Benign liver neoplasms encompass a variety of liver lesions, each with distinct pathologic, radiographic, and molecular characteristics. These include hemangioma, focal nodular hyperplasia, hepatocellular adenoma, and other less commonly seen lesions (Table 57-1). Benign liver lesions occur in up to 20% of the population and far surpass the incidence of malignant liver lesions. With the increased utilization of cross-sectional imaging, these tumors are being identified more frequently. Benign liver lesions are usually asymptomatic and are generally observed. Surgical intervention is warranted in symptomatic patients, cases where malignancy cannot be excluded, or if there is a potential for malignant transformation or associated complications (Fig. 57-1). Liver lesions with equivocal imaging characteristics can lead to diagnostic uncertainty resulting in important therapeutic ramifications. As a result, a thorough understanding of benign liver neoplasms is necessary to more accurately and appropriately screen patients for expectant management versus surgical intervention.
Hepatic hemangiomas are the most common benign tumor of the liver, affecting up to 20% of the population.1,2 Hemangiomas occur predominantly in females (60%-80%) and typically present in the third to fifth decades of life.3-5 Hemangiomas may be isolated to the liver or associated with systemic syndromes.6 In nearly 50% of patients, hepatic hemangiomas are multifocal.7
The etiology of hepatic hemangiomas is poorly defined. They are thought to arise either as congenital lesions that enlarge due to vascular ectasia or as vascular enlargement from previously normal hepatic vasculature. The blood supply of hepatic hemangiomas is derived from the hepatic artery. Commonly seen characteristics suggest a role of estrogens in their pathogenesis. These include female predominance, increase in size during pregnancy, and change in size while taking oral contraceptive pills (OCPs), as well as association of estrogen replacement therapy with hemangioma recurrence.8-10 The association of hormones and hemangioma does, however, remains controversial. For example, a case-control study showed no association between liver hemangioma and a history of OCP use.11 However, Glinkova et al. performed a prospective evaluation of 94 women with 181 hemangiomas and concluded that hormone therapy increased the risk of hemangioma enlargement.7 Although the association of hormone therapy and hepatic hemangioma pathogenesis remains poorly understood, hepatic hemangioma patients with nonphysiologic exposure to sex hormones warrant serial close observation.
Grossly, cavernous hemangiomas are soft, compressible, blood filled, and well-defined tumors. Hemangiomas can be multifocal and of variable size. Hemangiomas greater than 5 cm in size are typically considered “giant” hemangiomas. Microscopically, hemangiomas consist of a single layer of benign endothelium along vascular channels separated by thin connective tissue. Cystic degeneration, thrombosis, fibrosis, and calcifications may be present. Estrogen and progesterone receptors are not typically seen.12
Hepatic hemangiomas are generally not encapsulated and the liver parenchyma–hemangioma interface varies (Fig. 57-2). Zimmerman and Baer described this variation and its clinical implications.13 Most commonly, the liver parenchyma–hemangioma interface consists of capsule-like, avascular fibrous lamellae, making hemangiomas more amenable to enucleation. The second pattern, known as the interdigitating interface, has components of the hemangioma projecting into the associated liver parenchyma without fibrous lamellae. A third pattern is characteristic of highly irregular borders without clear delineation of liver parenchyma versus the hemangioma. The fourth and final interface lacks fibrous tissue between the hemangioma and liver parenchyma. In this case, the hemangioma is in direct contact with the liver parenchyma leading to a smooth, regular interface that includes dilated portal vein branches and microhemangiomas. The lack of a fibrous capsule in the latter three patterns complicates attempts at enucleation.
Hepatic hemangiomas are typically asymptomatic and found incidentally during imaging studies obtained for other reasons. Symptomatic hemangiomas occur in 10% to 50% of cases. Not unexpectedly, larger lesions are more likely to produce symptoms.3,14 Associated symptoms include right upper quadrant pain, general abdominal pain, nausea, and early satiety as a result of compression of adjacent structures. Potential life-threatening complications may include hemorrhage, hemobilia, rupture, and hemangioma-associated heart failure; however, each of these complications is very uncommon. In fact, less than 5% of hemangiomas will present after spontaneous rupture. However, if a patient does present with rupture, disseminated intravascular coagulation, hemodynamic instability, and hypovolemic shock are seen in up to one-third of these cases.4,15 Spontaneous rupture of a liver hemangioma has an estimated mortality of approximately 35%.16 Kasabach–Merritt syndrome is a life-threatening complication often triggered by a dental or surgical procedure in patients with giant hemangiomas. It consists of thrombocytopenia and disseminated intravascular coagulopathy, and patients present with acute right upper quadrant pain and bleeding.
