Hepatocellular adenomas (HCAs) are monoclonal, hepatocellular

Transcription

1 REVIEW ARTICLE Hepatocellular Adenomas: Current Update on Genetics, Taxonomy, and Management Alampady Shanbhogue, MD,* Shetal N. Shah, MD,Þ Atif Zaheer, MD,þ Srinivasa R. Prasad, MD,* Naoki Takahashi, MD, and Raghunandan Vikram, MBBS Abstract: Hepatocellular adenomas (HCAs) are uncommon, benign hepatocellular neoplasms that commonly occur in young women. Recent advances in pathology and cytogenetics have thrown fresh light on the pathogenesis of HCAs leading to classification of HCAs into 3 distinct subtypes, each with a characteristic epidemiology, histopathology, oncogenesis, and imaging findings. The aim of the article was to provide a comprehensive review of contemporary taxonomy of HCAs, with an emphasis on cross-sectional imaging findings and management. Key Words: genetics, pathology, hepatocellular adenoma, computed tomography, magnetic resonance imaging (J Comput Assist Tomogr 2011;35: 159Y166) Hepatocellular adenomas (HCAs) are monoclonal, hepatocellular neoplasms that are typically described in young women on oral contraceptives (OCs). 1,2 Because HCAs have an increased proclivity to hemorrhage and to rarely undergo malignant transformation, they warrant either a periodic imaging surveillance or surgical resection. Until recently, the pathogenesis of HCAs was poorly understood. Recent developments in pathology and genetics have contributed to better understanding of the distinctive oncological pathways involved in HCAs. It has now been shown that HCAs possess unique genetic signatures that are distinct from normal hepatocytes and hepatocellular carcinomas (HCCs). 3 Hepatocellular adenomas are now classified into 4 subtypes, including a miscellaneous category. The 3 distinct subtypes of HCAs include hepatocyte nuclear factor-1> (HNF-1>)Ymutated HCAs (HNF-HCAs), HCAs characterized by A-catenin mutations, and inflammatory HCAs (I-HCAs; due to mutations involving interleukin-6 [IL-6] signal transducer). Inflammatory HCAs and HNF-HCAs are the 2 major subgroups of HCAs, which together constitute for up to 80% of all adenomas. The characteristic genetic abnormalities, epidemiologic features, histomorphology, clinicobiologic behavior, and imaging findings of these 3 subtypes of HCAs are summarized in Table 1 and discussed in detail in the article. 4Y7 EPIDEMIOLOGY, PATHOLOGY, AND CLINICAL MANIFESTATIONS Hepatocellular adenoma is the second most common benign hepatocellular neoplasm after focal nodular hyperplasia. The exact incidence and prevalence of HCA is difficult to predict because of the introduction of low-dose OCs and increased rate From the *Departments of Radiology, University of Texas Health Science Center at San Antonio, TX; Cleveland Clinic, OH; Johns Hopkins University, Baltimore, MD; Mayo Clinic, Rochester, MN; and University of Texas MD Anderson, Houston, TX. Received for publication October 22, 2010; accepted December 15, Reprints: Srinivasa R. Prasad, MD, Radiology Department, University of Texas HSC at San Antonio, 7703 Floyd Curl Dr, San Antonio, TX ( prasads@uthscsa.edu). Copyright * 2011 by Lippincott Williams & Wilkins of incidental detection at least in part due to the widespread use of multiphase computed tomography (CT) or magnetic resonance imaging (MRI). Sex hormones, particularly estrogens, seem to play a major role in the pathogenesis of HCAs. Hepatocellular adenomas occur predominantly in young adult women (with a male-to- female ratio of 1:8Y10), with a mean age of 41 years, and are rarely seen in adult men and children. 4,8 Hepatocellular adenomas have been reported to grow remarkably during pregnancy, sometimes with catastrophic consequences from rupture or hemorrhage. 9 Medical literature has shown a strong association between long-term OC use and the development of HCA. Although early data suggested a definitive correlation between the duration of OC pill use and risk of HCAs (relative risk ranging from 1.3 for 1Y3 years of OC use to 25 for 911 years of OC use), 10 more recent data on low-estrogen OC pills have demonstrated an overall relative risk of 1.25 to ,12 Overall, 85% to 95% of patients with HCAs have history of OC pill use of more than 2 years. 6,13 Hepatocellular adenomas may show partial or complete regression after cessation of OC pills. 4,14 There is an increased risk of development of HCAs in patients with anabolic androgen steroid intake, familial adenomatosis polyposis and metabolic liver diseases such as glycogen storage diseases (types Ia, III, VI), tyrosinemia, galactosemia, steatohepatitis, and hemochromatosis. 6,15Y17 Pathologically, most HCAs are solitary, unencapsulated tumors that vary from less than 1 cm to up to 20 cm in size. Multiple HCAs may occur in adenomatosis, a condition defined by the presence of more than 10 adenomas in a normal liver (discussed later). Histologically, HCAs are characterized by layers of mildly thickened or irregular liver cell cords and plates separated by sinusoids. In contradistinction to HCCs, the hepatocytes in HCAs lack cytologic and nuclear atypia. Typically, HCAs do not have portal tract elements including the portal venules or bile ductules. Clinically, up to 50% of patients with HCAs may present with a wide variety of symptoms of which right upper quadrant discomfort is the most common (up to 43%). 18 Palpable mass and severe abdominal pain (secondary to rupture or hemorrhage) are uncommonly encountered. 