Journal Watch July-August 2016

A large cross-sectional study of 1181 members of the Framingham Third Generation cohort was performed to investigate the accuracy of serum levels of aminotransferases in detection of hepatic steatosis. Participants with hepatic steatosis had higher ALT and AST levels compared with those without steatosis. The ratio of ALT/AST identified people with hepatic steatosis with highest c-static value of 0.728. A Framingham Steatosis Index (FSI) was derived to include patient age, sex, BMI, triglycerides, hypertension, diabetes, and ratio of ALT/AST for prediction of hepatic steatosis. The model was validated in the external multiethnic NHANES III cohort of 4489 participants, which demonstrated good discrimination and calibration.

No biomarker is available for detection acute-on-chronic liver failure (ACLF), a syndrome occurs in cirrhosis characterized by acute decompensation of the disease, organ failure(s) and high mortality rate.  Ariza et al. evaluated urine and plasma levels of neutrophil gelatinase-associated lipocalin (NGAL) in 716 patients with complications of cirrhosis. 148 patients (20.7%) had ACLF, 568 (79.3%) had acute decompensation without ACLF. Patients with ACLF had significantly higher levels of both urine and plasma NGAL compared with those without ACLF. Multivariate analysis revealed uNGAL and pNGAL were independent predictors for ACLF as well as 28 day transplant-free mortality.  Additional gene expression study in 29 liver biopsies found lipocalin-2 gene (LCN2) was marked upregulated in patients with ACLF compared to all other groups. The data suggests that NGAL is a diagnostic and prognostic biomarker for ACLF.

Conventional subtyping of intrahepatic cholangiocarcinoma (ICC) into large duct-type and small duct-type ICC subtypes can sometimes be challenging due to ambiguity in anatomic location and overlapping histologic features.  Given these limitations, these authors investigated the utility of grouping ICC into 2 subtypes: Type 1 (mucin production; S100P positive; N-cadherin and NCAM negative) and Type 2 (absent or scant mucin production; S100P negative; N-cadherin and/or NCAM positive;).  102 consecutive ICCs were evaluated by Alcian blue and immunohistochemistry staining; KRAS and IDH1/IDH2 mutation testing by PCR and FGFR2 translocation testing by FISH were performed on the majority of cases.  By conventional subtyping, cases were categorized as follows: 33 large duct-type, 36 small duct-type, 33 indeterminate ICC.  Using the proposed classification system, the same cases were grouped as follows: 42 Type 1, 56 Type 2, 4 indeterminate.  Type 1 usually had higher levels of serum CEA (median 6.2 ng/mL) and CA 19-9 (median 1159 U/mL) and tumors more often showed perineural invasion and lymph node metastasis.  KRAS mutation was significantly more frequent in Type 1 (29%) whereas IDH1/IDH2 mutation (40%) and FGFR2 translocation (11%) were exclusively seen in Type 2.  Type 2 tumors were more frequently present in patients with chronic liver diseases.  Of the 3 IHC markers, S100P was the most consistent and discriminatory; N-cadherin and NCAM should be reserved for cases where S100P is indeterminate.  While the authors demonstrated clinicopathologic differences between Type 1 and Type 2 ICC, the significance of these findings on outcome is unclear given that multivariate Cox regression showed no prognostic differences.

The immunostain panel of liver fatty acid-binding protein (LFABP), serum amyloid A (SAA) protein, C-reactive protein (CRP), and glutamine synthetase (GS) are used to subtype hepatocellular adenoma (HCA).  In practice, however, some pathologists use these immunostains as an aid in diagnosis when distinguishing between HCC and HCA in a manner that is akin to “off label” use of medications for certain treatments by physicians in specific clinical settings.  This study sought to determine the validity of such practice by examining expression of these markers in 159 HCCs (47 full section tissue blocks, 112 from tissue microarray).  Proportion of HCCs showing positive staining are as follows: SAA 17%, CRP 54%, GS 49%; LFABP showed loss of expression in 23%.  An additional 7 cases of fibrolamellar HCC were positive for CRP in 100% and SAA in 43%; LFABP was lost in 57%.  Although 79% of HCCs that showed loss of expression of LFABP also stained positively for GS, there was no association between marker expression and other clinicopathologic features.  This study reaffirms that the immunostain panel comprised of LFABP, SAA, CRP, and GS is intended for use in subtyping HCA.  This panel should only be used after the diagnosis of HCA has been established and not for distinguishing HCC from HCA.