Diagnosis of a hepatic hemangioma is typically accomplished via imaging (Table 57-2). On ultrasound, hemangiomas appear as a well-defined, hyperechoic mass (Fig. 57-3A). The echogenicity can vary secondary to internal fibrosis, thrombosis, and necrosis.2 Unenhanced abdominal computed tomography (CT) depicts a hypodense or isodense lesion within the liver parenchyma, and an unenhanced scan can often miss hemangioma lesions. Multiphasic CT and magnetic resonance imaging (MRI) are highly sensitive and are most useful in differentiating hemangiomas from other hepatic lesions. Early peripheral enhancement that proceeds toward central enhancement is generally diagnostic. MRI has a better sensitivity (91%) and specificity (92%) than CT2 (Fig. 57-3B-F). MRIs with heavily weighted T2 imaging typically show delayed relaxation times and can reliably differentiate hemangioma from metastatic disease with an accuracy of 97%.17 Positron emission tomography scans may be helpful in differentiating hemangiomas from metastatic disease given that hemangiomas are not 18F-fluorodeoxyglucose (FDG)-avid—although this is rarely necessary. Technetium-99m-labeled red blood cell scans can also be used in instances where CT and MRI are nondiagnostic, although this test is also seldom needed.
|Hepatic hemangioma||Well-defined, hyperechoic||Hypoattenuating on noncontrast CT||T1: hypointense, T2: hyperintense with respect to the liver||Tagged red blood cell scan with larger lesions can show increased activity on delayed images|
|Arterial phase shows peripheral, discontinuous enhancement||T1+ Gadolinium: peripheral, discontinuous nodular enhancement with centripetal filling on delayed images|
|Portal venous phase shows centripetal filling and hyperattenuating lesion||T1 + Eovist: variable appearance|
|FNH||Variable echogenicity and difficult to visualize||Hypo or isoattenuating lesion with hypoattenuating central scar on noncontrast CT|
T1: isointense with hypointense central scar
T2: isointense with hyperintense central scar
|Technetium 99 m sulfur colloid scan positive (80% of cases)|
|Arterial phase shows hyperattenuating lesion with hypoattenuated central scar||T1 + Gadolinium: intense arterial phase, isointense to liver on portal venous phase|
|Enhanced on delayed phases in 80% of cases||T1 + Eovist: early arterial enhancement, persists in delayed phases, remaining slightly enhanced on hepatobiliary phase|
|Hepatic adenoma||Well-defined, typically solitary lesion with variable echogenicity depending on fat and blood content||Well demarcated with variable attenuation depending on blood (hyperattenuating) and fat (hypoattenuating) content||T2: hyperintense|
|T1+ Gadolinium: early arterial enhancement followed by rapid isointensity to liver on delayed images|
|T1 + Eovist: hypointense on hepatobiliary phase|
Hemangioma in a 42-year-old female. (A) Ultrasound image in the sagittal plane showing a large echogenic lesion (arrows), which is a feature of hemangioma. (B) Axial T2-weighted image shows the lesion to be very bright (arrow), known as the “light bulb” sign, and compatible with hemangioma. Axial T1-weighted images in the hepatic arterial phase (C), portal venous phase (D), and delayed phase (E) show peripheral nodular enhancement increasing centrally (arrowheads), characteristic of hemangioma. (F) Coronal T1-weighted image better demonstrates enhancing nodules (arrows) at the periphery of the lesion. (Used with permission from Ihab Kamel MD and Sepideh Besharati MD.)
Invasive diagnostic procedures are unnecessary and are typically contraindicated given the high likelihood of complications. Liver function tests are generally normal except in cases of obstructive jaundice secondary to parenchymal compression or in cases of Kasabach-Merritt syndrome. Biopsy of hepatic hemangioma is highly unreliable and is associated with a high risk of bleeding, and therefore is contraindicated when hepatic hemangioma is suspected.3
Hepatic hemangiomas typically remain stable in size, and radiographic changes are unlikely.18 As a result, observation is warranted in both asymptomatic and minimally symptomatic patients.3 Gestational hepatic hemangiomas behave similarly to lesions in nonpregnant patients. Accordingly, asymptomatic lesions may be followed in both pregnant patients and patients taking hormone replacement therapy.
