A concurrent primary hepatic MALT lymphoma and hepatocellular carcinoma.

Pathology (February 2015) 47(2), pp. 169–188

CORRESPONDENCE Intraductal tubulopapillary neoplasm of pancreas with associated invasive carcinoma, lymph node, rectal and hepatic metastases

Sir, Intraductal tubulopapillary neoplasm (ITPN) is a rare premalignant pancreatic tumour comprising 0.9% of all pancreatic exocrine neoplasms and 3% of intraductal neoplasms in the largest series published.1 Recognised as a specific diagnostic entity for over a decade by Japanese investigators,2 this tumour was not well recognised by Western pathologists until recently and only incorporated into the WHO classification of tumours of the digestive system in 2010 as an intraductal neoplasm distinct from other intraductal tumours, in particular, intraductal papillary mucinous neoplasms (IPMN).3 An invasive component, designated ‘ITPN with an associated invasive carcinoma’ in the 2010 WHO Classification was found in a small minority of reported cases and the malignant potential and behaviour of this tumour is not well characterised. We describe the histological, immunohistochemical and cytological features of an intraductal tubulopapillary neoplasm of pancreas with associated invasive carcinoma, and subsequent rectal and liver metastases in a 55-year-old man. The patient was previously well and presented with left sided upper abdominal pain. Endoscopic ultrasound and subsequent abdominal computed tomography (CT) scan showed a cystic lesion in the pancreatic tail. A distal pancreatectomy, splenectomy and partial gastrectomy was performed. Gross examination of the surgical specimen showed a friable, poorly fixed, white-grey tumour, 100 mm in maximum dimension, centred within the pancreas with infiltration through the wall of the adjacent stomach and into splenic parenchyma (Fig. 1A). Microscopic examination (Fig. 1B–F) showed areas typical for ITPN plus extensive associated invasive carcinoma. The intraductal component consisted of well formed, closely packed tubules and cribriform structures (Fig. 1B). Neither intra-cellular nor extra-cellular mucin was present on haematoxylin and eosin (H&E) and periodic acid-Schiff (PAS)-diastase staining. Occasional tubulopapilliform structures were identified. The tumour cells were cuboidal to columnar with rounded nuclear contours, pale chromatin and small nucleoli (Fig. 1C,D). The cytoplasm varied from palely eosinophilic to amphophilic. Patchy necrosis was present in the intraductal component and as a focal finding in the invasive component of the tumour, including within metastatic tumour in local lymph nodes. The invasive portion strongly resembled the intraductal component with a crowded tubular and cribriform architecture and infiltrated widely throughout the pancreatic tissue, into gastric muscularis propria and splenic parenchyma. Infrequent mitotic figures were present in both the intraductal and invasive components [averaging 1/10 high power fields (HPF)]. Two peripancreatic lymph nodes contained metastatic tumour deposits with a tubular and cribriform appearance. On immunohistochemical examination the neoplastic cells displayed positivity for CK7, CK19 and MUC1, while MUC5AC, MUC2, synaptophysin, chromogranin, CA19-9, CD56, CD10, nuclear b-catenin and BRAF V600E were negative. Print ISSN 0031-3025/Online ISSN 1465-3931

#

Approximately 2 years after pancreatic resection the patient complained of tenesmus. CT scanning demonstrated a 6 cm mass in the rectovesicle pouch with uptake on a positron emission tomography (PET) scan. At laparoscopy no mass was evident, colonoscopy demonstrated an extrinsic mass indenting the anterior rectal wall, and core biopsies were performed. The lesion appeared isolated and resectable and he underwent neoadjuvant chemoradiation. Subsequent abdominal CT scan demonstrated multiple low density hepatic lesions consistent with metastases. Core and fine needle aspiration biopsies of one of the liver lesions were performed. Core biopsies from the rectal mass showed a few clusters of hyperchromatic cells with cribriform architecture, strong immunostaining for CK7, negative immunostaining for CA19-9 and no mucin. Liver core biopsy showed tumour deposits consisting of variably sized tubular and cribriform structures with a small amount of fibrotic stroma (Fig. 1F). Many tubules contained neutrophils and foci of necrosis. The neoplastic cells were cuboidal to low columnar with high nucleocytoplasmic ratios, nuclear crowding and small nucleoli, closely resembling those seen in the previously resected pancreatic tumour. FNA smears of the hepatic tumour were highly cellular and contained medium-sized cells arranged in sheets, branching tubular structures and some papilliform groups (Fig. 1E). Nuclei were moderately pleomorphic, round to oval with occasional indentations, a fine chromatin pattern and small nucleoli. No mucin vacuoles were seen and the cytoplasm was pale and poorly defined. Given the confirmed irresectable metastatic disease, the rectal lesion was not removed. He was commenced on combination chemotherapy and remains working and relatively asymptomatic almost 3 years after original diagnosis. The primary tumour showed morphological features diagnostic of ITPN with the typical immunohistochemical profile of that entity. The tumour lacked intraluminal and cytoplasmic mucin, the architecture was tubular and tubulopapillary rather than papillary, and there was positive immunostaining for MUC1 while MUC2 and MUC5AC were negative, features which distinguish ITPN from IPMN.1,3 The differential diagnosis of acinar cell carcinoma was excluded by the absence of PAS-diastase positive cytoplasmic granules combined with strong diffuse CK7 and CK19 immunostaining, while the lack of staining with neuroendocrine markers ruled out a neuroendocrine tumour. Solid pseudopapillary neoplasm was excluded based on the absence of typical morphological features of that entity, such as nuclear grooves, cytoplasmic PASdiastase positive inclusions and aggregates of foamy macrophages, as well as negative immunostaining with CD10, CD56 and for nuclear b-catenin. ITPN occurs over a wide age range in adults with approximately equal sex distribution and is most commonly found in the pancreatic head; however, tumours are also reported in the pancreatic body and tail, and rarely involve the entire duct system.1,4,5 As in our case, the typical macroscopic findings are of solid, nodular masses within the dilated main pancreatic duct with or without involvement of branch ducts, the accessory pancreatic duct or common bile duct and an absence of intraductal mucin.1,2,4,6 The histological features of ITPN are well illustrated by this case. Key features include the presence of tumour nodules

2015 Royal College of Pathologists of Australasia

Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.

170

CORRESPONDENCE

Pathology (2015), 47(2), February

Fig. 1 (A) Macroscopic appearance of the pancreatic tail lesion. The tumour consists of a poorly fixed large intraductal portion (*) and adjacent solid invasive nodules of tumour (arrow). (B) The intraductal portion of the tumour (*) shows tubulopapillary architecture with rounded contours and the adjacent invasive portion consists of irregularly shaped nodules of crowded tubular structures (H&E). (C) The invasive component adjacent to a thick walled pancreatic artery retains a tubular and cribriform appearance (H&E). (D) High power view of typical intraductal tubular and cribriform arrangements composed of cuboidal and low columnar tumour cells with eosinophilic cytoplasm, raised nucleocytoplasmic ratios and mild-moderate nuclear pleomorphism (H&E). (E) Smear preparations from aspiration of a liver lesion contain crowded papilliform tissue fragments with focal tubule formation. Cells show moderate nuclear variability with round to oval nuclei, fine chromatin and small nucleoli. Cytoplasm is pale and poorly defined and no mucin vacuoles were identified (Papanicolaou stain). (F) Metastatic disease within liver showing retention of tubular and cribriform architecture (H&E).

within expanded ducts showing crowded tubular and cribriform structures, absence of cytoplasmic or extra-cellular mucin, foci of necrosis and cuboidal tumour cells with small nucleoli.1,3,7 Cystic areas are uncommon, but when present may contain tubulopapillary or papillary projections. Mitotic activity is variable but often higher than in this case with counts up to 9 per 10 HPF reported; high mitotic score and Ki-67 proliferative index are associated with invasion.1 Characteristically ITPNs have uniformly high grade dysplasia/atypia without adjacent or intermixed lower grade areas.1,8 Tumours with a component of invasive carcinoma are described in a minority of published cases and retain a tubulopapillary appearance1 which in this case is preserved between the primary invasive carcinoma and lymph node, rectal and liver metastases. Preservation of tubulopapillary and cribriform architecture and the limited

desmoplasia create difficulty in recognising invasion. Involvement of adjacent structures, such as duodenum, or stomach as in our case, is helpful if present and the presence of venous invasion is diagnostic of invasion.1,2 Foci of tumour adjacent to a pancreatic blood vessel implies extraductal spread since the pancreatic ducts do not normally accompany muscular vessels.9 Immunohistochemical profiling may be used as an adjunct to diagnosis and is summarised in a recent review paper.7 ITPNs show strong positive immunoreactivity with either cytokeratin 7 and/or 19, plus membrane positivity with MUC1 and sometimes MUC6.1 Immunopositivity for CA19-9 is present in approximately half of cases whereas CEA is negative.5 Analysis of the underlying molecular aberrations of intraductal pancreatic tumours is another potential diagnostic aid: ITPNs frequently show strong expression of phosphorylated AKT by


