reprinted from december 2016

pathology ◆ laboratory medicine ◆ laboratory management Isolated hepatic neuroendocrine tumor expressing albumin mRNA and arginase-1 CAP TODAY and the Association for Molecular Pathology ment. The following report comes from the University of have teamed up to bring molecular case reports to CAP Pittsburgh Medical Center. If you would like to submit a TODAY readers. AMP members write the reports using case report, please send an email to the clinical cases from their own practices AMP at [email protected]. For more informa- that show molecular testing’s important tion about the AMP and all previously role in diagnosis, prognosis, and treat- published case reports, visit www.amp.org. case report Michael A. Nalesnik, MD the time of operation it was felt to lular markers. Arginase-1 (Fig. D, Aatur D. Singhi, MD, PhD represent a hepatocellular . page 2) showed patchy uptake in Liver has historically been Pathologic examination showed a approximately 10 percent of cells subdivided into hepatocellular carci- well-differentiated epithelial neo- with large areas devoid of stain, noma and . plasm with a regular appearance and whereas immunostains for alpha- Combined types are recognized, sug- an “organoid” architecture (Fig. A, fetoprotein, HepPar-1 (carbamoyl gesting a common cell of origin, and page 2). Given this typical appear- phosphate synthetase 1) (Fig. E, page indeed a spectrum of such tumors ance for neuroendocrine tumor, tar- 2), polyclonal CEA, and CD10 were has been proposed.1 In contrast, pri- geted immunohistochemical stains all negative. Glutamine synthetase, mary neuroendocrine tumor of the were performed and showed diffuse an antigen characteristic of but not liver is an extremely rare positivity (Fig. B, specific for hepatocytes, was ex- that can be diagnosed only following page 2), with trace PGP9.5 and no pressed in about half of the tumor an extensive search for extrahepatic chromogranin uptake. A diagnosis of cells. After diligent search, a single primary sites, since no markers spe- low-grade neuroendocrine tumor on small bile plug was found within the cific for liver origin of these tumors the basis of low mitotic count was tumor (Fig. F, page 2). exist.2 Neuroendocrine tumor is not made, and the patient continues to Discussion. The advent of immu- even listed as a primary liver tumor do well following surgical resection nohistochemistry and, more recently, type by the World Health Organiza- with no evidence of tumor within or molecular diagnostics has disrupted tion,3 although some authors recog- outside the liver to date. morphology-based tumor taxonomy nize a rare neuroendocrine variant of Tumor studies. Approximately one by demonstrating common features cholangiocarcinoma.4 We present a year after surgery, tumor sections among tumor types previously con- case in which in situ detection of al- were randomly selected as negative sidered unrelated. In this case, a tu- bumin mRNA strongly suggested control for a validation study mor diagnosed as neuroendocrine primary hepatocellular origin of a of albumin RNA ISH assay using the tumor by an experienced pathologist liver-based neuroendocrine tumor. automated Affymetrix branch chain and supported by immunohisto- Case. A middle-aged woman under- RNA ISH technology. Surprisingly, chemical stains was serendipitously went examination for an the tumor showed albumin mRNA found to have features associated unrelated condition and was found to expression that surpassed some of with hepatocellular origin using a have a 9.5-cm liver mass. No other the positive samples (hepatocellular molecular-based approach. lesions were identified. The back- carcinoma or normal liver) (Fig. C, In situ hybridization for albumin ground liver was unremarkable and page 2). This prompted reevaluation mRNA has long been recognized as the tumor was resected surgically. At of the lesion for additional hepatocel- a marker for benign and malignant 2 | CAP TODAY