6 Nearly 60% of patients who are symptomatic have signs of intratumoral or peritumoral hemorrhage. 19Y21 Overall, up to 30% of HCAs may bleed and less than 10% undergo malignant transformation to HCCs. 22 Detailed descriptions of specific pathogenesis, pathology, clinical, and imaging manifestations of individual HCA subtypes are presented below. INFLAMMATORY HCAs Inflammatory HCAs, which form the largest subgroup of HCAs, comprise 40% to 55% of all HCAs. These tumors are predominantly seen in women, in association with obesity, alcohol use, and hepatic steatosis. More than 90% of women with I-HCAs have a history of OC use. 15 Inflammatory HCAs are also described in men in whom these tumors are almost always solitary. 4 Clinically, patients with I-HCAs may manifest a J Comput Assist Tomogr & Volume 35, Number 2, March/April

2 Shanbhogue et al J Comput Assist Tomogr & Volume 35, Number 2, March/April 2011 TABLE 1. Salient Epidemiological, Clinical, Pathological, and Imaging Features of 3 Major Subtypes of HCA Epidemiology Clinical features Pathological features Imaging features MDCT/MRI CEUS Inflammatory HCA (I-HCA) 40%Y55% of HCAs; majority are women; rarely seen in men. History of OC use in 990% cases. Increased risk of bleeding and small risk of malignant transformation. Associated with obesity, alcohol abuse, and inflammatory syndrome. Marked sinusoidal dilatation/peliosis, polymorphous inflammatory infiltrates, thick tortuous arteries. Less extensive and variable steatosis. Prominent ductular reaction. Previously misclassified as telangiectatic FNHs. No cytologic and nuclear atypia. Strong expression of serum amyloidyassociated protein A2 (SAA-2) and CRP on immunohistochemistry. Hyperintense on T2-weighted image; intense arterial enhancement that persists on venous and delayed phases. Arterial hypervascularity with centripetal filling; peripheral rim of sustained enhancement with delayed central washout. HCA With HNF-1> Gene Mutation 35%Y50% of HCAs; almost exclusively in women. History of OC use in 990% cases. May bleed; no risk of malignant transformation. Tend to develop familial adenomatosis and MODY-3 diabetes mellitus. No inflammatory infiltrates or significant peliosis. Excessive lipid accumulation within the tumor hepatocytes. No portal tract elements. No cytologic and nuclear atypia. Lack of expression of liver fatty acid binding protein on immunohistochemistry. Diffuse steatosis within the lesion; arterial enhancement that does not persist into the venous phase. Homogeneously hyperechoic on grayscale imaging with mild to moderate arterial hypervascularity; becomes isoechoic on the portal venous phase. *Malignant HCAs may show arterial enhancement with venous washout, mimicking HCC. HCA With A-Catenin Activation 10%Y18% of HCAs. Affects men and women. Definite increased risk of malignancy. Associated with androgen therapy and glycogen storage disease. No significant peliosis or steatosis. No portal tract elements. Cytological and nuclear atypia may be seen (in malignant HCA). Strong diffuse overexpression of glutamine synthetase and nuclear A-catenin staining on immunohistochemistry. Arterial enhancement with portal venous washout. V* systemic inflammatory syndrome including fever, leukocytosis, and elevated serum C-reactive protein (CRP) levels. 23 Inflammatory HCAs carry a definite increased risk of bleeding (up to 30%) and a small risk of malignant transformation (5%Y9%). 6 Inflammatory HCAs comprise a prototype example of tumors induced by hepatobiliary inflammation. More than 40% of around 285 genes overexpressed in I-HCAs are associated with inflammation and immune response. 24 Interleukin-6 is a proinflammatory cytokine that regulates acute phase inflammatory response and liver regeneration through activation of the JAK-STAT (Janus kinase- signal transducers and activators of transcription) pathway. Sustained activation of IL-6 receptor signaling due to somatic mutations of IL-6 signal transducer gene (IL-6ST gene), which encodes glycoprotein-130 (gp-130), is one of the major pathogenetic mechanisms proposed in the development of I-HCAs. 24 The resultant overactivation of acute phase inflammatory response results in overexpression of markers such as serum amyloid A and CRP in neoplastic hepatocytes and, in some cases, in serum. Serum levels of F-glutamyl transferase may also be elevated in these patients. 25 Somatic, activating mutations of IL-6ST gene are seen in up to 60% of I-HCAs. The remainder of the I-HCAs show STAT3 activation independent of gp-130 activation. 26 Interestingly enough, activation of IL-6YSTAT3 pathway is also described in development of HCCs especially when combined with inactivation of the transforming growth factor-a signaling pathway. 27 Around 10% of I-HCAs may also show mutations involving A-catenin gene. Histologically, I-HCAs are characterized by marked sinusoidal dilatation, polymorphous inflammatory infiltrates, peliosis, and thickened tortuous arteries. The presence of prominent ductular reaction is a distinct histologic feature. Steatosis within I-HCA is less extensive (compared with HNF-HCA) and variable. 6 Historically, HCAs with these histologic features were misclassified as telangiectatic focal nodular hyperplasias (FNHs). Polymorphous inflammatory infiltrate seen in I-HCAs likely results from a significant increase in CCL20, a chemokine that attracts a wide spectrum of immune cells. 