CORRESPONDENCE

immunohistochemistry and may contain somatic mutations in PIK3CA identifying activation of the phosphoinositide 3-kinase (PI3K)-AKT pathway.8,10 These mutations have not been described in IPMN and highlight the molecularly distinct nature of ITPN. Somatic mutations of KRAS and GNAS are seen more frequently in IPMN than ITPN.8,10 Our case showed cytological features in keeping with the two previous cases of ITPN reported from endoscopic ultrasound guided fine needle aspirate (EUS-FNA) smears, such as complex branched cellular groups with enlarged nuclei, variably sized nucleoli, granular debris within tubular lumens and the absence of mucin.11,12 While differentiation from other lesions such as IPMN may be challenging, the presence of thick mucin and well-formed papillae favour an IPMN. Analysis of cell block preparations and immunohistochemistry may be helpful in these cases. Few published reports exist of ITPN with an associated invasive carcinoma1,2 and the biological behaviour of these lesions is not clear. The present case confirms that the tubulopapillary growth pattern typical of this neoplasm is retained in the locally invasive component. Furthermore, our case also shows retention of the classical architecture and cytological features in metastatic deposits in peri-pancreatic lymph nodes, rectal wall and liver. When encountering an intra-abdominal neoplasm exhibiting a fairly monotonous tubular and tubulopapillary pattern it is important to consider the diagnosis of ITPN, particularly in the setting of a co-existing pancreatic tumour. The relatively slow progression of disease in our case is consistent with observations of other investigators. In the largest series published to date, only 3 of 10 cases of ITPN showed associated invasive carcinoma, and survival times up to 7 years following diagnosis are documented even in patients with metastatic disease.1,2 This is in marked contrast to the dismal prognosis of conventional pancreatic ductal adenocarcinoma.3 The intramural location of the rectal metastasis in our patient raises the possibility of transcoelemic spread. In conclusion, ITPN is a recently recognised diagnostic entity with characteristic morphology, immunophenotype and molecular aberrations. Identification of this rare neoplasm is important given the significantly more favourable prognosis compared to invasive pancreatic ductal adenocarcinoma, both in terms of overall survival and the relatively indolent clinical course observed in the few reported instances of ITPN with associated invasive carcinoma. However, a few patients do die of disease and, as in this case, metastatic disease may develop. The presence of molecular aberrations in these tumours involving the PI3K-AKT pathway may provide potential therapeutic targets in future. Acknowledgements: We thank Dr Tina Selinger for assistance with the images. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Yasmin Matthews1,2 Catriona McKenzie1 Christopher Byrne3 James G. Kench1,4 1

Department of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Camperdown, 2Douglass Hanly Moir Pathology, Macquaire Park, 3Department of Colorectal

171

Surgery, Royal Prince Alfred Hospital, Camperdown, and Central Clinical School, University of Sydney, Camperdown, Sydney, NSW, Australia

4

Contact Professor James Kench. E-mail: [email protected] 1. Yamaguchi H, Shimizu M, Ban S, et al. Intraductal tubulopapillary neoplasms of the pancreas distinct from pancreatic intraepithelial neoplasia and intraductal papillary mucinous neoplasms. Am J Surg Pathol 2009; 33: 1164–71. 2. Suda K, Hirai S, Matsumoto Y, et al. Variant of intraductal carcinoma (with scant mucin production) is of main pancreatic duct origin: a clinicopathological study of four patients. Am J Gastroenterol 1996; 91: 798–800. 3. Bosman FT, Carneiro F, Hruban RH, et al. WHO Classification of Tumours of the Digestive System. Lyon: IARC; 2010. 4. Jokoji R, Tsuji H, Tsujimoto M, et al. Intraductal tubulopapillary neoplasm of pancreas with stromal osseous and cartilaginous metaplasia; a case report. Pathol Int 2012; 62: 339–43. 5. Tajiri T, Tate G, Kunimura T, et al. Histologic and immunohistochemical comparison of intraductal tubular carcinoma, intraductal papillary-mucinous carcinoma, and ductal adenocarcinoma of the pancreas. Pancreas 2004; 29: 116–22. 6. Del Chiaro M, Pozzi Mucelli R, Bloomberg J, et al. Is intraductal tubulopapillary neoplasia a new entity in the spectrum of familial pancreatic cancer syndrome? Familial Cancer 2014; 13: 227–9. 7. Cooper CL, O’Toole SA, Kench JG. Classification, morphology and molecular pathology of premalignant lesions of the pancreas. Pathology 2013; 45: 286–304. 8. Yamaguchi H, Kuboki Y, Hatori T, et al. The discrete nature and distinguishing molecular features of pancreatic intraductal tubulopapillary neoplasms and intraductal papillary mucinous neoplasms of the gastric type, pyloric gland variant. J Pathol 2013; 231: 335–41. 9. Green KB, Sharma S. The pancreatic duct and its arteriovenous relationship. Am J Surg Pathol 2004; 28: 613–20. 10. Yamaguchi H, Kuboki Y, Hatori T, et al. Somatic mutations in PIK3CA and activation of AKT in intraductal tubulopapillary neoplasms of the pancreas. Am J Surg Pathol 2011; 35: 1812–7. 11. Aslan DL, Jessurun J, Gulbahce HE, et al. Endoscopic ultrasound-guided fine needle aspiration features of a pancreatic neoplasm with predominantly intraductal growth and prominent tubular cytomorphology: intraductal tubular carcinoma of the pancreas? Diagn Cytopathol 2008; 36: 833–9. 12. Guan H, Gurda G, Marie Lennon A, et al. Intraductal tubulopapillary neoplasm of the pancreas on fine needle aspiration: case report with differential diagnosis. Diagn Cytopathol 2014; 42: 156–60.

DOI: 10.1097/PAT.0000000000000228

Purpuric exanthem caused by Ross River virus infection Sir, The Ross River virus (RRV) is a species of the genus Alphavirus, of the family Togaviridae. It is endemic to Australia and New Guinea, and is the aetiological agent of the syndrome known as epidemic polyarthritis. It is the most common arbovirus infection in Australia, with national annual incidence rates of between 14 and 50 per 100,000 population.1–3 Large macropods such as wallabies are thought to be a significant vertebrate reservoir host, although other animals may also play a role, particularly in urban centres.4 Most human infections occur in people aged between 20 and 60 years, and can result in a number of symptoms, with the most common being fever, arthralgia, myalgia and rash. At least half of patients experience a debilitating polyarthritis, which may last for 6–12 months with a persisting lethargy.3 Serious complications, such as encephalitis, have been reported5 but seem to be rare, and while there is an early report of a death in


172


CORRESPONDENCE

association with viral infection,6 this is an exceedingly rare event. Approximately 50% of infected people develop a rash, which may be the sole manifestation of the disease. The rash is predominantly located on the limbs and the trunk, but may also be seen on the face, scalp or acral sites. Typically the exanthem is maculopapular in appearance, although in some cases it may be vesicular or, as demonstrated in this case, purpuric.1,4 The histological features of the RRV-associated rash are not well described, with the literature being limited to a single report in 1983.1 We have recently encountered a biopsy from a patient with proven RRV infection, and we present our histological findings herein. The patient was a 60-year-old female residing in the southwestern portion of Western Australia. This area of the country has been associated with several reported outbreaks of the disease.3 Prior to her presentation she had been in good health, and there was no history of recent travel. She developed generalised arthralgia, muscle cramps and lethargy for several months, and presented to her general practitioner after the appearance of purpuric lesions on her back, legs and palms. An indirect fluorescent antibody test (IFA) was positive for RRV IgM, allowing a presumptive diagnosis of RRV infection in the setting of an appropriate clinical presentation. Serology for other viral infections such as measles and rubella was negative. ANA and anti SSA/Ro were negative and no cryoglobulins were identified. As part of the diagnostic workup a 4 mm punch biopsy of skin was performed on an area of the rash from the left foot (Fig. 1). Microscopy revealed acral skin incorporating an area of lichenoid interface inflammatory change, with accompanying vacuolar damage to the basal layer of the epidermis. Underlying these changes was a perivascular inflammatory infiltrate within the superficial dermis, which was predominantly lymphohistiocytic in nature. There was prominent extravasation of erythrocytes into the dermis and focally into the lower reaches of the epidermis. No fibrinoid change was present around dermal blood vessels, and no haemosiderin deposition was identified, either with routine haemotoxylin and eosin or Perl’s Prussian blue reaction. There were no significant inflammatory changes in the deeper portions of the biopsy. The findings presented here represent the second report describing the histological features of an RRV-associated rash. The previous report, authored by Fraser and colleagues, described the features of two cases: one example of a typical maculopapular rash, and one example of a purpuric eruption.1

