The most commonly encountered differential diagnostic challenge in the liver is HCC versus intrahepatic cholangiocarcinoma or metastatic adenocarcinoma
. Some of these diagnostic challenges can be attributed to: a) The liver represent one of the three most common sites of metastasis, b) HCCs may show a variety of histologic patterns, mimicking a wide variety of malignant tumors. In addition, a number of metastatic tumours, notably from the breast, pancreas, kidney and adrenals may mimic the trabecular, liver-like pattern of HCC, c) Cholangiocarcinoma and HCC often share overlapping morphologic appearances, d) Complicating the diagnostic process is that pathologists are frequently asked to handle and diagnose tiny liver needle core biopsies with various biopsy artifacts
[9, 22]. A limited number of diagnostically useful immunohistochemical markers have been applied in an attempt to differentiate HCC from liver metastases or cholangiocarcinoma including; HepPar-1, polyclonal carcinoembryonic antigen (CEA), and CD10, with alfa-fetoprotein (AFP) and glypican-3 labeling some HCCs
. However, the utility of each of these markers has significant diagnostic limitations
A recent study of Hajósi-Kalcakosz et al.
 published in 2012 investigated enhancer of zeste homologue 2 (EZH2) as a new marker of HCC. They reported that EZH2 was detected by immunohistochemistry in nearly all the investigated HCC, CC, hepatoblastoma, metastatic liver tumors and several other childhood cancers. On the contrary, none of the hepatocellular or biliary adenomas, high grade dysplastic or cirrhotic nodules was positive. Thus, this study concluded that EZH2 is a sensitive and reliable immune marker of hepatocellular carcinoma, compared to non-malignant hepatocellular lesions. However, EZH2 is not specific for HCC, since almost all the investigated malignant liver tumors were positive as well regardless of their histogenesis. Consequently, this marker does not provide help in differentiating the specific histogenesis of liver tumors, but it may well be very useful to differentiate malignant hepatocellular and cholangiocellular tumors from benign tumors and reactive lesions.
Moreover, special stains, such as reticulin stain and CD34 immunostain, are very helpful in the diagnosis of well differentiated HCC. Most studies have shown that absent or decreased reticulin stain or an abnormal reticulin pattern with widened trabeculae is reliable for the diagnosis of well-differentiated HCC. However, Hong et al.
 reported two cases of well-differentiated HCC with an unusual reticulin staining pattern in their primary biopsies. They suggested that HCC may have diverse reticulin patterns in different portions of the tumor. In a small specimen, such as core biopsy, if only the portion of tumor with well preserved reticulin network is present, the diagnosis can be challenging. Thus, it is important to recognize the presence of different reticulin staining patterns in the evaluation of small biopsies for the diagnosis of HCC.
Arginase-1 has been described in recent literature as a new potential immunohistochemical marker of hepatocellular differentiation
. Only few studies investigated arginase −1 expression in HCC and most of these reports performed on fine needle aspiration cytology
[5, 25, 26] with some variation in their interpretations as regards its sensitivity and specificity. Therefore; the primary purpose of the current study was to examine the immunohistochemical staining of arginase-1 in cases of HCC, metastatic carcinoma involving the liver and cholangiocarcinoma as compared to HepPar-1. This is in an attempt to further define its diagnostic utility as a reliable positive marker in differentiating these tumors. HepPar-1 was selected to be compared with this new marker as it is conventially used and has been found to be overrated as a hepatoma marker. The present study examined arginase-1 and HepPar-1 expression in 50 HCC cases, 38 cases of metastatic carcinomas to the liver from varying sites, 12 cases of cholangiocarcinoma and 10 specimens of normal liver tissues. In addition, the non-neoplastic liver tissues adjacent to HCC or metastatic carcinomas were also investigated.
The results revealed that arginase -1 showed a significantly higher overall sensitivity for diagnosis of HCC (84%) compared to HepPar -1 (70%). This confirm the conclusion of the previous studies
[5, 6, 25–27]. It is worth mentioning that there were no cases were positive for HepPar-1, with concurrent negative arginase-1 staining. In addition, arginase-1 showed more diffuse staining in HCC (76.2%) than HepPar-1 (57.1%). This makes interpretation of arginase -1 easier especially in limited liver biopsies.
Furthermore, arginase-1 gave a sensitivity of 100%, 90%, and 44.4% in well, moderately, and poorly differentiated HCCs, respectively, whereas, in comparison, HepPar-1 demonstrated sensitivities of 100%, 73.3%, and 22.2% for well, moderately, and poorly differentiated tumors, respectively. Therefore, arginase-1 showed better sensitivity compared with HepPar-1 in identifying higher grade HCC. This is relatively in accordance with the original paper describing the antibody of Yan et al.
 who found more marked difference between both immunomakers in poorly differentiated HCCs, in which the sensitivities of arginase-1 and HepPar-1 were 85.7% and 46.4%, respectively. This finding is very useful because one of the most frequent diagnostic challenges facing a pathologist examining liver focal lesion is distinguishing between poorly differentiated HCC from a metastasis, especially in small biopsy specimen. The lower diagnostic sensitivity in our study as compared to that of Yan et al.
 may be because of the smaller sample size. In contrast, Timek et al.
 failed to demonstrate a better sensitivity of arginase-1 for higher–grade HCC compared with HepPar-1 and they explained that by the small sampling of the cytologic specimens in the moderately to poorly differentiated HCC category (n =7), limited amount of sample for each case, and patchy/focal staining for arginase-1 in higher-grade HCC.