Indications for surgical resection include severely symptomatic lesions, patients with hemangioma-related complications (ie, rupture or Kasabach–Merritt syndrome), and patients with indeterminate liver lesions where malignancy cannot be ruled out. Of note, size is not a criterion for operative intervention. In a retrospective series of 492 patients with giant hemangiomas, long-term outcomes were available in 289 survey responders. Clinical observation resulted in a similar incidence of hemangioma-related complications/symptoms versus operatively managed patients19 (Table 57-3).
|Nonoperative Group (n = 233)||Operative Group (n = 56)|
|Type of Adverse Eventa||n||%||n||%||p Value|
|Persistence of symptoms or complications||26||11||0||0||<0.001|
|New onset of symptoms or complications||20||9||0||0||0.02|
|Need for intervention or reoperation||14||6||0||0||0.08|
|Total patients with adverse events||46||20||8||14||0.45|
Operative approaches for hepatic hemangiomas include enucleation and formal anatomic or nonanatomic liver resection. Hemangioma enucleation is associated with reduced loss of functional hepatic parenchyma, less perioperative blood loss, and fewer complications compared with formal liver resection.3,20 As a result, enucleation is the most commonly employed surgical technique. Peripheral lesions are particularly amenable to this approach. Enucleation typically begins with control of hepatic artery inflow via a Pringle maneuver. The ipsilateral hepatic artery is identified. For larger lesions, ligation of the ipsilateral hepatic artery may be necessary, whereas in smaller hemangiomas, hepatic artery inflow can be controlled with ligation of more distal branches. Division of hepatic parenchyma is then performed via gross identification of the hemangioma to normal parenchyma border. Control of bile ducts and small blood vessels is obtained while transecting through the compressed sheath of liver tissue representing the hemangioma–normal liver parenchyma interface.
Follow-up imaging as suggested by Mezhir et al. is recommended.21 This consists of follow-up MRI 3 months after diagnosis along with a repeat MRI in 3 to 6 months for atypical lesions with low suspicion for malignancy. More infrequent follow-up is appropriate once the diagnosis and stability of the lesion have been documented. Recurrence is rare after definitive treatment. These instances are commonly associated with estrogen therapy.9 As noted above, hemangiomas in pregnant patients or patients receiving exogenous estrogen may have a higher likelihood to increase in size.7 As a result, these patients warrant careful observation.7,22
FOCAL NODULAR HYPERPLASIA
Focal nodular hyperplasia (FNH) is the second most common benign solid liver tumor. Its prevalence is estimated to be between 0.3% to 3% based on autopsy analysis.23 FNH occurs predominantly in females (8:1 female-to-male ratio). The average age of presentation is 35 years old. FNH tumors are composed of a cluster of benign-appearing hepatocytes within the background of normal-appearing hepatic parenchyma.24
FNH is thought to arise from a vascular malformation or a vascular injury and subsequent expansion of proliferative hepatocytes.25 Its association with vascular abnormalities is further supported by studies demonstrating a higher prevalence of FNH in families with hereditary hemorrhage telangiectasia.26 Furthermore, in a series of 148 patients with FNH, 20% of patients had concomitant hepatic hemangiomas.27 An association with dysregulation of angiopoietin genes ANGPT1 and ANGPT2 has also been suggested.28
FNH and its association with oral contraceptives remains controversial. Approximately 50% to 75% of patients with FNH report a history of oral contraceptive use. Moreover, estrogen receptor expression within FNH specimens has been reported.29,30 However, other studies indicate that there is no direct association between FNH and oral contraceptives or sex hormones.31-33 For example, in their analysis of 216 patients, Mathieu and colleagues concluded that neither the size nor the number of FNH lesions was influenced by OCP use.31 Size changes during follow-up were rare and were not associated with OCP use.
FNH lesions are benign lesions without malignant potential. They grossly appear as brown or tan nodules composed of proliferative, polyclonal hepatocytes and lack the presence of a true capsule24 (Fig. 57-4A). Approximately 80% of patients present with solitary lesions that are variable in size (1 mm-20 cm).25,34 A central scar (80%) and bile duct proliferation (100%) are characteristics of FNH lesions34 (Fig. 57-4B,C).
(A) Surgical specimen of focal nodular hyperplasia showing a globular, lobulated mass with a prominent central scar. Microscopic appearance of focal nodular hyperplasia at high magnification showing (B) hepatocytes with a central fibrous septum and proliferating bile ductules at the interface between the hepatocytes and fibrous tissue. (Used with permission from Dr. Robert Anders, MD, PhD.)