A

Both cases showed a perivascular infiltrate of leucocytes (predominantly T lymphocytes and cells of the monocytic lineage), with the additional features of extravasated erythrocytes in the purpuric example. Our observations were similar, with the additional finding in our case of a prominent lichenoid interface reaction pattern. Fraser et al. identified RRV antigen in the perivascular tissues of the maculopapular rash, but not in the purpuric example, and saw no viral particles on ultrastructural examination.1 They reasoned that the virally infected cells had been cleared, and that the subsequent extravasation of erythrocytes may reflect damage to the vessel wall via the cytotoxic inflammatory reaction, rather than lodgement of viral particles or antigen. With the exception of the three major families of DNA viruses (Poxviridae, Herpesviridae and Papovaviridae) most cutaneous viral exanthems (including those associated with the Togaviridae) are histologically non-specific. Features typically seen in these cases include a superficial and perivascular chronic inflammatory infiltrate, mild epidermal spongiosis, occasional apoptotic keratinocytes and on occasions extravasation of red blood cells.7 These changes can also be seen in other conditions, particularly drug reactions, and part of the difficulty lies in the fact that there is little information regarding the spectrum and/or severity of changes which are acceptable for a viral aetiology.8 The findings described by both Fraser et al.1 and us are indeed relatively non-specific, and it would be difficult to exclude a purpuric drug reaction on histology alone. In addition, the possibility of a condition within the spectrum of pigmented purpuric dermatoses (PPD) might also be considered. This term encompasses a number of conditions, including Schamberg disease, Majocchi purpura, Gougerot–Blum purpura, lichen aureus and eczematoid purpura of Doucas and Kapetanakis. These conditions share the histological features of superficial lymphocytic infiltration with erythrocyte extravasation and haemosiderin deposition.9 While the absence of haemosiderin may help in this differential diagnosis, it is conceivable that in early lesions of PPD this feature may be absent.8 Vasculitic processes such as Henoch–Scho¨nlein purpura have been described as being of a similar clinical appearance to the purpuric rash of RRV,1 however histologically the abscence of fibrinoid change around the walls of dermal postcapillary venules would argue against this diagnosis. In conclusion, we have described the histological features of a purpuric rash occurring in the setting of RRV infection. The findings can be summarised as a lichenoid interface inflammatory reaction pattern, with a superficial and perivascular

B

Fig. 1 (A) Low power view of the lesion, showing a superficial and perivascular inflammatory infiltrate within a fairly localised region of the dermis (H&E, scale bar ¼ 0.5 mm). (B) Higher power view showing details of the inflammatory process. The infiltrate is predominantly lymphohistiocytic in nature, with basal vacuolar degeneration consistent with a lichenoid interface reaction pattern. In addition there are numerous extravasated erythrocytes (H&E, scale bar ¼ 50 mm).


CORRESPONDENCE

lymphohistiocytic infiltrate and extravasation of erythrocytes. These changes are similar to the only other previously reported example, although the lichenoid interface reaction was not described in that case. The changes are relatively non-specific, and correlation with the clinical scenario and serological findings may be required for a definitive diagnosis. This case serves to expand the literature on the histological features of RRVinduced cutaneous manifestations, and also reminds pathologists of another potential cause for a purpuric rash in Australia. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Benjamin Allanson1 Nathan Tobias Harvey1,2 Peter John Beaton3 Benjamin Andrew Wood1,2 1

Department of Anatomical Pathology, PathWest, QEII Medical Centre, Nedlands, 2School of Pathology and Laboratory Medicine, The University of Western, Australia, Crawley, and 3John Parry Medical Centre, Narrogin, WA, Australia Contact Dr Nathan Harvey. E-mail: [email protected] 1. Fraser JR, Ratnamohan VM, Dowling JP, et al. The exanthem of Ross River virus infection: histology, location of virus antigen and nature of inflammatory infiltrate. J Clin Pathol 1983; 36: 1256–63. 2. Mylonas AD, Brown AM, Carthew TL, et al. Natural history of Ross River virus-induced epidemic polyarthritis. Med J Aust 2002; 177: 356–60. 3. Russell RC. Ross River virus: ecology and distribution. Annu Rev Entomol 2002; 47: 1–31. 4. Harley D, Sleigh A, Ritchie S. Ross River virus transmission, infection,;1; and disease: a cross-disciplinary review. Clin Microbiol Rev 2001; 14: 909–32. 5. Lucas RE, Qiao M. A case of encephalitis in central Australia due to Ross River virus? Aust NZ J Med 1999; 29: 268–70. 6. Scrimgeour EM. Suspected Ross River virus encephalitis in Papua New Guinea. Aust NZ J Med 1999; 29: 559. 7. Weedon D. Weedon’s Skin Pathology. 3rd ed London: Elsevier Health Sciences, 2009. 8. Leboit PE. What don’t we know, and when did we know that we didn’t know it? Am J Dermatopathology 2006; 28: 89–90. 9. Crowson AN, Magro CM, Zahorchak R. Atypical pigmentary purpura: a clinical, histopathologic, and genotypic study. Hum Pathol 1999; 30: 1004–12.

DOI: 10.1097/PAT.0000000000000222

Use of immunohistochemistry for SSTR2A to support a diagnosis of phosphaturic mesenchymal tumour Sir, Tumour induced osteomalacia (TIO), also known as oncogenic osteomalacia, is a rare paraneoplastic syndrome associated with tumours which secrete phosphaturic hormones, usually fibroblast growth factor 23 (FGF23), and therefore present with secondary osteomalacia.1 Historically the causative tumours of TIO were considered a diverse group, however in 2004 Folpe et al.2 proposed that the majority of cases actually represented a unique class of tumour which they termed ‘phosphaturic mesenchymal tumour mixed connective tissue variant’, now

173

commonly shortened to ‘phosphaturic mesenchymal tumour’ (PMT). Although the histological features of PMT are distinct, the morphology can be subtle and difficult to distinguish from other tumour types particularly in small biopsies. Furthermore the concept of non-phosphaturic PMTs (that is tumours which are morphologically PMT but in which there is no evidence of phosphate wasting), further clouds the differential diagnosis.2 One approach to confirming the diagnosis of PMT is to demonstrate production of FGF23 by immunohistochemistry or RT-PCR (summarised by Houang et al.3). However neither technique is widely available. Furthermore, whilst up to 94% of PMTs associated with osteomalacia and 75% of PMTs not associated with osteomalacia will be shown to produce FGF23 by these techniques,4 FGF23 mRNA has also been found in up to 44% of otherwise typical aneurysmal bone cysts and 29% of chondromyxoid fibromas not associated with osteomalacia.5 That is, FGF23 immunohistochemistry and RT-PCR appear to be highly sensitive but not specific. In 2013 Houang et al.3 reported that immunohistochemistry for the somatostatin receptor 2A (SSTR2A) was highly sensitive for the diagnosis of PMT, being positive in 15 of 15 PMTs with typical morphology and confirmed phosphate wasting. Although not specific (SSTR2A expression was found in occasional synovial sarcomas, haemangiomas, aneurysmal bone cysts and osteosarcomas), SSTR2A showed stronger and more diffuse expression than FGF23. Therefore, it was proposed that SSTR2A expression could be used to confirm a diagnosis of PMT if FGF23 immunohistochemistry or RT-PCR was unavailable or apparently negative due to focal expression. We present a case of morphologically typical PMT which was negative for FGF23 by immunohistochemistry but demonstrated positive staining for somatostatin receptor 2A (SSTR2A) which was useful to definitively confirm the diagnosis. A 12-year-old female, who was well other than for occasional migraines, presented with a painless mass on the buccal aspect of the left mandible. Imaging (CT and MRI) demonstrated an intra- and extra-osseous component with cortical breach. She was not known to be suffering from osteomalacia, and serum FGF23 and phosphate were not measured prior to surgery. The tumour underwent excision biopsy and was submitted for pathological assessment as a single 19 mm tissue fragment. The tumour was composed of plump spindled cells in a somewhat myxoid and hyalinised stroma (Fig. 1A). Of note, in areas there were prominent haemangiopericytomatous vessels (Fig. 1B) and a peculiar ‘grungy’ calcification (Fig. 1C,D), both of which have been reported as characteristic of PMT.2–4 After synovial sarcoma was excluded with negative RT-PCR studies for SYT/SSX1 and SYT/SSX2, the possibility of PMT was considered. Immunohistochemistry for FGF23 using previously published methods3 was negative. However RT-PCR for FGF23 expression was positive and, although there was some uncertainty given the relatively low specificity of RTPCR,5 a diagnosis of PMT was favoured. The patient underwent further investigation. Serum calcium and phosphate (measured 15 days after surgery) were normal. FDG-PET performed 7 days post-operatively demonstrated increased uptake at the site of surgery attributed to granulation tissue. MRI performed 12 days after surgery demonstrated mass effect and oedema at the operative site along with an 11 mm lesion thought to be either residual tumour or post-operative reactive change.


174


CORRESPONDENCE

A

B

C

D

Fig. 1 The tumour demonstrated morphological features compatible with PMT including (A) being composed of plump oval to spindled cells in a somewhat myxoid and hyalinised stroma, (B) a haemangiopericytomatous vascular pattern, and (C,D) ‘grungy’ calcification (H&E).

