primary solid liver lesion is suggestive of ICC.

A recent meta-analysis showed that factors associated with shorter overall survival included older

patient age, larger tumor size, multiple tumors, lymph node metastasis, vascular invasion, and poor

tumor differentiation. Because up to 30% of patients have lymph node metastases, routine

lymphadenectomy is recommended,102 although the survival benefit remains controversial.

METASTATIC LIVER TUMORS

1556