28 On imaging, I-HCAs manifest as hypervascular hepatic masses with persistent enhancement in the portal venous and delayed phases. Inflammatory HCAs are markedly hyperintense on T2-weighted images corresponding to areas of sinusoidal dilatation (Fig. 1). Marked T2 hyperintensity and persistent delayed enhancement have been shown to have a sensitivity of up to 85.2% and specificity of up to 87.5% for detection and characterization of I-HCAs. 29 Focal areas of microscopic fat may be seen in a small subset of patients (11%). 29 On contrastenhanced ultrasound (CEUS), I-HCAs show arterial hypervascularity with centripetal filling and peripheral rim of sustained enhancement with delayed central washout. These findings have a diagnostic significance given that other benign hepatic lesions rarely exhibit such a pattern. 30 HNF-1>YMUTATED HCAs Hepatocyte nuclear factor-1>ymutated HCAs (HNF-HCAs) account for 35% to 50% of all HCAs and arise because of biallelic inactivation of transcription factor 1 gene located in chromosome 12. Hepatocyte nuclear factor-1>ymutated HCAs * 2011 Lippincott Williams & Wilkins

3 J Comput Assist Tomogr & Volume 35, Number 2, March/April 2011 Hepatocellular Adenomas: Current Update FIGURE 1. Multiple inflammatory HCAs in a 35-year-old woman. A, Axial T2-weighted image reveals multiple focal hepatic lesions (white arrows) that are hyperintense on T2-weighted images. After administration of intravenous gadolinium, the lesions show intense arterial enhancement (white arrows in B), which persists into the venous and delayed phases (white arrows C). Histopathologic examination from percutaneous core biopsy revealed marked peliosis, thick-walled vessels, and bile ductular proliferation consistent with telangiectatic HCA. are exclusively seen in women; more than 90% patients have history of OC use. 15 Hepatocyte nuclear factor-1>ymutated HCAsmaybemultipleinupto50%ofcases. 6 Hepatocyte nuclear factor-1> mutations may be somatic (up to 90% of cases) or germline (5%Y10% of cases) in origin. Germline mutations of HNF-1> gene result in maturity-onset diabetes of the young- type 3 (MODY-3), a type of diabetes mellitus syndrome that is also characterized by familial adenomatosis. 7,31,32 Transcription factor 1 encodes the tumor suppressor gene, hepatocyte nuclear factor-1> (HNF-1>). Hepatocyte nuclear factor-1> mutation promotes lipogenesis by suppression of gluconeogenesis, activation of glycolysis, and promotion of fatty acid biosynthesis. 33 The down-regulation of fatty acid binding protein-1 leads to faulty transport of fatty acids and to intracellular deposition of fat. Therefore, HNF-HCAs are characterized by diffuse intralesional steatosis. The putative role of OCs in the induction and progression of HCAs is partly explained by genotoxic effects of estrogen, which result in HNF- 1> mutations. Alternatively, HNF-1> mutation may be the primary inciting event that results in the accumulation of estrogen metabolites that unconditionally stimulate hepatocyte proliferation. 34,35 On multidetector computed tomography (MDCT), HNF- HCAs may appear as liver masses with either macroscopic fat or low-density masses on unenhanced scans because of microscopic fat. Steatosis within HCAs is better demonstrated on MRI than on CT scans because of superior tissue differentiation on MRI. In contrast to a high (35%Y77% of adenomas) albeit variable detection of fat within HCAs on MRIs, fat deposition was seen in only 7% of 44 adenomas studied by CT scanning. 36Y38 On MRI, HNF-HCAs manifest as heterogeneous lesions with areas of T1 high signal intensity that may correspond to the presence of fat, glycogen or, less commonly, hemorrhage. The characteristic finding of HNF-HCAs is that of diffuse intralesional steatosis that is classically demonstrated as diffuse signal dropout on out-ofphase T1-weighted gradient echo imaging (Fig. 2). The sensitivity, specificity, positive predictive value, and negative predictive value of diffuse signal drop on T1 out-of-phase chemical shift FIGURE 2. Diffusely steatotic adenomas in a 31-year-old woman. Axial T2-weighted (A) and T1-weighted in-phase (B) and out-of phase (C) images reveal multiple focal hepatic lesions (white arrows), which are isointense on T2-weighted and in-phase T1-weighted images with signal loss on out-of-phase images. The lesions are hardly discernible on T2-weighted and in-phase T1-weighted images. After administration of intravenous gadolinium, the lesions show intense arterial enhancement (white arrows in D), which does not persist into the venous and delayed phases (white arrows in E). Histopathologic examination from the wedge biopsy from the left hepatic lobe confirmed the diagnosis of hepatic adenoma. Note that some of the lesions are best appreciated on postgadolinium images (arrowhead in D and E). * 2011 Lippincott Williams & Wilkins 161