This is the largest case series to date investigating clinicopathologic features in 18 immunocompetent patients with Cystoisospora belli infection of the gallbladder.  Histologic features evaluated included the presence, type and/or degree of cholecystitis, epithelial disarray (displacement of nuclei from their normal basal location by parasite-containing vacuoles), architectural distortion, mural thickening/serositis, intraepithelial lymphocytosis, and intramucosal eosinophilia.  Subjects ranged from 17 to 61 years of age (mean 33 years) with a female predominance (4.3:1).  Clinical indications were not unique and included usual reasons for cholecystectomy.  Each case was diagnosed on light microscopy by observation of the characteristic eosinophilic, oval-to-banana-shaped intraepithelial parasitic forms within perinuclear vacuoles, although such findings were focal in 8.  94% of cases displayed epithelial disarray and 83% showed rare intraepithelial lymphocytosis.  Of the 11 cases that had follow-up information, none had evidence of Cystoisospora infection within the biliary or tubular GI tract.  Biliary cystoisosporiasis is likely an underrecognized entity that presents with nonspecific symptoms and appears to be a self-limited infection in immunocompetent patients.

IgG4-related disease involving the pancreas/pancreatobiliary tree can pose particularly difficult diagnostic challenges, and can closely mimic primary sclerosing cholangitis (PSC) and/or pancreatobiliary malignancy. Patients often have elevated serum IgG4 levels; however, approximately 30% of patients with IgG4-related pancreatitis (and/or cholangiopathy) have normal levels of serum IgG4, and elevated IgG4 levels can be seen in up to 30% of patient with PSC or pancreatobiliary malignancy. The authors demonstrate the peripheral blood presence of IgG4 B-cell receptor clones in IgG4-related pancreatitis/cholangitis patients as compared to patients with PSC and pancreatobiliary malignancies. They also demonstrate the diagnostic accuracy of IgG4/IgG RNA ratio by PCR in peripheral blood. The IgG4/IgG RNA ratio performed better than serum IgG4 in sensitivity (94% vs 86%) and specificity (99% vs 73%), and correlated with treatment response.

The authors report an unusual case of a 41 year-old woman with a past medical history of neonatal portal vein thrombosis with cavernous transformation of the portal vein, and multiple hepatocellular lesions including large regenerative nodules, focal nodular hyperplasias, and HNF1A inactivated adenomas measuring up to 11.9 cm. The largest resected lesion was an HNF1A inactivated adenoma with several foci of well-differentiated HCC. The tumor did not show diffuse glutamine synthetase expression, or nuclear/cytoplasmic staining with B-catenin.

Two common mouse models for chronic liver disease — choline-deficient, ethionine-supplemented diet (CDE) and thioacetamide supplementation (TAA) — were compared. CDE mice showed portal-based injury, whereas TAA mice had centrizonal-based disease and a more severe phenotype.

In this study, the impact of coffee consumption was investigated in regard to progression of end-stage liver disease (ESLD) and long-term survival in patients who had liver transplantation. Coffee consumption habits in 379 patients with ESLD awaiting liver transplantation and 260 patients after liver transplantation were studied. One hundred ninety-five patients with ESLD consumed coffee on a daily basis, and 184 did not. Actuarial survival was lower (P = 0.041) in non-coffee drinkers (40.4 ± 4.3 months, 95% confidence interval [CI]: 32.0–48.9) compared to coffee drinkers (54.9 ± 5.5 months, 95% CI: 44.0–65.7). In subgroup analysis, the survival of patients with alcoholic liver disease (ALD; P = 0.020) and primary sclerosing cholangitis (PSC; P = 0.017) was improved with coffee intake while unaffected in patients with chronic viral hepatitis (P = 0.517) or other liver disease entities (P = 0.652). Multivariate analysis showed that coffee consumption in PSC and ALD patients was an independent risk factor (odds ratio [OR]: 1.94; 95% CI: 1.15–3.28; P = 0.013) along with MELD score (OR: 1.13; 95% CI: 1.09–1.17; P = 0.000). Following liver transplantation, long-term survival was longer in coffee drinkers (61.8 ± 2.0 months, 95% CI: 57.9–65.8) than non-drinkers (52.3 ± 3.5 months, 95% CI: 45.4 –59.3; P = 0.001). Coffee consumption reduced disease progression in ALD and PSC patients with ESLD and improved long-term survival after liver transplantation. Thus, regular coffee intake might be recommended for these patients.