A B reported positivity in intrahepatic, but not extrahepatic, cholangiocarci- noma, suggesting a shared feature that may reflect a common cell of origin. Terris, et al.6 recently chal- lenged this interpretation, maintain- ing that such positivity may more likely reflect combined hepatocellu- lar-cholangiocarcinoma with stem cell features. Nevertheless, both viewpoints support the concept of plasticity of an intrahepatic tumor progenitor cell able to differentiate in C D either a hepatocellular or biliary direction. This case, although a single re- port, raises the possibility that this plasticity may extend to neuroendo- crine differentiation as well. Primary neuroendocrine tumor of the liver is not recognized in the current WHO classification, although neuroendo- crine tumors arising in the gallblad- der are acknowledged. Identification of neuroendocrine tumors showing specific evidence of primary hepatic origin, if supported by more exten- E F sive studies, would not only be of biologic interest but also have direct clinical implications since a diagno- sis of primary liver neuroendocrine tumor would preclude extensive radiologic and perhaps surgical ex- ploration in search of a primary tu- mor elsewhere. Retrospective screen- ing of intrahepatic neuroendocrine tumors for albumin mRNA and ar- ginase-1 in specimens from patients with no recognized extrahepatic tu- A. The intrahepatic tumor (left) was composed of monotonous small epithelioid cells arranged as nests and mors is currently underway in our sheets with an “organoid” appearance. Background hepatic parenchyma is seen on the right (H&E×10). laboratory. B. Uniform synaptophysin uptake (lower half) was present within tumor cells (synaptophysin immunoperoxidase ×20). C. Detection of albumin mRNA by branch chain in situ hybridization showed uniform positivity within tumor cells with areas of hyperintense uptake (albumin branch chain ISH alkaline phosphatase ×20). D. Arginase was 1. Shafizadeh N, Kakar S. Hepatocellular car- present with varying intensity in scattered tumor cells (arginase immunoperoxidase ×20). E. HepPar-1 expression cinoma: histologic subtypes. Surg Pathol was uniformly absent in tumor cells. Normal liver uptake is seen on the left (HepPar-1 immunoperoxidase ×20). Clin. 2013;6(2):367–384. F. A single bile plug (arrow) was detected within the tumor (H&E×40). 2. Park CH, Chung JW, Jang SJ, et al. Clinical features and outcomes of primary hepatic hepatocytes. The branch chain RNA phosphatase labeled detector probes, neuroendocrine . J Gastroen- ISH technology uses multiple tiling leading to a several hundred-fold terol Hepatol. 2012;27(8):1306–1311. probes that serve as the foundation amplification process. Shahid, et al.5 3. Bosman FT, Carneiro F, Hruban RH, for hybridization to pre-amplifier found this technique to be more sen- Theise ND, eds. WHO Classification of molecules. These in turn hybridize to sitive than either HepPar-1 or argi- Tumours of the Digestive System; vol 3. 4th ed. Lyon, France: IARC Press; 2010. multiple amplifier molecules, each of nase staining for detecting hepatocel- which attaches to multiple alkaline lular carcinoma. These authors also 4. Nakanuma Y, Sato Y, Harada K, Sasaki CAP TODAY | 3

M, Xu J, Ikeda H. Pathological classifica- chain in situ hybridization for albumin tion of intrahepatic cholangiocarcinoma mRNA? based on a new concept. World J Hepatol. Test yourself a) It has limited utility because of the 2010;2(12):419–427. Here are three questions taken from the case widespread presence of albumin in serum 5. Shahid M, Mubeen A, Tse J, et al. Branched report. Answers are online now at www.amp. and interstitial tissue. b) It requires fresh frozen tissue as a chain in situ hybridization for albumin as org/casereviews and will be published next a marker of hepatocellular differentiation: substrate. month in CAP TODAY. evaluation of manual and automated in c) It has high sensitivity for mRNA detection situ hybridization platforms. Am J Surg 1. Which of the following is true regard- in routinely fixed paraffin-embedded tissue Pathol. 2015;39(1):25–34. ing primary hepatic neuroendocrine­ due to probe tiling and signal amplification. d) Fluorescence microscopy is required for 6. Terris B, Hergli I, Lin-Marq N, Rubbia- tumors? a) They represent the third most common analysis. Brandt L. Letter to the editor with regard e) The results must be confirmed by protein to article titled: “Branched chain in situ type of primary liver cancer. b) Features of neuroendocrine differentiation immunohistochemistry due to mRNA cross- hybridization for albumin as a marker of reactivity with pseudo-albumin. hepatocellular differentiation.” Am J Surg are frequently found in both hepatocellular carcinomas and . Pathol. 2015;39(8):1156–1157. 3. Detection of albumin mRNA in a docu- c) They have a tendency to metastasize to mented neuroendocrine tumor restricted endocrine organs. to the liver is suggestive of: Dr. Nalesnik is a professor of pathology, d) A neuroendocrine tumor presenting in the a) contamination by background hepatocytes. Division of Transplantation Pathology, liver will prompt an extensive search for a b) contamination by extracellular albumin. and Dr. Singhi is an assistant professor of primary tumor site. c) the presence of a housekeeping gene e) They are known to arise from neural pathology, Division of Anatomic Pathol- common to all neuroendocrine tumors. crest cells that migrate to the liver during ogy, both in the Department of Pathology, d) an incorrect diagnosis. embryogenesis. University of Pittsburgh Medical Center. e) multipotential plasticity of a primary hepatic 2. From the case report, which of the fol- tumor stem/progenitor cell. lowing may be inferred regarding branch