The patient was observed for 2 years as the residual tumour gradually enlarged on MRI, being 14 mm at 5 months and 17 mm at 18 months after surgery. Serum FGF23 determined 7 months post-operatively was 147 pg/mL (normal range 20% Little/scanty Inconspicuous Moderate Marked

AC, atypical carcinoid; HPF, high power field; SmCC, small cell neuroendocrine carcinoma; TC, typical carcinoid.

only four cases of TC tumours3–5 and few cases of SmCC have been reported in the nasopharynx. This is the first case of a unique primary AC of the nasopharynx to be reported in the English literature. Clinically, tumours in the nasopharynx can cause a progressive chronic nasal obstruction that can be unilateral or bilateral according to size and extent of the mass. In our reported case, the patient presented with a chronic bilateral asymmetrical nasal obstruction due to the size of the mass (2.8 2.2 cm). The microscopic morphology of this tumour was consistent with a NET (Fig. 2A), and on immunohistochemistry, tumour cells were positive for cytokeratin AE1/AE3, as well as for chromogranin A and synaptophysin (Fig. 2C,D), ruling out other possible malignant tumours in this location (lymphoma, nasopharyngeal carcinoma, adenocarcinoma, salivary glandtype carcinomas, etc). Table 1 shows histological characteristics of laryngeal NET according to the WHO classification.1,5,6 In our case, we had a NET with punctuate areas of necrosis, mitotic activity of 1/10 HPF and a proliferation index of 3%. The diagnosis of AC was considered. AC of the nasopharynx should be distinguished from other neuroendrocrine neoplasms such as olfactory neuroblastoma (ONB), medullary thyroid carcinoma (MTC), paraganglioma, secondary involvement of the nasopharynx by a NET, and nonneuroendocrine neoplasms such as malignant melanoma. ONB arises most commonly in the upper nasal cavity, but other ‘ectopic’ localisations may occur.9 ONB may present morphological and immunohistochemichal similarities with AC.5 However, in ONB, a variable amount of neurofibrillary stroma is usually present, and Flexner–Wintersteiner rosettes may be noted. The nuclei are small round, and vesicular, with fine ‘salt-and-pepper’ chromatin. Necrosis is uncommon, and is generally seen in poorly differentiated tumours with high mitotic counts. On immunohistochemistry, ONB cells express NSE, CD56, synaptophysin, and chromogranin. Focal immunoreactivity for cytokeratin may be found in approximately 20–25% of tumours, generally in areas of gland-like or olfactory differentiation, whereas AC are diffusely positive for epithelial markers. In our case, tumour cells were diffusely positive for cytokeratin AE1/AE3. Furthermore, ONB is characterised by the presence at the periphery of the cell nests of spindle or stellate (sustentacular) cells, highlighted by S-100 staining.9,10 These cells are absent in AC as in our case. Distinguishing MTC and AC may be more problematic, since these two tumours share many histological and immunohistochemical features.6,11 Clinicopathological correlation should establish the proper diagnosis. Serum calcitonin levels are rarely elevated in AC, which may limit the reliability of this

test in distinguishing between these two tumours. Both tumours are positive for synaptophysin, and carcinoembryonic antigen (CEA). Calcitonin is reported to be expressed in 80% of laryngeal AC but no data is found regarding nasopharyngeal AC. TTF-1 is helpful with this differential diagnosis, because intense positive staining for this marker is frequently seen in MTC, whereas AC is typically negative or only focally weakly positive.6,11 In our case, the negativity of TTF-1 and calcitonin favours the diagnosis of AC of the nasopharynx over MTC. Paragangliomas have been reported as primary tumours in the nasopharynx or as secondary lesions.5,6 Positive staining with an epithelial marker (cytokeratin AE1/AE3), as well as the absence of sustentacular cells in our case make the diagnosis of AC more plausible.5,6 Secondary involvement of the nasopharynx by a digestive (positive CDX-2) or pulmonary (positive TTF-1) NET was ruled out in our case, based on the negativity of the PET-CT scan, a normal cervico-thoracic CT scan, as well as negative TTF-1 and CDX-2 staining. Malignant melanoma may be confused with AC. However, the degree of cytological atypia and pleomorphism, generally seen in melanoma, are absent in AC. S-100, a common marker for melanoma, may be positive in laryngeal neuroendocrine carcinoma. However, melanomas are usually positive for HMB-45 and Melan A, and negative for neuroendocrine markers and cytokeratin.6 In our case, the positivity of cytokeratin and the absence of S-100 favour AC over the diagnosis of malignant melanoma. Currently, the management of NET is not well established in this location.3–5 In our case, a complete endoscopic resection of the tumour followed by adjuvant radiotherapy was sufficient to cause a complete clinical remission 16 months after the end of the treatment. In conclusion, this case highlights the need for diagnostic awareness and the use of appropriate ancillary studies in the diagnosis of a NET of the head and neck if suspected. Tumours of the nasopharynx should be submitted in toto for pathological examination, in order not to miss the presence of an aggressive small cell component. Thorough examination for small areas of punctuate necrosis, and mitosis, is necessary in order to establish an accurate diagnosis. Although NET of the nasopharynx are rare, they should be included in the WHO classification of nasopharyngeal tumours in order to be considered in the differential diagnosis. Currently, no appropriate management is established in the case of an AC of the nasopharynx. In our case, surgery followed by radiotherapy was the option adopted. Sixteen months after the end of the treatment, the patient is asymptomatic but the results should be assessed on a long-term basis.


178

CORRESPONDENCE

Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Chadi Farah1 Samah El Naderi2 Bassam Tabchy1 Claude Ghorra2 1

Otolaryngology – Head and Neck Surgery Department, and Pathology Department, Hotel Dieu de France Hospital, Saint Joseph University, School of Medicine, Beirut – Lebanon

2

Contact Chadi Farah, MD. E-mail: [email protected] 1. Kao H-L, Chang W-C, Li W-Y, et al. Head and neck large cell neuroendocrine carcinoma should be separated from atypical carcinoid on the basis of different clinical features, overall survival, and pathogenesis. Am J Surg Pathol 2012; 36: 185–92. 2. Klimstra DS, Modlin IR, Coppola D, et al. The pathologic classification of neuroendocrine tumors: a review of nomenclature, grading, and staging systems. Pancreas 2010; 39: 707–12. 3. Warman M, Halperin D, Poria Y, et al. Nasopharyngeal carcinoid tumor: Successful therapy for a tumor in a unique location. Otolaryngol Head Neck Surg 2009; 140: 437–8. 4. Vandist V, Deridder F, Waelput W, et al. A neuroendocrine tumour of the sphenoid sinus and nasopharynx: a case report. B-ENT 2010; 6: 147–51. 5. Weinreb I, Perez-Ordon˜ez B. Non-small cell neuroendocrine carcinoma of the sinonasal tract and nasopharynx. Report of 2 cases and review of the literature. Head Neck Pathol 2007; 1: 21–6. 6. Barnes L. Neuroendocrine tumours. In: Barnes L, Eveson J, Reichart P, Sidransky D, editors. World Health Organization Classification of Tumours: Pathology and Genetics of Head and Neck Tumours. Lyon: IARC Press, 2005; 135–139. 7. Lewis JS, Ferlito A, Gnepp DR, et al. Terminology and classification of neuroendocrine neoplasms of the larynx. Laryngoscope 2011; 121: 1187–93. 8. Chan J, Pilch B, Kuo T, et al. Tumours of the nasopharynx: Introduction. In: Barnes L, Eveson J, Reichart P, Sidransky D, editors. World Health Organization Classification of Tumours: Pathology and Genetics of Head and Neck Tumours. Lyon: IARC Press, 2005. 9. Wening B, Dulguerov P, Kapadia S, et al. Neuroectodermal tumours. In: Barnes L, Eveson J, Reichart P, Sidransky D, editors. World Health Organization Classification of Tumours: Pathology and Genetics of Head and Neck Tumours. Lyon: IARC Press, 2005; 65–103. 10. Prasad M, Perez-Ordon˜ez B. Nonsquamous lesions of the nasal cavity, paranasal sinuses, and nasopharynx. In: Gnepp D, editor. Diagnostic Surgical Pathology of the Head and Neck. 2nd ed. Philadelphia: Saunders, 2009;111–89. 11. Brandwein-Gensler M, Mahadevia P, Gnepp D. Nonsquamous pathologic diseases of the hypopharynx, larynx, and trachea. In: Gnepp D, editor. Diagnostic Surgical Pathology of the Head and Neck. Philadelphia: Saunders, 2009.

