A Limited Immunohistochemical Panel Can Subtype Hepatocellular Adenomas for Routine Practice

Brent K. Larson, DO; Maha Guindi, MD

Disclosures

Am J Clin Pathol. 2017;147(6):557-570. 

In This Article

Results

A total of 46 specimens from 33 patients were obtained. Forty-five were surgical specimens, and one specimen was from autopsy material. Two specimens from one patient were sent from an outside institution for consultation; all others were in-house cases. Clinical characteristics of the patient cohort are summarized in Table 2. In brief, specimens from females outnumbered those from males by a ratio of 3.1:1. Age ranged from 2 to 81 years, with a median of 39 years. Nineteen female patients had a clinical history of oral contraceptive use, ranging in duration from 2 to 34 years and varying from remote to ongoing use. No patients had a definitive clinical history of anabolic steroid use. Thirteen of 25 patients for whom data were available had no history of alcohol use. Consumption by 10 was variously described as occasional, infrequent, social, or less than three drinks per week. Two patients had six or more drinks per week. BMI was available for 27 patients and ranged from 16.5 to 33.5 kg/m2, with a median value of 24.9 kg/m2. Thirteen patients were overweight or obese (BMI ≥25 kg/m2). Chart review revealed seven patients with diagnoses associated with metabolic syndrome. One had the complete metabolic syndrome. One patient had a personal history of an inflammatory condition (Crohn disease), and one other had a family history of an inflammatory condition (first-degree relative with ulcerative colitis). One patient had glycogen storage disease type I.

General pathologic adenoma characteristics are summarized in Table 3. HCA size on gross examination ranged from 0.2 to 24 cm, with a median of 4.5 cm. Sixteen of 27 patients for whom data were available had multiple masses on gross or radiologic examination, including one with more than 10 (adenomatosis). Four patients had hemorrhage/rupture of adenomas.

Morphologic examination produced a working morphologic diagnosis of IHCA in 25 (54%) of 46 cases, b-HCA in three (7%) cases, and untypable HCA in 18 (39%) cases. Immunohistochemical staining produced final diagnoses of IHCA in 16 (35%) cases, b-HCA in four (24%) cases, b-IHCA in seven (15%) cases, and untypable HCA in 19 (41%) cases. Image 1, Image 2, Image 3, and Image 4 illustrate the typical findings in IHCA, b-HCA, untypable HCA, and b-IHCA, respectively.

Image 1.

Features of inflammatory hepatocellular adenoma. A, Telangiectatic sinusoids and inflammatory infiltrate (H&E, ×100). B, A pseudoportal tract with multiple arterioles, admixed inflammatory cells, and a proliferation of biliary ductal epithelial cells (H&E, ×400). C, Perivenular glutamine synthetase positivity (×40). D, Diffuse serum amyloid A positivity (×40). E, Serum amyloid A positivity in hepatocytes with an adjacent pseudoportal tract (upper left) (×200).

Image 2.

Features of β-catenin–activated hepatocellular adenoma. A, Pseudoglandular formations (H&E, ×400). B, Diffuse glutamine synthetase positivity (×40). C, Nuclear β-catenin positivity (×400). D, Negative serum amyloid A staining (×100).E, Glutamine synthetase positivity at high power (×400). F, Strong glutamine synthetase staining, less diffuse than shown in E (×400).

Image 3.

Features of untypable hepatocellular adenoma. A, Abundant steatosis (H&E, ×40). B, Negative glutamine synthetase staining (×40). C, Membranous β-catenin staining (×400). D, Patchy serum amyloid A staining (×40). E, Patchy serum amyloid A at high power (×200).

Image 4.

Features of β-catenin–activated inflammatory hepatocellular adenoma. A, Telangiectatic sinusoids (H&E, ×40). B, Diffuse glutamine synthetase positivity (×40). C, Patchy nuclear β-catenin positivity (×400). D, Diffuse serum amyloid A positivity (×40).