4 Shanbhogue et al J Comput Assist Tomogr & Volume 35, Number 2, March/April 2011 imaging for predicting the HNF-HCA is 86.7%, 100%, 100%, and 94.7%, respectively. 29 On T2-weighted MR images, HNF- HCAs appear as homogenous masses that may be hypointense or hyperintense to the surrounding liver. 29 On contrast-enhanced CT or MRI, moderate contrast enhancement is seen in the arterial phase, which does not persist on to the portal venous and delayed phases. On CEUS, HNF-HCAs are homogeneously hyperechoic on grayscale imaging (likely because of diffuse steatosis) with mild to moderate arterial hypervascularity and isoechoic appearance to the liver on the portal venous phase. 30 A-CATENIN MUTATED HCAs A small subset of HCAs (up to 10%Y18%) originates from sustained activation of A-catenin because of mutations involving the CTNNB1 gene (catenin, beta-1). These tumors primarily affect patients with glycogen storage disease and on androgen treatment and have a greater propensity than the other subtypes of HCAs to undergo malignant transformation to HCCs. 4,7 A- Catenin plays a major role in hepatocyte development, differentiation, zonation, proliferation, and regeneration. Activation of A-catenin in normal hepatocytes is usually transient and is regulated by its rapid degradation. Adenomatous polyposis coli (APC) and Axin family genes in concert with glycogen synthase kinase-1a forms the cytoplasmic destruction complex responsible for prompt degradation of A-catenin. Excessive nuclear accumulation and sustained activation of A-catenin may result from mutations in A-catenin itself (with the mutant A-catenin resisting the degradation) or from mutations involving the cytoplasmic degradation complex. Biallelic mutations of the APC gene seen in familial polyposis syndrome therefore predisposes an individual for the development of HCAs. 39 Excessive A- catenin activity results in autonomous growth of hepatocytes and accelerates HCA formation. Hepatocellular adenomas with A-catenin activation show cytologic abnormalities and pseudoglandular formation on histology. 40 Aberrant activation of A- catenin is seen in HCCs and hepatoblastomas as well. Although A-catenin mutations are implicated in malignant transformation of HCAs, only 20% to 30% of malignant HCAs show A-catenin mutations suggesting alternate pathways of malignant transformation of HCAs. 18,41,42 Glycogen storage disease is a well-described independent risk factor for the development of HCAs and HCCs on account of background chronic inflammation. Up to 75% of patients with glycogen storage disease (IA, II, IV, and VI) may develop HCAs. 42 Several chromosomal and genetic alterations have been described in HCAs in glycogen storage disease, which include gain of chromosome 6p involving multiple oncogenes and loss of chromosome 6q, involving multiple tumor suppressor genes like IGF2R (insulin-like growth factor 2 receptor) and LATS1 (large tumor suppressor, homolog 1). 43 These chromosomal/ genetic changes are also frequently described in HCCs and preneoplastic dysplastic nodules, thereby explaining the higher risk of malignant transformation associated with HCAs in patients with glycogen storage disease. 44,45 Given the rarity of glycogen storage disease, the exact incidence of malignant transformation of HCA is, however, difficult to predict in this context. Hepatocellular adenomas with A-catenin gene mutations are less frequently encountered in glycogen storage disease (4/14 cases reported so far in the literature). 4,7,43,46 Patients with glycogen storage disease manifest with diffusely increased attenuation of the liver on MDCT (Fig. 3). Hepatocellular adenomas due to A-catenin mutations may appear as homogeneous or heterogeneous hypervascular masses and lack intratumoral fat (Fig. 4). Intense arterial enhancement is seen, which may or may not persist into the delayed phase. 29 Frankly malignant, nonyinflammatory type of HCAs with A- catenin mutations (up to 10%Y15% of all HCAs) simulates HCCs on imaging. MISCELLANEOUS HCAs (UNCLASSIFIED) The miscellaneous group of HCAs include the remainder of HCAs that have none of the distinct genetic alternations described above. These HCAs are poorly understood; detailed studies are required to evaluate specific pathogenetic and clinical characteristics of this disparate entity. HEPATIC ADENOMATOSIS Hepatic adenomatosis (HA) is defined as the presence of more than 10 adenomas in an otherwise normal liver. First described by Flejou et al 47 in 1985, patients with glycogen storage disease are typically excluded from this definition. Hepatic adenomatosis may be characterized by the development of either a steatotic or an inflammatory type of HCAs and shows a definite female preponderance (Figs. 5 and 6). 48 The role of OCs in the etiopathogenesis of HA is, however, still unclear. Hepatic adenomatosis was initially thought to be an independent risk factor FIGURE 3. Hepatocellular adenoma with malignant change in a 39-year-old man with glycogen storage disease. A, Axial unenhanced CT image of the liver showing a large heterogeneous lesion in the right lobe of the liver (white arrow). Note the mild diffuse hyperattenuation of the background liver. Multiple additional steatotic as well as nonsteatotic focal lesions were seen (not shown). Percutaneous ultrasound-guided biopsy of the largest lesion revealed HCC in the background of glycogen storage disease. B and C, Images from digital subtraction angiograms of the right hepatic artery injection before (B) and after (C) chemoembolization show intense vascular tumor blush in the right hepatic lobe on the pre-embolization image (arrows in B), which disappears after successful embolization (C) * 2011 Lippincott Williams & Wilkins