DOI: 10.1097/PAT.0000000000000217

A concurrent primary hepatic MALT lymphoma and hepatocellular carcinoma Sir, Primary hepatic lymphoma is uncommon and defined as an extranodal non-Hodgkin lymphoma (NHL) arising in and confined to the liver. The majority of primary hepatic lymphomas are diffuse large B-cell lymphoma.1 Although extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma) is the second most common primary hepatic lymphoma, it is an exceedingly rare disease with only 49 reported cases in the literature.2,3


Herein, an even rarer case with coexisting primary hepatic MALT lymphoma and hepatocellular carcinoma (HCC) is described. A 59-year-old man had chronic hepatitis B (HBV), diabetes mellitus and hypertension. He complained of on-and-off right upper quadrant pain for 2 years. Physical examination showed mild hepatomegaly. Multiphasic contrast computed tomography revealed a 16 mm hypodense nodule with enhancement pattern at segment 7/8, which was compatible with HCC radiologically (Fig. 1A–C). There was another 14 mm hypovascular nodule at segment 5 (Fig. 1D–F). Right hepatectomy was subsequently performed. The resected liver lobe weighed 523 g and measured 150 130 70 mm. The segment 5 mass was a 25 mm, soft, whitish lesion, whereas the segment 7/8 mass was a 25 mm, soft to firm, friable, pale yellowish lesion. They were 70 mm apart and located 22 mm and 14 mm from the resection margin, respectively. Histology of the segment 5 lesion showed diffuse sheets of dense atypical lymphoid infiltration replacing normal hepatic parenchyma and entrapping small islands of residual hepatocytes, bile ducts and reactive bile ductules (Fig. 2A–C). The atypical lymphoid cells were mainly small-sized lymphoid cells with slightly irregular nuclear contour, dense chromatin, and scanty cytoplasm. The atypical lymphoid cells were mainly B cells (immunoreactive to CD20; Fig. 2D) with aberrant expression of CD43/MT-1. These lymphoid cells were negative for CD5, CD10, CD23, cyclin D1, Bcl-2, Bcl-6 and CD56. Although light chain restriction was not shown by immunohistochemistry, clonal IgH gene rearrangement was demonstrated by polymerase chain reaction. Fluorescent in situ hybridisation revealed absence of MALT1 gene translocation (Fig. 2E). The overall features of the segment 5 lesion were most compatible with MALT lymphoma. On the other hand, the segment 8 lesion was a typical hepatocellular carcinoma (Fig. 2F). It was a solitary tumour without any satellite nodule or vascular invasion. The non-tumorous hepatic tissue in the background showed chronic hepatitis B with mild necroinflammatory activity and mild portal fibrosis. No atypical haematolymphoid infiltrate was present in the remaining liver. The subsequent bone marrow examination and whole body imaging revealed no other suspicious focus of MALT lymphoma, a diagnosis of coincidental primary hepatic MALToma and HCC was established. The patient underwent surveillance in the clinic with regular imaging, and was free of disease recurrence after 4 years of follow-up. MALT lymphoma is an extranodal low grade B-cell lymphoma and comprises 7–8% of all B-cell lymphomas. The gastrointestinal tract is the most common site of MALT lymphoma, accounting for 50% of all cases, and the stomach is the most commonly involved organ (85%) in the gastrointestinal tract.4 MALT lymphoma primarily involving liver is rare. Since Isaacson et al. first presented four cases of primary hepatic MALT lymphoma in 1995, 49 cases have been reported in the literature.2,3 Primary hepatic MALT lymphoma affects both genders equally and patients in adulthood (36–85 years) with a median age of 62 years. The majority of patients (77%) presented with a solitary mass. Although no universal consensus on treatment is available, most patients received surgical resection. The clinical outcome is favourable, as 91% of patients survived in a median follow-up period of 24 months (1 to 96 months) and none died from the disease. Our case is the second reported case with synchronous primary hepatic MALT lymphoma and HCC. This was first described by Takeshima


CORRESPONDENCE

179

Fig. 1 Multiphasic contrast computed tomography of hepatic lesions. (Upper panel) The segment 7/8 lesion showed typical enhancement pattern of hepatocellular carcinoma with arterial enhancement (A) and washout in portal venous (B) and equilibrium (C) phases. (Lower panel) The segment 5 lesion showed no definite arterial enhancement (D) and was hypoenhancing in portal venous (E) and equilibrium (F) phases.

et al.5 in 2004 in a 65-year-old woman with a solitary HBVrelated HCC and two separated nodules of primary hepatic MALT lymphoma. MALT lymphoma usually arises in background of chronic inflammation associated with infective agents, e.g., Helicobacter pylori-associated chronic gastritis, or autoimmune diseases, e.g., Sjo¨gren’s syndrome and Hashimoto’s thyroiditis.4 However, underlying risk factors of primary hepatic MALT lymphoma remain uncertain. Among patients with primary hepatic MALT lymphoma and reported viral hepatitis status, 35% (8/23) and 24% (6/24) of them had HBV and hepatitis C (HCV), respectively. Six of 50 (12%) patients had underlying primary biliary cirrhosis. Chronic inflammation of the liver caused by HBV, HCV and primary biliary cirrhosis has been hence postulated to be involved in pathogenesis of primary hepatic MALT lymphoma. In fact, various studies have shown that HBV and HCV are associated with NHL in various sites. A meta-analysis, including 22 studies with >42,000 patients, demonstrated that individuals with chronic HBV had a pooled odds ratio of 2.24 [95% confidence interval (CI) 1.80–2.78] of developing NHL.6 Another meta-analysis, involving seven studies with >11,000 patients, summarised individuals with chronic HCV had pooled odds ratios of 1.78 (95% CI 1.40–2.25) and 2.47 (95% CI 1.44–4.23) of developing NHL and marginal zone lymphoma, respectively.7 In contrast, the association between primary biliary cirrhosis and

NHL is inconclusive.8 HBV and HCV are well-established risk factors for HCC and are directly involved in the carcinogenesis of HCC but the exact role of chronic viral hepatitis in tumourigenesis of primary hepatic MALT lymphoma still requires further investigation.9 To diagnose primary hepatic MALT lymphoma, radiological and pathological examinations are crucial. Although MALT lymphoma does not exhibit distinctive radiological features,10 radiological investigation is important to exclude other primary sites of MALT lymphoma. Histology with the aid of immunohistochemistry is essential in establishing a diagnosis of MALT lymphoma. Several important differential diagnoses need to be considered during pathological examination: inflammatory pseudotumour (IPT), lymphoepithelioma-like HCC (LEL-HCC) and lymphoepithelioma-like cholangiocarcinoma (LEL-CC). Hepatic IPT is an uncommon benign tumour-like lesion. It was first reported by Pack and Baker in 1953 and less than 100 cases are subsequently described.11 Histologically, it is featured by proliferation of fibroblasts, myofibroblasts and collagen fibres in background of dense mixed inflammatory infiltrate. The inflammatory infiltrate is composed of mainly polyclonal plasma cells with variable numbers of lymphocytes, neutrophils, eosinophils and histiocytes. A subgroup of hepatic IPT belongs to the spectrum of immunoglobulin G4 (IgG4) diseases.11 Entrapped bile ducts and reactive bile ductules may be present. Spindle cell proliferation and mixed population of


180

CORRESPONDENCE


Fig. 2 (A,B) Primary hepatic MALT lymphoma featured dense lymphoid infiltration replacing normal hepatic parenchyma and entrapping small islands of residual hepatocytes, bile ducts and many reactive bile ductules. (C) Rare lymphoepithelial lesion was present and evidenced as bile duct/ductule destructed by intraepithelial lymphoid infiltrate. (D) The lymphoid infiltrate was composed of predominantly B-cells (CD20 immunohistochemistry). (E) Break-apart fluorescent in situ hybridisation for MALT1 gene. The 5’ end and 3’ end of MALT1 gene was labelled red and green, respectively. No break-apart signal of MALT1 gene was detected. (F) Typical hepatocellular carcinoma featured thickened trabeculae of atypical hepatocytes with sinusoidal vasculatures. An unpaired arteriole was present.

inflammatory cells distinguish IPT from MALT lymphoma. Entrapped islands of hepatocytes and reactive bile ductules in MALT lymphoma may mimic well-differentiated LEL-HCC and well-differentiated LEL-CC, respectively. LEL-HCC is a rare variant of HCC with only 22 cases described in the literature, and characterised by dense cytotoxic T-cell infiltration and favourable prognosis.12 LEL-CC is a rare variant of intrahepatic cholangiocarcinoma with only 23 cases reported in the literature, and frequent association (74%) with Epstein– Barr virus (EBV).13 Those EBV-associated LEL-CC are characterised by marked female predominance, favourable overall survival and distinctively frequent DNA hypermethylation. Absence of architectural and cytological atypia among hepatic and biliary components, and the presence of aberrant immunophenotype of lymphoid component, differentiate MALT lymphoma from LEL-HCC and LEL-CC. Other B-cell

lymphomas involving the liver also need to be considered in the differential diagnoses. Diffuse large B-cell lymphoma is the most common primary hepatic lymphoma and can be easily distinguished from MALT lymphoma by the presence of largesized lymphoid cells and absence of lymphoepithelial lesion.1 Primary hepatic Burkitt’s lymphoma has been reported and can be differentiated from MALT lymphoma by the presence of medium-sized B-cells with high proliferative pool (100% Ki-67 index) and expression of CD10.1 Secondary hepatic involvement by other low-grade small B-cell lymphomas (follicular lymphoma, mantle cell lymphoma and small lymphocytic lymphoma) can be distinguished from MALT lymphoma by the morphological growth pattern and immunophenotype specific to individual entities. Absence of the following characteristic immunohistochemical markers is crucial in establishing the diagnosis of MALT lymphoma: CD10 and Bcl-6 for follicular