Immunohistochemistry confirmed the diagnosis in 15 (33%) cases. Thirteen (52%) of 25 cases with a morphologic diagnosis of IHCA were confirmed as either IHCA or b-IHCA, and two (67%) of three cases with a morphologic diagnosis of b-HCA were confirmed as either b-HCA or b-IHCA (both were b-IHCA). Eleven (61%) of 18 cases with a morphologic diagnosis of untypable HCA remained untyped after immunostaining. Figure 1 details the morphologic-immunophenotypic concordance and enumerates all changes of diagnoses.

Figure 1.

Morphologic-immunophenotypic concordance and changes of diagnoses. b-HCA, β-catenin–activated hepatocellular adenoma; b-IHCA, β-catenin–activated inflammatory hepatocellular adenoma; HCA, hepatocellular adenoma; IHCA, inflammatory hepatocellular adenoma.

Specific morphologic features were also associated with different adenoma subtypes Table 4. Telangiectatic sinusoids, pseudoportal tracts, ductular reaction, and inflammation were present in 10 (63%), 13 (81%), four (25%), and 13 (81%) of the 16 immunophenotypic IHCAs, respectively, and in one (14%), two (29%), two (29%), and four (57%) b-IHCAs, respectively. Pseudoportal tracts, ductular reaction, and inflammatory infiltrates were significantly correlated with an immunophenotypic diagnosis of IHCA or b-IHCA (P = .0169, .0216, and .0012, respectively). The presence of ductular reaction was 100% specific for IHCA or b-IHCA, but sensitivity was poor. Neither nuclear atypia nor pseudoglandular architecture was significantly correlated with b-HCA or b-IHCA. Similarly, macrovesicular steatosis was not significantly correlated with untypable HCAs.

Five HCAs harbored foci of HCC Image 5. The size of these HCAs ranged from 3.0 to 24.0 cm, with only one being smaller than 5.0 cm. Two patients each had only one examined HCA (one b-HCA and one b-IHCA), each with foci of HCC. One patient had one benign HCA and a separate adenoma with foci of HCC (both untypable HCAs). One patient had one benign adenoma (IHCA) and two with foci of HCC (both b-HCAs). One patient had two benign adenomas (both untypable HCAs) and a separate nodule of HCC without associated HCA.

Image 5.

Features of hepatocellular carcinoma ex hepatocellular adenoma. A, Hepatocellular carcinoma (right) arising in a hepatocellular adenoma (left) (H&E, ×100). B, This adenoma (left) shows patchy strong glutamine synthetase positivity. The neighboring nodule of carcinoma (right) shows only rare glutamine synthetase positivity (×100). C, Membranous β-catenin staining in a hepatocellular adenoma (left) and adjacent hepatocellular carcinoma (×100). D, Patchy serum amyloid A positivity in a nodule of carcinoma (below) and an adjacent adenoma (above) with no staining (×100).

Of the nine patients with multiple adenomas confirmed by histopathologic examination, eight had HCAs that shared the same immunophenotype. Three patients had multiple IHCAs, one patient had multiple b-HCAs, and four patients had multiple untypable HCAs. The aforementioned patient with two b-HCAs harboring malignant foci was the only patient to have a separate adenoma of an entirely different subtype (IHCA).

Patients with IHCAs or b-IHCAs had a greater average BMI than patients with other types of HCAs (26.3 vs 22.9 kg/m2), although this difference was not statistically significant (P = .2519). Nine patients with IHCAs or b-IHCAs had clinical profiles that could be consistent with an inflammatory syndrome (including diabetes mellitus, hypertension, hyperlipidemia, or a diagnosed inflammatory disorder), whereas six did not. Five patients with other adenomas had a clinical profile that could be consistent with an inflammatory syndrome and seven did not, with no significant association found between an inflammatory clinical profile and IHCA/b-IHCA (P = .4495).

Nine previous studies examining immunohistochemical subtyping of HCAs were found.[6,9,18–24] Three series were excluded from analysis, as they were small (fewer than 20 total adenomas) and/or reported on HCAs in East Asian populations.[9,20,23] Table 5 presents a comparison of the rates of each subtype using this study's limited immunohistochemical panel compared with the rates previously published using larger panels. Using a limited panel, the overall distribution of subtypes was significantly different from five of six previous studies.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....