Moreover, we observed diffuse and strong immunostaining for both arginase-1 and HepPar-1 in the non- neoplastic cirrhotic liver tissues adjacent to HCC as well as those adjacent to MC. This supports the study of Fujiwara et al.
 and Timek et al.
 who reported that arginase-1 has no role in distinguishing well-differentiated hepatocellular carcinoma from benign hepatic lesions.
Two very recent studies examined the immunohistochemical expression of L1 cell adhesion molecule (L1CAM)
 and SOX9
 in HCC cases and their adjacent non- neoplastic liver tissues and they reported that immunoreactivity of these markers was significantly increased in substantial proportion of HCC cases compared with their adjacent non- neoplastic liver tissue. Additionally, they suggested that L1CAM expression in HCC was significantly correlated with the advanced tumor progression and was an independent poor prognostic factor for both overall survival and disease-free survival in patients with HCC. Furthermore, SOX9 overexpression in HCC tissues is of predictive value on tumor progression and poor prognosis. Moreover, Schmilovitz-Weiss et al.
 reported that squamous cellular carcinoma antigen (SCCA) is overexpressed in HCC and it is associated with tumor differentiation, cell proliferation and apoptosis. The results of their study confirm a potential association of negative SCCA expression with other markers of poor outcome in HCC.
In our study, the specificity of arginase-1 for diagnosis of HCC was higher (96%) than that of HepPar -1 (84%). Only one case of pancreatic adenocarcinoma out of 38 (2.6%) cases of MC and one of 12(8.3%) cases of CC showed positive immunoreactivity for arginase-1. However, the staining was focal and weak in these two positive cases. In contrast, HepPar-1 immunoreactivity was detected in 6 of 38 (15.8%) cases of MC (3 from colon and 3 from stomach) and in 2 of 12 (16.7%) cases of CC. Although, neither arginase-1 nor HepPar-1 immunostaining demonstrated 100% diagnostic specificity to distinguish HCC from MC in the liver and CC, our analysis of the combination of both immunomarkers among all studied tumors, raised the diagnostic specificity for HCC to 100% if both showed positive immunostainings. This high specificity of arginase-1 and HepPar-1 combination because the staining patterns of both immunomarkers in adenocarcinomas were mutually exclusive (i.e. arginase-1 - positive adenocarcinomas always lacked HepPar-1 immunoreactivity and vice versa)
These findings are in agreement with the study of Fujiwara et al.
 which showed that arginase-1 is not entirely specific for hepatic differentiation, as immunoreactivity can be identified in adenocarcinomas, particularly of pancreatic origin. The authors reported that it is not surprising to find a subset of the pancreatic adenocarcinomas included in their analysis demonstrated arginase-1 immunoreactivity. This is because a recent analysis of arginase-1 immunohistochemical expression in rats demonstrated that it was expressed at high levels in the liver and at moderate levels in the pancreas
. Moreover, Yan et al.
 found that only one case of prostatic adenocarcinoma demonstrated arginase-1 immunoreactivity. Of note, their study did not include pancreatic adenocarcinomas in their analysis. In contast, Timek et al.
 and McKnight et al.
 reported negativity of arginase-1 in all their cases of MC.
The positive immunostaining of HepPar-1 in our 6 cases of MC (3 from colon and 3 from stomach) was in concordance with the results of Yan et al.
 who detected HepPar-1 reactivity in 2 colonic adenomas, 8 colonic adenocarcinomas, 2 pulmonary adenocarcinomas, 1 chromophobe RCC, and 9 gastric adenocarcinomas (47.4% of cases). HepPar-1 immunoreactivity in gastric adenocarcinomas is reported in previous studies in which it was expressed in 47% to 83% of gastric cancers
[10, 13, 31]. Moreover, Timek et al.
 reported that the expression of HepPar-1 in nonhepatocellular tumors is well documented in the literature and they assumed that caution should be taken when using HepPar-1 to confirm a diagnosis of HCC.
In our study, out of 12 cases of CC, only one (8.3%) was positive for arginase-1, while 2 (16.7%) were positive for HepPar-1. This supports the study of Yan et al. as regards arginase -1 reactivity
. In addition, Fujiwara et al.
 reported negative immunoreactivity in all their cases for both immunomarkers. However, the positivity of HepPar-1 in our study is consistent with previous studies
[14, 22, 32]. Shiran et al.
 claimed that the presence of this occasional positivity should not be surprising considering the common progenitor cell of HCC and CC
. On the contrary, Iida et al.
 concluded that HepPar-1 was rarely but definitely expressed in hilar and peripheral intrahepatic CC, while arginase-1 was expressed at a high rate in both hilar and peripheral intrahepatic CC, irrespective of their histology. They assumed that care should be taken when using arginase-1 as a hepatocyte marker for distinguishing between a poorly differentiated hepatocellular carcinoma and a mass-forming peripheral intrahepatic CC showing the histology of poorly differentiated adenocarcinoma.
One of the important findings in the present study was that arginase-1 showed diffuse and strong nuclear reactivity along with cytoplasmic staining which was observed more in some HCC cases and their adjacent non-neoplastic cirrhotic liver tissues compared with other studied cases. It could be explained as all our HCC cases are associated with HCV. This possible explanation is supported by the findings of Cao et al.
 who reported that elevated arginase-1 staining is associated with chronic HCV infection as they found that arginase-1 expression was elevated in more than 75% of HCV infected liver samples compared to paired HCC from the same patients (> 33% positive) and to uninfected liver tissues (0% positive). The authors suggested that up-regulated expression of arginase- 1 was associated with HCV infected liver, and to a lesser extent in tumor, but not in uninfected liver. They assumed that an important part of the mechanism whereby HCV regulates hepatocellular growth and survival may be through altering arginine metabolism. However, further studies in large scale are worth-while to confirm these observations.