CORRESPONDENCE

lymphoma, CD5 and cyclin D1 for mantle cell lymphoma, and CD5 and CD23 for small lymphocytic lymphoma.4 Primary hepatic MALT lymphomas are uncommon and those with synchronous HCC are exceptionally rare. The prognosis is generally good. Further studies are required to elucidate the roles of chronic viral hepatitis and primary biliary cirrhosis in pathogenesis of primary hepatic MALT lymphoma. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Ronald C. K. Chan1 C. M. Chu2 C. Chow3 Stephen L. Chan4 Anthony W. H. Chan3 1

Li KaShing Faculty of Medicine, The University of Hong Kong, Departments of 2Imaging and Interventional Radiology, 3Anatomical and Cellular Pathology, and 4 Clinical Oncology, Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong Contact Dr Anthony W. H. Chan. E-mail: [email protected] 1. Jaffe ES, Muller-Hermelink HK, Delabie J, et al. Lymphoma of the liver. In: Bosman FT, Carneiro F, Hruban RH, editors. WHO Classification of Tumours of the Digestive System. Lyon: IARC, 2010; 239–40. 2. Isaacson PG, Banks PM, Best PV, et al. Primary low-grade hepatic B-cell lymphoma of mucosa-associated lymphoid tissue (MALT)-type. Am J Surg Pathol 1995; 19: 571–5. 3. Yu YD, Kim DS, Byun GY, et al. Primary hepatic marginal zone B cell lymphoma: a case report and review of the literature. Indian J Surg 2013; 75: 331–6. 4. Isaacson PG, Chott A, Nakamura S, et al. Extranodal marginal zone lymphoma of mucosa-associated lymphoid tissue. (MALT lymphoma). In: Swerdlow SH, Campo E, Harris NL, editors. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Lyon: IARC, 2008; 214–7. 5. Takeshima F, Kunisaki M, Aritomi T, et al. Hepatic mucosa-associated lymphoid tissue lymphoma and hepatocellular carcinoma in a patient with hepatitis B virus infection. J Clin Gastroenterol 2004; 38: 823–6. 6. Dalia S, Chavez J, Castillo JJ, et al. Hepatitis B infection increases the risk of nonHodgkin lymphoma: a meta-analysis of observational studies. Leuk Res 2013; 37: 1107–15. 7. de Sanjose S, Benavente Y, Vajdic CM, et al. Hepatitis C and non-Hodgkin lymphoma among 4784 cases and 6269 controls from the International Lymphoma Epidemiology Consortium. Clin Gastroenterol Hepatol 2008; 6: 451–8. 8. Liang Y, Yang Z, Zhong R. Primary biliary cirrhosis and cancer risk: a systematic review and meta-analysis. Hepatology 2012; 56: 1409–17. 9. European Association for the Study of the Liver, European Organisation For Research and Treatment of Cancer. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2012; 56: 908–43. 10. Foschi FG, Dall’Aglio AC, Marano G, et al. Role of contrast-enhanced ultrasonography in primary hepatic lymphoma. J Ultrasound Med 2010; 29: 1353–6. 11. Leung HH, Cho CC, Chan AW. A pancreatic mass with multiple hepatic lesions. Clin Gastroenterol Hepatol 2014; 13: e3–4. 12. Shinoda M, Kadota Y, Tsujikawa H, et al. Lymphoepithelioma-like hepatocellular carcinoma: a case report and a review of the literature. World J Surg Oncol 2013; 11: 97. 13. Chan AW, Tong JH, Sung MY, et al. Epstein-Barr virus-associated lymphoepithelioma-like cholangiocarcinoma: a rare variant of intrahepatic cholangiocarcinoma with favourable outcome. Histopathology 2014; 65: 674–83.

DOI: 10.1097/PAT.0000000000000220

181

Metastatic colon adenocarcinoma to the placenta Sir, A 31-year-old woman delivered a set of dichorionic diamniotic twins after medically assisted conception with clomiphene for anovulatory cycles. This was her second pregnancy after a previous miscarriage a year before. She had an unremarkable medical and family history and was not known to be taking any long term medications. Apart from occasional constipation and abdominal discomfort, her early pregnancy was otherwise uneventful. Intrauterine growth retardation was diagnosed in the third trimester and an elective Caesarean section was performed at 35 weeks gestation. The placenta was submitted for routine histological examination as is the case with all multiple pregnancies delivered in our institution. Grossly the placental examination was unremarkable with two separate discs weighing 321 g and 248 g. The combined weight (569 g) was just below the 25th centile for twin placentas of the stated gestational age. In particular, no macroscopic lesion was identified in either disc even after repeated slicing in retrospect. Microscopic examination showed multiple solid foci of highly atypical epithelioid cells with enlarged nuclei, prominent nucleoli and ample, eosinophilic cytoplasm that was occasionally vacuolated due to intracellular mucin. There was associated patchy cellular necrosis as well as mitoses, including atypical forms. The malignant cells were seen in intervillous spaces, surrounding chorionic villi and were not found in fetal blood vessels, villous stroma, maternal bed or umbilical cord (Fig. 1A,B). Up to 75 sections of the placenta were examined and tumour cells were seen in both placental discs. Due to the unusual nature of these lesions, a broad panel of immunohistochemistry was carried out. The tumour cells were strongly and diffusely positive for AE1/AE3, CAM5.2, EMA, CDX2 and CK20 (Fig. 1C,D). There was weak and/or patchy staining with CK7, p63 and CD10. The following stains were negative: inhibin, b-hCG, S100, HMB45, Melan-A, hPL, Napsin, TTF-1, glypican-3, CD30, oestrogen receptor, mammoglobin, GCDFP-15, PAX-8, PAX-2, GATA-3, WT1, OCT4 and SALL4. The morphology and immunohistochemical staining pattern was interpreted as favouring metastatic colorectal adenocarcinoma. Subsequent clinical work-up after the placental examination including radiological imaging and endoscopies revealed a 90 mm, poorly differentiated hepatic flexure colonic adenocarcinoma with similar histological features to the placental deposits, arising in a background of tubular adenoma with high grade dysplasia (Fig. 2A,B). The tumour involved the serosa and multiple metastatic deposits in the adjacent epiploic fat were also present in the right hemicolectomy specimen. Nineteen of 20 lymph nodes were involved by metastatic adenocarcinoma and extracapsular extension was widespread. The American Joint Committee on Cancer (AJCC, 2010) pathological stage of the colorectal cancer at resection was pT4a, pN2b, pM1b (Stage 4B). A radiological metastasis screen was performed post-surgery and showed a duodenal metastatic deposit with extensive mesenteric, retroperitoneal, abdominal retrocrural and left supraclavicular lymph node involvement. A 7 mm left gluteal soft tissue metastasis was also identified on PET scan. The enteric and supraclavicular involvement were confirmed on subsequent biopsy and histopathological examination.


182


CORRESPONDENCE

A

B

C

D

Fig. 1 (A) Low power and (B) high power views of the placenta showing a focus of adenocarcinoma in the intervillous compartment. (C) Intracellular mucin in tumour cells highlighted by PAS-D. (D) The tumour cells are CK20 positive.

A

B Fig. 2 (A) Low power and (B) high power views of the adenocarcinoma in the right hemicolon.

Immunohistochemistry for mismatch repair genes (PMS2, MSH2, MLH1 and MSH6) was performed on the colon tumour and no loss of staining was identified. Our patient received chemotherapy after bowel resection and subsequently died 4 months postpartum from cancer-related causes. Our offer to perform a post-mortem examination was declined by the family. Cancer diagnosis during pregnancy is a rare event with a frequency of just over 1:1000 but is on the rise with the increase in late onset pregnancies in Western societies.1 Lee et al. has recently looked at the incidence and outcomes of pregnancyassociated cancers (i.e., cancers diagnosed during pregnancy or within 12 months of delivery) in 781,907 women who gave birth in New South Wales, Australia, between 1994 and 2008.2 The population-based cohort study identified 1798 pregnancy associated cancers in 1,309,501 pregnancies. This corresponds to an incidence rate of 137.3 per 100,000 pregnancies and more than two-thirds of these cancers were diagnosed in the 12 months post-partum. This study also found significantly increased rates of Caesarean section, planned preterm induction of labour and large for gestational age infants among the cancer related pregnancies. Colorectal cancer is the third leading cause of cancer-related death in women but only 8% of these are diagnosed in patients below 40 years of age.1,3 Among the 1798 cancers found by Lee et al. in the aforementioned study between 1994 and 2008, only 62 were colorectal type. Diagnosing colorectal cancer in pregnancy is clinically challenging as the usual symptoms such as nausea, vomiting, abdominal pain, rectal bleeding, diarrhoea, constipation and anaemia, are often attributed to the pregnancy, thereby delaying recognition and investigation. Not surprisingly, colorectal cancers diagnosed in pregnancy are often advanced in stage and the outcome is dismal.3 The incidence of metastatic involvement of the placenta by maternal disease is difficult to establish because not all