5 J Comput Assist Tomogr & Volume 35, Number 2, March/April 2011 Hepatocellular Adenomas: Current Update FIGURE 4. Hepatocellular adenoma with malignant change in a 42-year-old man. Axial contrast-enhanced CT images in the arterial (A) and delayed (B) phase after administration of intravenous contrast showing a large arterially enhancing mass in the liver, which become isodense to slightly hypodense to the liver on the delayed phase image. Surgical resection was performed because of the large size of the lesion. Surgical pathology confirmed the diagnosis of HCA with focus on carcinomatous change. for complications like hemorrhage or malignant transformation with a reported risk of bleeding of up to 63%. 49 However, a more recent analysis has shown that the rates of intrahepatic bleeding and intratumoral bleeding in asymptomatic incidentally discovered HA are 2% and 13%, respectively; rates that are similar to solitary HCAs. 50 Recent studies thus suggest that the potential risk of complications is related to the underlying histologic subtype and tumor size rather than the number of HCAs. HCAs: DIFFERENTIAL DIAGNOSES AND SUBTYPE DIFFERENTIATION The major imaging differential diagnoses of HCAs include FNH and HCC. Focal nodular hyperplasia, a polyclonal tumorlike lesion that occur predominantly in young women, commonly manifests as a lobulated, homogenous, solid liver mass that shows intense arterial phase enhancement and becomes isoattenuating to the liver on portal venous phase imaging. 51 The presence of a central scar is a useful sign, which can be readily detected in FNH on MRI in up to 80% of large FNHs (93 cm). 51 In contrast to FNHs, HCAs manifest as heterogeneous masses with variable presence of fat, hemorrhage, or necrosis. 51 Magnetic resonance imaging is the current criterion standard imaging technique in noninvasively differentiating FNH from HCA, especially with the advent of hepatobiliary-specific contrast agents like gadoxetic acid (EOVIST; Bayer Schering Pharma, Germany). 22 Unlike HCAs that are hypointense in the hepatobiliary phase (20 minutes after injection), most FNHs remain isointense or become hyperintense on delayed phase imaging after administration of gadoxetic acid (Fig. 7). 22 Gadobenate dimeglumine enhanced MRI has also been found to be highly accurate with sensitivity of up to 96.9% and specificity of up to 100% for differentiation of HCA and FNH. 52 In contradistinction to HCAs, HCCs develop against the background of liver cirrhosis or chronic hepatitis. Hepatocellular carcinomas may have variable imaging manifestations; they commonly appear as heterogeneously enhancing solid masses that show arterial phase enhancement and washout during portal venous phase. 53 Hepatocellular carcinomas may have variable proportion of fat, hemorrhage, and necrosis and may at times be indistinguishable from HCAs. The presence of infiltrative growth pattern, portal or hepatic venous thromboses, lymph node, and/or distant metastases helps discriminate HCCs from HCAs. Accurate differentiation may warrant biopsy and histopathologic evaluation. Traditionally, HCAs were considered as lesions with widely pleomorphic imaging appearance. Hepatocellular adenomas were usually described as heterogeneous, hypervascular masses with areas of fat, hemorrhage, and necrosis on multiphase CT or MRI. 54 Recent findings suggest that 2 major subtypes of HCAs, namely I-HCA and HNF-HCAs, can be reliably differentiated FIGURE 5. Hepatic adenomatosis in a 35-year-old woman. Axial unenhanced CT scan of the liver showing multiple fat-containing masses scattered throughout the liver consistent with multiple steatotic HCAs. FIGURE 6. Hepatic adenomatosis in a 38-year-old woman. Axial T2-weighted MR image demonstrating multiple T2 hyperintense lesions throughout the liver (white arrows) consistent with multiple inflammatory/telangiectatic HCAs. Note marked obesity (diverging white arrows). * 2011 Lippincott Williams & Wilkins 163

6 Shanbhogue et al J Comput Assist Tomogr & Volume 35, Number 2, March/April 2011 FIGURE 7. Role of gadoxetic acidyenhanced MRI in the differentiation of focal nodular hyperplasia from HCA. A, Axial gadoxetic acidyenhanced MR image in the hepatobiliary phase (20 minutes after injection) in a 35-year-old woman revealing a focal hepatic lesion that is isointense to slightly hyperintense to the hepatic parenchyma (white arrow) consistent with focal nodular hyperplasia. A central scar is also seen. B, Axial gadoxetic acidyenhanced MR image in the hepatobiliary phase (20 minutes after injection) in a 32-year-old woman revealing multiple focal hepatic lesions throughout the liver, which are hypointense to the hepatic parenchyma (white arrows). Histopathology confirmed the diagnosis of HCA. from one another on MRI (described above under specific subtypes) thereby significantly affecting patient management. Immunohistochemistry is also a reliable tool to differentiate various subtypes of HCAs. Lack of staining for liver fatty acid binding protein is 100% sensitive and specific for HNF-1>Ymutated HCA. Strong diffuse overexpression of glutamine synthetase (which is encoded by the A-catenin gene) and nuclear A-catenin staining is highly specific (100% specificity) and sensitive (85% sensitivity) for the detection of A-cateninYactivated HCAs and hepatocellular expression of serum amyloid A2 correlates closely with inflammatory adenomas (91% sensitivity and specificity). 25 MANAGEMENT OF HCAs: CURRENT UPDATE Current taxonomical schemata and genotype-phenotype correlation have resulted in a paradigm shift in the management of HCAs. Management guidelines are now based on the histologic subtype and size rather than the number of HCAs. It has been shown that HCAs more than 5 cm tend to bleed, and tumors more than 8 cm tend to undergo malignant change. 