This study reports a series of 17 cases of a novel vascular tumor of the liver designated as ‘hepatic small vessel neoplasm’ (HSVN). These tumors occur in adults, have infiltrative borders and are comprised of small sized vascular channels lined by endothelial cells that can show hobnail appearance. Immunohistochemistry for endothelial markers (CD31, CD34, FLI-1) is positive. The hobnail appearance along with infiltrative borders raises the possibility of angiosarcoma. The lack of significant cytologic atypia and mitoses along with low Ki-67 (typically <10%) favors HSVN. Strong p53 and diffuse c-Myc staining is more common in angiosarcoma.  HSVNs showed an activating hotspot GNAQ mutation in 2 of 3 (67%) cases tested, while a hotspot mutation in PIK3CA was seen in one case. Complications like rupture, hemorrhage, disseminated intravascular coagulation, Kasabach-Merritt syndrome, recurrence or metastasis have not been reported with HSVN, but the study recommends complete resection and close follow-up as the outcome data is based on limited number of cases.

The authors reviewed discrepancies between the primary and specialist liver pathologists in 1265 liver biopsy reports over 4 years. They concluded that specialist review was necessary but not in all cases, as non-specialist pathologists did well with cases like viral hepatitis, fatty liver, and malignancies, while most of the 59% cases with discrepancies were biliary, autoimmune hepatitis, and cases with vascular/architectural abnormalities. Aims: Liver pathology is a challenging subspecialty, with histopathologists frequently seeking specialist opinions. This study aims to determine the impact of specialist reviews on the final diagnosis and patient management. Methods and results: Agreement with the initial reporting centre in the histopathological diagnosis of 1265 liver biopsies was determined. The nature of differences was explored in more depth for 103 discrepant cases. Differences in the histopathological interpretation were present in 749 of 1265 (59%) biopsies, of which 505 of 749 (67%) were predicted at the time of reporting to impact upon patient management. Agreement was good in cases with chronic viral hepatitis, fatty liver disease, malignancy and minimal pathological changes, while diagnostic differences occurred in more than 70% with biliary disease, autoimmune hepatitis or vascular/architectural changes. A clinical review of a subset of reports with histopathological differences predicted changes in patient management in 63 of 103 (61%). Conclusions: Clinically significant differences in liver biopsy interpretation between local pathologists and subspecialists are common. Diagnoses with frequent discrepancies, such as biliary disease, may benefit from a specialist review as standard when diagnosed initially, while cases requiring specialist advice from disease subgroups where discrepancies are less common, such as chronic viral hepatitis, could be selected during the clinicopathological conference process.

The authors investigated the expression of ezrin in cholangiocarcinoma (intra-hepatic), a plasma membrane protein restricted to cholangiocytes in the liver. Using immunohistochemistry on 94 tumor micro-arrays, they found that 48% of their cases had weak/lost ezrin expression, more likely in peripheral than perihilar tumors, and correlated with a more aggressive phenotype. Aims: Ezrin connects proteins from the plasma membrane to the subcortical cytoskeleton, and contributes to epithelial integrity by interacting with the cell–cell adhesion molecule E-cadherin. In the liver, ezrin is restricted to cholangiocytes, where it regulates biliary secretory functions. During carcinogenesis, ezrin expression is impaired and associated with enhancement of cell migratory activity in cancer cells; therefore, we aimed to analyze ezrin in cholangiocarcinogenesis. Methods and results: Ezrin expression was evaluated by immunohistochemistry on tissue microarrays from 94 surgical specimens of intrahepatic cholangiocarcinoma (CCA), and correlated with clinicopathological factors and E-cadherin expression. Ezrin function was also analyzed in human CCA cell lines. In CCA, ezrin was negative/weakly expressed in 49 cases (52%) and moderately/strongly expressed in 45 cases (48%), mostly in cell cytoplasm. The negative/weak expression of ezrin was more frequent in peripheral than in perihilar CCA (P = 0.002), and was associated with high tumour size (P = 0.001), low mucus secretion (P = 0.042), the presence of satellite nodules (P = 0.024), and ectopic cytoplasmic expression of E-cadherin (P = 0.005). In vitro, silencing of ezrin in CCA cells caused internalization of E-cadherin and favored cell migration. Conclusions: Ezrin is down-regulated during cholangiocarcinogenesis, and its loss results in a more aggressive phenotype.