CORRESPONDENCE

placentas are submitted for histological examination. Of those examined microscopically, multiple sections and levels through the placenta may be required, as in our case, to confirm the diagnosis. It is also believed that metastases predispose to spontaneous abortions and therefore many cases may go undiagnosed. Just over 100 cases of maternal metastases to the placenta have been reported in the last century with melanomas constituting about one-third of the cases.4 Leukaemia/lymphomas, breast and lung cancers have comparable frequencies of about 13–15% each.5 Gastrointestinal cancers and soft tissue sarcomas complete the top six with 9% and 8%, respectively.1 Of the gastrointestinal cancers, the majority are of gastric origin with a few pancreatic and hepatocellular. Only one case of colorectal carcinoma metastasising to the placenta has been reported to date.6 In 1973, Rothman et al. reported a case of a 25-year-old African-American woman with advanced signet-ring cell type rectal adenocarcinoma that had widespread metastases including to the placenta, posterior vaginal wall, liver, lungs, adrenals and bone marrow. The patient died of cancer related illness two days after Caesarean section with pelvic exenteration. Her offspring was free of metastatic disease 8 months later. To the best of our knowledge, our case report is the first documentation of a right sided colon adenocarcinoma metastasising to the placenta in the English literature. The mechanism of metastatic spread to the placenta is thought to be haematological due to the histological finding of tumour cells in the intervillous spaces of the placenta mixed with maternal blood. There is some evidence suggesting that colorectal carcinomas arising in pregnancy tend to be poorly differentiated and behave more aggressively than those in nonpregnancy related settings.7 Histological examinations of placentas constitute a significant proportion of routine sign-outs in anatomical pathology but immediately ‘game-changing’ cancer diagnoses are rarely made. This had led some anatomical pathologists to overprioritise other surgical cases and biopsies at the expense of placental examinations.8 Cases such as this one justify the routine microscopic examination of placentas as well as the importance of meticulous histopathological and immunohistochemical interrogation equivalent to that applied to potentially malignant cases.9 Notable histological differential diagnosis of adenocarcinoma in the placenta includes primary intraplacental (gestational) choriocarcinoma,10 non-trophoblastic primary placental lesions such as heterotopic liver and adrenocortical tissue. Metastatic fetal tumours to the placenta, such as hepatoblastoma and neuroblastoma, should also be considered. For obstetricians and general health workers managing pregnant women, this case highlights the potential pitfall of misdiagnosing signs and symptoms of malignancy as pregnancy related complaints. All placentas delivered to mothers with known or suspected malignancies should be submitted for histological examination even if they are macroscopically normal. In one series more than half of the involved placentas were normal on gross examination.6 Cases positive for malignancy in either the placenta or fetus should be reported to the cancer registry as well as being published.11 Maternal malignancies during pregnancy can rarely metastasise to the fetus. The most common cancer to reach the fetal circulation is melanoma.12 All cases of fetal involvement reported to date had histologically confirmed placental involvement with invasion of chorionic villi by the tumour cells.1 Recommendations for management and follow up of babies

183

born to mothers with gestational malignancies have been published.13 The neonates should be screened for malignancy by clinical examination, imaging and serology as necessary. Subsequently, biannual paediatric follow-up for two years is recommended. In our case, both twins remain asymptomatic at 24 months and are being followed up with serological tumour markers monitoring. In conclusion, gestation associated malignancies are rare and in the case of colorectal cancers, difficult to diagnose due to pregnancy-related physiological changes mimicking the symptoms of neoplastic disease. Therefore, these patients have late presentations and, not surprisingly, poorer prognosis. While rare in the child-bearing population without germline predispositions, colorectal cancers in pregnancies are expected to increase in frequency and clinicians should remain vigilant in their assessment of gastrointestinal symptoms in pregnant women. Nonobstetricians should refer if they have clinical suspicion that the symptoms reported cannot simply be explained by physiological changes of pregnancy. Finally, for the anatomical pathologist this case highlights the importance of attending to placentas with the same circumspection extended to any other surgical specimens in which life-threatening malignancies may arise. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Admire Matsika1 Garfield R. O. Wright2 Maneesh N. Singh2 Jason D. Smith3 Judith F. Bligh4 1

Anatomical Pathology, Pathology Queensland and University of Queensland, Brisbane, 2Department of Obstetrics and Gynaecology, Gold Coast Hospital, 3Histopathology Department, Sullivan and Nicolaides Pathology, Tugun, and 4 Histopathology Department, Sullivan and Nicolaides Pathology, Brisbane, Qld, Australia Contact Dr Admire Matsika. E-mail: [email protected] 1. Pentheroudakis G, Pavlidis N. Cancer and Pregnancy. New York: Springer, 2003. 2. Lee YY, Roberts CL, Dobbins T, et al. Incidence and outcomes of pregnancy-associated cancer in Australia, 1994–2008: a population-based linkage study. BJOG 2012; 119: 1572–82. 3. Hoellen F, Reibke R, Hornemann K, et al. Cancer in pregnancy. Part II: treatment options of breast and other non-gynecological malignancies. Gynecol Obstet 2011; 284: 1481–94. 4. Alexander A, Samlowski WE, Grossman D, et al. Metastatic melanoma in pregnancy: risk of transplacental metastases in the infant. J Clin Oncol 2003; 21: 2179–86. 5. Tan K, Sinclair E, Angus G, Hill J, Snyman N, Russell P. Mucinous adenocarcinoma of the breast metastatic to the placenta. Pathology 2010; 42: 688–90. 6. Rothman LA, Cohen CJ, Astarloa J. Placental and fetal involvement by maternal malignancy: a report of rectal carcinoma and review of the literature. Am J Obstet Gynecol 1973; 116: 1023–34. 7. Al-Ibrahim A, Parrish J, Dunn E, Swallow C, Maxwell C. Pregnancy and maternal outcomes in women with prior or current gastrointestinal malignancies. J Obstet Gynaecol Can 2014; 36: 34–41. 8. Roberts DJ, Oliva E. Clinical significance of placental examination in perinatal medicine. J Matern Fetal Neonatal Med 2006; 19: 255–64. 9. Miller K, Zawislak A, Gannon C, Millar D, Loughrey MB. Maternal gastric adenocarcinoma with placental metastases: what is the fetal risk? Pediatr Dev Pathol 2012; 15: 237–9.


184


CORRESPONDENCE

10. Hovav Y, Almagor M, Golomb E, Beller U. Intraplacental choriocarcinoma in residual placenta 8 months postpartum. Eur J Obstet Gynecol Reprod Biol 2014; 176: 199–200. 11. Can NT, Robertson P, Zaloudek CJ, Gill RM. Cervical squamous cell carcinoma metastatic to placenta. Int J Gynecol Pathol 2013; 32: 516–9. 12. Jhaveri MB, Driscoll MS, Grant-Kels JM. Melanoma in pregnancy. Clin Obstet Gynecol 2011; 54: 537–45. 13. Voulgaris E, Pentheroudakis G, Pavlidis N. Cancer and pregnancy: a comprehensive review. Surg Oncol 2011; 20: 5.

DOI: 10.1097/PAT.0000000000000224

Oxyntic gland polyposis Sir, Polyps of the oxyntic mucosa are commonly referred to as fundic gland polyps (FGPs). These are composed of cystically dilated glands lined by fundic (parietal and chief) cells, admixed with normal glands. The overlying foveolar zone is often shortened. Distorted smooth muscle fibres may be seen around the cystic glands. Parietal cell hypertrophy and hyperplasia, often protruding into the lumen of the pits and producing a serrated profile may be observed. There is usually no evidence of inflammation or atypia in these polyps.1 Very little is known about non-cystic polyps of the oxyntic mucosa.2 This report

details the clinical findings and morphological appearances of a case of oxyntic gland polyps (OGPs), discusses the morphological differences between FGPs and OGPs and attempts to explore the possible link between these polyps to a very rare type of gastric adenocarcinoma with chief cell differentiation. The patient, a 43-year-old female, underwent sleeve gastrectomy for morbid obesity. She had an endomorphic body habitus with a body mass index (BMI) of 65. She had been on proton pump inhibitors (PPI) for 3 years. There was no personal or family history of polyposis. The resected specimen consisted of a portion of stomach which measured 240 mm along the outer curvature, 200 mm along the inner curvature and 60 mm in width. Figures 1 and 2 show the gross and microscopic features. The mucosal surface was diffusely covered with over 100 polyps ranging from 1 mm to 10 mm with the majority being approximately 5 mm (Fig. 1A and 2A). No other macroscopic changes were noted in the stomach. Microscopically, numerous polyps were recognised. These were made up of parietal and chief cells with slight attenuation of the overlying foveolar layer (Fig. 1B and 2B,D). The polyps often merged and coalesced to form larger areas of polypoid tissue (Fig. 1A,C). No nuclear pleomorphism, increased nuclear to cytoplasm ratio or mitotic figures were seen in these polyps. Minute foci of oxyntic gland proliferation and numerous areas of minute neuroendoendocrine cell hyperplasia were also seen. The latter was seen only after immunostaining for neuroendocrine markers (Fig. 3A–C). The background gastric mucosa showed features of chronic gastritis Macroscopic x1

A

Fig. 1 Main features of oxyntic gland polyposis. (A) Macroscopic appearances of the polyps. Numerous polypoid protrusions ranging from 1 mm to 10 mm are seen. Most of the polyps have indistinct outlines because they have coalesced to form large mucosal masses. (B) Three mucosal polyps each made up mostly of eosinophilic oxyntic glands and covered by a somewhat attenuated surface epithelium. Also a few lymphoid aggregates are seen in the basal part of the lamina propria. (C) Proliferation of the oxyntic gland forming a polypoid lesion.