6 Risk of malignancy within HCAs is also higher in patients with glycogen storage disease and in men (up to 50% of HCAs in men). 18,42 Hepatocyte nuclear factor-1>ymutated HCAs do not carry any significant risk of malignant transformation; the risk of bleeding in HNF-HCA was 9% versus 16% in I-HCAs according to a study. 6,18 Hemorrhage associated with HCAs may be intratumoral or may result in subcapsular hematoma and/or hemoperitoneum (Fig. 8). Hemodynamic instability is rarely associated with ruptured HCAs, and selective hepatic artery embolization is sufficient to restore hemodynamic stability and to reduce tumor size as well as operative risk during surgery. 55Y57 Embolization may also obviate the need for aggressive surgical resection, which is associated with an increased morbidity. 6,58 Hepatocellular adenomas less than 5 cm in size are rarely associated with risk of hemorrhage or malignant transformation and hence can safely be managed conservatively with serial imaging follow-up. 7,33 Stopping the offending drugs (OC pills, androgens, barbiturates) and dietary modification as in glycogen storage disease are the recommended initial steps that may help by halting HCA growth and reducing the tendency to bleed. Most small HCAs remain stable during surveillance, and a small number of them may disappear. 4 Currently, there are no clear-cut guidelines for the optimal interval and duration of imaging surveillance in HCAs. Yearly surveillance with multiphase CT or MRI seems to be the most appropriate strategy, which should probably be continued until menopause. 4,59 Currently, surgical resection is recommended for HCAs more than 5 cm, HCAs that do not regress after stopping the offending drugs, HCAs with malignant change or evidence of A-catenin activation as demonstrated on biopsy, and all HCAs in men. 60 Rates of tumor recurrence are found to be very low after surgical resection. 6,18 Alternative approaches for the treatment of HCAs include transarterial embolization and radiofrequency ablation. Radiofrequency ablation has been found to be safe and effective in achieving complete tumor ablation in a recently published series of 10 patients with HCAs. 61 A recent analysis showed that, for small HCAs that continue to grow despite stopping the offending drugs, radiofrequency ablation may be the most cost-effective option, in comparison to surgery, transarterial embolization, and watchful waiting. 62 Image-guided biopsy with immunohistochemical and cytogenetic evaluation has been proposed to enable identification of patients at risk for malignant transformation. Biopsy may be performed for small I-HCAs that continue to grow despite FIGURE 8. Ruptured HCA in a 42-year-old woman presenting with acute right upper quadrant pain. Axial contrast-enhanced image of the hepatic arterial phase reveals a large right hepatic mass with focal disruption along the right anterolateral aspect, intralesional and perihepatic hematoma (arrowheads), and hemoperitoneum (arrows). Note residual enhancement of the lesion along its medial aspect * 2011 Lippincott Williams & Wilkins

7 J Comput Assist Tomogr & Volume 35, Number 2, March/April 2011 Hepatocellular Adenomas: Current Update stopping the offending drugs. 29 Widespread application of the guidelines on percutaneous biopsy of HCAs is still controversial because of several factors including lack of sophisticated techniques in pathology and cytogenetics in most hospitals around the world. Other potential problems with biopsy include sampling errors, the risk of hemorrhage or tumor seeding along the needle-track, and the occasional difficulty in differentiating HCA from normal hepatocytes as well as well-differentiated HCCs on histopathologic examination of limited core biopsy samples. Management of HCAs in women of child-bearing age is more complicated given the increased risk of growth and bleeding during pregnancy. Adenomas that bleed or rupture during pregnancy are frequently found to be large (96.5 cm). Major catastrophic events because of HCAs in pregnancy comes from case reports from 1970s to 1990s, which may be attributable to the delay in clinical and imaging diagnoses. 9 Overall, up to 59% of large HCAs may rupture with resultant maternal and fetal mortality in the range of 30% to 40%. 9 Although pregnancy is not contraindicated in patients with HCAs, 6 given the risk of growth and subsequent increased risk of hemorrhage during pregnancy, HCAs in young women should be managed more aggressively. Some authors recommend early intervention even for small HCAs (G5 cm) in young women who wish to become pregnant, 63 although others reserve it for HCAs larger than 5 cm or HCAs, which continue to grow despite stopping the offending drugs. 64 Management of HA should be similar to solitary HCAs with aggressive approach (surgery or embolization) for symptomatic lesions or asymptomatic lesions larger than 5 cm and watchful waiting for smaller lesions. 50,59 Withdrawal of exogenous steroid intake has been found to decrease the risk of bleeding but does not affect the lesion size or growth. 59,65 Liver transplantation is no more indicated as the primary treatment strategy for all patients with HA and is reserved for cases with progressive liver failure or malignant transformation. 20 CONCLUSIONS Recent insights into distinctive cytogenetic abnormalities and genotype-phenotype correlation in HCAs have brought paradigm shifts in our understanding of the imaging findings with significant implications for management. Adenomas with specific genetic abnormalities demonstrate characteristic clinicobiologic behavior, natural history, response to treatment, and prognosis. Although I-HCAs frequently undergo hemorrhage, HCAs with A-catenin mutations show increased predisposition to transformation to HCCs. Dynamic, multiphase MDCTand MRI play a key role in the diagnosis, characterization, and staging of HCAs as well as in surveillance after management with one of many therapeutic options. REFERENCES 1. Rosenberg L. The risk of liver neoplasia in relation to combined oral contraceptive use. Contraception. 1991;43:643Y Nime F, Pickren JW, Vana J, et al. The histology of liver tumors in oral contraceptive users observed during a national survey by the American College of Surgeons Commission on Cancer. Cancer. 1979;44:1481Y Staib FKM, Krupp M, Maass T, et al. Comparative genomics analysis of hepatic adenoma, normal liver, and HCC revealed closer relationship of adenoma to HCC with favorable outcome. JHepatol. 2010;52:S319YS Bioulac-Sage P, Laumonier H, Couchy G, et al. Hepatocellular adenoma management and phenotypic classification: the Bordeaux experience. Hepatology. 2009;50:481Y Bioulac-Sage P, Laumonier H, Laurent C, et al. Hepatocellular adenoma: what is new in Hepatol Int. 2008;2:316Y Dokmak S, Paradis V, Vilgrain V, et al. A single-center surgical experience of 122 patients with single and multiple hepatocellular adenomas. Gastroenterology. 2009;137:1698Y Zucman-Rossi J, Jeannot E, Nhieu JT, et al. Genotype-phenotype correlation in hepatocellular adenoma: new classification and relationship with HCC. Hepatology. 2006;43:515Y Farges O, Dokmak S. Malignant transformation of liver adenoma: an analysis of the literature. Dig Surg. 2010;27:32Y Cobey FC, Salem RR. A review of liver masses in pregnancy and a proposed algorithm for their diagnosis and management. Am J Surg. 2004;187:181Y Edmondson HA, Henderson B, Benton B. Liver-cell adenomas associated with use of oral contraceptives. N Engl J Med. 1976;294:470Y Scalori A, Tavani A, Gallus S, et al. Oral contraceptives and the risk of focal nodular hyperplasia of the liver: a case-control study. Am J Obstet Gynecol. 2002;186:195Y Heinemann LA, Weimann A, Gerken G, et al. Modern oral contraceptive use and benign liver tumors: the German Benign Liver Tumor Case-Control Study. Eur J Contracept Reprod Health Care. 1998;3:194Y Bioulac-Sage P, Balabaud C, Bedossa P, et al. Pathological diagnosis of liver cell adenoma and focal nodular hyperplasia: Bordeaux update. J Hepatol. 2007;46:521Y Aseni P, Sansalone CV, Sammartino C, et al. Rapid disappearance of hepatic adenoma after contraceptive withdrawal. J Clin Gastroenterol. 2001;33:234Y Bioulac-Sage P, Balabaud C, Zucman-Rossi J. Focal nodular hyperplasia, hepatocellular adenomas: past, present, future. Gastroenterol Clin Biol. 2010;34:355Y Carrasco D, Barrachina M, Prieto M, et al. Clomiphene citrate and liver-cell adenoma. N Engl J Med. 1984;310:1120Y Vazquez JJ, Marigil MA. Liver-cell adenoma in an epileptic man on barbiturates. Histol Histopathol. 1989;4:301Y Zucman-Rossi J. Genetic alterations in hepatocellular adenomas: recent findings and new challenges. JHepatol. 2004;40: 1036Y Terkivatan T, de Wilt JH, de Man RA, et al. Indications and long-term outcome of treatment for benign hepatic tumors: a critical appraisal. Arch Surg. 2001;136:1033Y Leese T, Farges O, Bismuth H. Liver cell adenomas. A 12-year surgical experience from a specialist hepato-biliary unit. Ann Surg. 1988;208:558Y Barthelmes L, Tait IS. Liver cell adenoma and liver cell adenomatosis. HPB (Oxford). 2005;7:186Y Zech CJ, Grazioli L, Breuer J, et al. Diagnostic performance and description of morphological features of focal nodular hyperplasia in Gd-EOB-DTPAYenhanced liver magnetic resonance imaging: results of a multicenter trial. Invest Radiol. 2008;43: 504Y Paradis V, Champault A, Ronot M, et al. Telangiectatic adenoma: an entity associated with increased body mass index and inflammation. Hepatology. 2007;46:140Y Rebouissou S, Amessou M, Couchy G, et al. Frequent in-frame somatic deletions activate gp130 in inflammatory hepatocellular tumours. Nature. 2009;457:200Y Bioulac-Sage P, Rebouissou S, Thomas C, et al. Hepatocellular adenoma subtype classification using molecular markers and immunohistochemistry. Hepatology. 2007;46:740Y748. * 2011 Lippincott Williams & Wilkins 165

8 Shanbhogue et al J Comput Assist Tomogr & Volume 35, Number 2, March/April Spannbauer MM, Trautwein C. Frequent in-frame somatic deletions activate gp130 in inflammatory hepatocellular tumors. Hepatology. 2009;49:1387Y Tang Y, Kitisin K, Jogunoori W, et al. Progenitor/stem cells give rise to liver cancer due to aberrant TGF-beta and IL-6 signaling. Proc Natl Acad Sci U S A. 2008;105:2445Y Schutyser E, Struyf S, Van Damme J. The CC chemokine CCL20 and its receptor CCR6. Cytokine Growth Factor Rev. 2003;14:409Y Laumonier H, Bioulac-Sage P, Laurent C, et al. Hepatocellular adenomas: magnetic resonance imaging features as a function of molecular pathological classification. Hepatology. 2008;48:808Y Laumonier HB-SP, Laurent C, Sa Cunha A, et al. Contrast-enhanced ultra sonography to identify the Two major subtypes of hepatocellular adenoma. J Hepatol. 2010;52:S183YS Yamagata K, Furuta H, Oda N, et al. Mutations in the hepatocyte nuclear factor-4> gene in maturity-onset diabetes of the young (MODY1). Nature. 1996;384:458Y Bacq Y, Jacquemin E, Balabaud C, et al. Familial liver adenomatosis associated with hepatocyte nuclear factor 1> inactivation. Gastroenterology. 2003;125:1470Y Rebouissou S, Imbeaud S, Balabaud C, et al. HNF1> inactivation promotes lipogenesis in human hepatocellular adenoma independently of SREBP-1 and carbohydrate-response element-binding protein (ChREBP) activation. J Biol Chem. 2007;282:14437Y Shih DQ, Bussen M, Sehayek E, et al. Hepatocyte nuclear factor-1> is an essential regulator of bile acid and plasma cholesterol metabolism. Nat Genet. 2001;27:375Y Cavalieri E, Frenkel K, Liehr JG, et al. Estrogens as endogenous genotoxic agentsvdna adducts and mutations. J Natl Cancer Inst Monogr. 2000;75Y Chung KY, Mayo-Smith WW, Saini S, et al. Hepatocellular adenoma: MR imaging features with pathologic correlation. AJR Am J Roentgenol. 1995;165:303Y Ichikawa T, Federle MP, Grazioli L, et al. Hepatocellular adenoma: multiphasic CT and histopathologic findings in 25 patients. Radiology. 2000;214:861Y Paulson EK, McClellan JS, Washington K, et al. Hepatic adenoma: MR characteristics and correlation with pathologic findings. AJR Am J Roentgenol. 1994;163:113Y Buendia MA. Genetic alterations in hepatoblastoma and hepatocellular carcinoma: common and distinctive aspects. Med Pediatr Oncol. 2002;39:530Y Longerich T, Schirmacher P. A new link between cancer and inflammation? J Hepatol. 2009;51:230Y Chen YW, Jeng YM, Yeh SH, et al. P53 gene and Wnt signaling in benign neoplasms: A-catenin mutations in hepatic adenoma but not in focal nodular hyperplasia. Hepatology. 2002;36:927Y Lee PJ. Glycogen storage disease type I: pathophysiology of liver adenomas. Eur J Pediatr. 2002;161(suppl 1):S46YS Kishnani PS, Chuang TP, Bali D, et al. Chromosomal and genetic alterations in human hepatocellular adenomas associated with type IA glycogen storage disease. Hum Mol Genet. 2009;18:4781Y De Souza AT, Hankins GR, Washington MK, et al. Frequent loss of heterozygosity on 6q at the mannose 6-phosphate/insulin-like growth factor II receptor locus in human hepatocellular tumors. Oncogene. 1995;10:1725Y Yamada T, De Souza AT, Finkelstein S, et al. Loss of the gene encoding mannose 6-phosphate/insulin-like growth factor II receptor is an early event in liver carcinogenesis. Proc Natl Acad Sci U S A. 1997;94:10351Y Cassiman D, Libbrecht L, Verslype C, et al. An adult male patient with multiple adenomas and a hepatocellular carcinoma: mild glycogen storage disease type IA. J Hepatol. 2010;53:213Y Flejou JF, Barge J, Menu Y, et al. Liver adenomatosis. An entity distinct from liver adenoma? Gastroenterology. 1985;89:1132Y Lewin M, Handra-Luca A, Arrive L, et al. Liver adenomatosis: classification of MR imaging features and comparison with pathologic findings. Radiology. 2006;241:433Y Greaves WO, Bhattacharya B. Hepatic adenomatosis. Arch Pathol Lab Med. 2008;132:1951Y Vetelainen R, Erdogan D, de Graaf W, et al. Liver adenomatosis: re-evaluation of aetiology and management. Liver Int. 2008;28: 499Y Prasad SR, Sahani DV, Mino-Kenudson M, et al. Benign hepatic neoplasms: an update on cross-sectional imaging spectrum. J Comput Assist Tomogr. 2008;32:829Y Grazioli L, Morana G, Kirchin MA, et al. Accurate differentiation of focal nodular hyperplasia from hepatic adenoma at gadobenate dimeglumineyenhanced MR imaging: prospective study. Radiology. 2005;236:166Y Kudo M. Multistep human hepatocarcinogenesis: correlation of imaging with pathology. J Gastroenterol. 2009;44(suppl 19):112Y Grazioli L, Federle MP, Brancatelli G, et al. Hepatic adenomas: imaging and pathologic findings. Radiographics. 2001;21:877Y892; discussion 892Y Erdogan D, Busch OR, van Delden OM, et al. Management of spontaneous haemorrhage and rupture of hepatocellular adenomas. A single centre experience. Liver Int. 2006;26:433Y Stoot JH, van der Linden E, Terpstra OT, et al. Life-saving therapy for haemorrhaging liver adenomas using selective arterial embolization. Br J Surg. 2007;94:1249Y Terkivatan T, de Wilt JH, de Man RA, et al. Treatment of ruptured hepatocellular adenoma. Br J Surg. 2001;88:207Y Marini P, Vilgrain V, Belghiti J. Management of spontaneous rupture of liver tumours. Dig Surg. 2002;19:109Y Ribeiro A, Burgart LJ, Nagorney DM, et al. Management of liver adenomatosis: results with a conservative surgical approach. Liver Transpl Surg. 1998;4:388Y Bioulac-Sage P, Balabaud C, Zucman-Rossi J. Subtype classification of hepatocellular adenoma. Dig Surg. 2010;27:39Y Rhim H, Lim HK, Kim YS, et al. Percutaneous radiofrequency ablation of hepatocellular adenoma: initial experience in 10 patients. J Gastroenterol Hepatol. 2008;23:e422Ye van der Sluis FJ, Bosch JL, Terkivatan T, et al. Hepatocellular adenoma: cost-effectiveness of different treatment strategies. Radiology. 2009;252:737Y Terkivatan T, de Wilt JH, de Man RA, et al. Management of hepatocellular adenoma during pregnancy. Liver. 2000;20:186Y van Aalten SM, Terkivatan T, de Man RA, et al. Diagnosis and treatment of hepatocellular adenoma in the Netherlands: similarities and differences. Dig Surg. 2010;27:61Y Chiche L, Dao T, Salame E, et al. Liver adenomatosis: reappraisal, diagnosis, and surgical management: eight new cases and review of the literature. Ann Surg. 2000;231:74Y * 2011 Lippincott Williams & Wilkins