B ABPAS

C

ABPAS x4

D

E

H&E x60

H&E x20

Fig. 2 (A) Full thickness strip of stomach. Several polyps can be seen. (B) Polyp covered with neutral mucin containing gastric foveolar epithelium as well as the base and neck of a polyp and the underlying submucosa. (C) Chronic gastritis with a focus of intestinal mucosa. (D,E) Higher magnification shows the predominance of eosinophilic parietal cells in the body of the polyp. A few slightly basophilic peptic (chief) cells can be seen in the centre (E).


CORRESPONDENCE

A

C

CD56 x2

CD56 x10

B

D

185

CD56 x4

B-catenin x20

Fig. 3 The immunohistochemical features of the process are highlighted. (A) CD56 showing positive staining at the base of the polyp and also localised strong positivity in the basal part of the lamina propria in the adjacent gastric mucosa. (B,C) Areas of brisk CD56 positive neuroendocrine cell proliferation (hyperplasia). (D) Strong cytoplasmic b-catenin positivity in the oxyntic cells and less intense staining in the overlying foveolar cells.

with focal intestinal metaplasia (Fig. 2C). No Helicobacter pylori were identified on routine sections, special stains or immunostains. b-catenin was strongly expressed in the cytoplasm of the oxyntic epithelium compared to the remaining mucosa (Fig. 3D). FGPs are one of the most common polyps to occur in the body and fundus of the stomach. They are usually incidentally found at upper gastrointestinal tract endoscopy for abdominal symptoms. They may cause nausea, vomiting and epigastric pain in patients with large or multiple polyps. FGPs are believed to develop from progressive dilatation and infolding of glandular buds to produce irregular tortuous glands and microcysts. The cysts gradually become atrophic as they dilate and are lined by cuboidal cells in which flattened parietal cells can still be identified.3,4 OGPs on the other hand have similar endoscopic appearances but are solid with no cystically dilated oxyntic mucosal gands.2 Both FGPs and OGPs occur in the fundus and upper body of the stomach. They can occur sporadically, in the setting of chronic gastritis, or associated with familial adenomatous polyposis (FAP) with or without colonic polyposis.3 – 5 However, there are morphological differences between them. FGPs are cystic whereas OGPs are solid. Sporadic and proton pump inhibitor related polyps have very weak malignant potential with very low frequency of dysplasia.4 However, dysplasia in the surface epithelium or the foveolar compartment has been described in over 40% of FGPs in the setting of FAP.6 The malignant potential of OGPs is unknown. However, a recent report by Tsukamoto and colleagues alludes to oxyntic gland polyp/adenoma as a precursor of gastric adenocarcinoma with chief cell differentiation.6 These tumours are located in areas of stomach with oxyntic mucosa (body and fundus). They present as small solitary polyps in the deep mucosa and are composed of tightly packed glands and cords of

predominantly chief cells. Nuclear pleomorphism and anisonucleosis have been noted but mitotic figures have not been seen in these tumours. Despite describing these lesions as carcinomas, none of the reported cases has resulted in true recurrence or metastasis. Singhi et al. believe these lesions to be benign and refer to them descriptively as ’oxyntic gland polyp or adenoma’.7 FGPs are considered neoplastic because of the frequent finding of genetic alteration involving the APC/b-catenin pathway with the sporadic cases showing mutations more often in the bcatenin than in the APC gene, whereas the reverse is true in the syndromic cases.2 Despite intense cytoplasmic b-catenin expression in the present case, it is unclear whether the case presented here is neoplastic or hyperplastic as a consequence of prolonged proton pump inhibitor use as neoplastic polyps of the stomach appear to show nuclear b-catenin staining with activating mutations in the b-catenin gene.8 The current case has similarities with the previously described oxyntic mucosa pseudopolyps. The latter was originally described as a presentation of atrophic autoimmune gastritis.9 In our case there was no evidence of autoimmune gastritis and no mucosal atrophy was identified. A more recent report by Celikbilek et al. describes gastric oxyntic mucosa pseudopolyposis in a patient with long-term protein pump inhibitor use and atrophic gastritis.10 We prefer the term oxyntic gland polyps (OGPs) to gastric oxyntic mucosa pseudopolyposis (GOMP) and feel that OGPs should be separated from FGPs and GOMP until further studies elucidate the exact nature of these lesions. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Ibrahim M. Zardawi1, William S. Munro3

2


186

CORRESPONDENCE

1

University of Newcastle, 2Douglass Hanly Moir Pathology, Gateshead, and 3General Surgeon, Gosford, NSW, Australia Contact Professor Ibrahim Zardawi. E-mail: [email protected]; izardawi@dhm. com.au 1. Park DY, Lauwers GY. Gastric polyps: classification and management. Arch Pathol Lab Med 2008; 132: 460–633. 2. Turner RJ, Odze RD. Oxyntic gland hyperplasia/adenoma. In: Odze RD, Goldblum JR, editors. Surgical Pathology of the GI Tract, Liver, Biliary Tract and Pancreas. 3rd ed. Philadelphia: Saunders Elsvier, 2014 . 566. 3. Tatsuta M, Okuda S, Tamura H, Taniguchi H. Polyps in the acid-secreting area of the stomach. Gastrointest Endosc 1981; 27: 145–9. 4. Odze RD, Marcial MA, Antonioli D. Gastric fundic gland polyps: a morphological study including mucin histochemistry, stereometry and MIB-1 immunohistochemistry. Hum Pathol 1996; 27: 896–903. 5. Bertoni G, Sassatelli R, Nigrisoli E, et al. Dysplastic changes in gastric fundic gland polyps of patients with familial adenomatous polyposis. Ital J Gastroenterol Hepatol 1999; 31: 192–7. 6. Tsukamoto T, Yokoi T, Maruta S, et al. Gastric adenocarcinoma with chief cell differentiation. Pathol Int 2007; 57: 517–22. 7. Singhi AD, Lazenby AJ, Montgomery EA. Gastric adenocarcinoma with chief cell differentiation: a proposal for reclassification as oxyntic gland polyp/adenoma. Am J Surg Pathol 2012; 36: 1030–5. 8. Abraham SC, Nobukawa B, Giardiello FM, Hamilton SR, Wu TT. Sporadic fundic gland polyps: common gastric polyps arising through activating mutations in the beta-catenin gene. Am J Pathol 2001; 158: 1005–10. 9. Krasinskas AM, Abraham SC, Metz DC, Furth EE. Oxyntic mucosa pseudopolyps: a presentation of atrophic autoimmune gastritis. Am J Surg Pathol 2003; 27: 236–41. 10. Celikbilek M, Deniz K, Torun E. Gastric oxyntic mucosa pseudopolyposis. Clin Gastroenterol Hepatol 2010; 8: e90.

DOI: 10.1097/PAT.0000000000000218

Hazards of a floating separator gel: a case study Sir, Gel separator tubes are widely used to separate plasma (or serum) from the cellular components of blood, allowing sample storage in the primary tube.1–3 The gel, typically a thixotropic polymer gel, has a density between that of plasma (1.026–1.031 g/cm3) and the cells (1.092–1.095 g/cm3)1–3 (although gel density varies between tubes from different manufacturers).4 ‘Floating gel’ is when the gel rises and sits at the top of the tube (i.e., above the serum or plasma component), after centrifugation. Although it is an uncommon occurrence, it poses great risks to both laboratory instrumentation and subsequently to the accuracy of patient results. Floating gel is caused by a marked increase in plasma density2,3 and has been reported in cases of sample contamination (with iodinated contrast media3,5,6 or ‘catheter locking’ anticoagulant solutions7) and in multiple myeloma patients with very high paraprotein concentrations.1,2,8 Two previous reports have described the floating gel problem and resultant analyser sampling probe occlusion,1,5 but none have detailed the effects of iodinated contrast material contamination on patient results. Here, we report the outcomes of such a case which resulted in analyser malfunction and we also describe the spurious biochemistry results obtained. A plasma sample, from a 69-year-old man who presented with acute myocardial infarction, gave a sample aspiration error


alert on our Abbott Architect analyser (attached to an Abbott Accelerator Automated Processing System; Abbott, USA). Visual inspection showed an abnormally positioned gel separator (Fig. 1). The sample had been collected in a BD Vacutainer PST II tube (heparin plasma, gel separator; BD, USA) and promptly centrifuged (at 2800 g for 10 min at 238C in a swinging-bucket centrifuge). Further enquiry revealed the sample was drawn from a radial artery catheter, during an urgent coronary angiogram and percutaneous coronary intervention procedure. During this procedure, 180 mL of Visipaque (iodixanol; GE Healthcare, USA), an iodinated contrast agent, was administered intra-arterially via the same arterial catheter (and 5 mL of blood was discarded prior to sample collection). The analyser sampling probe had entered and aspirated from the floating gel layer, contaminating several instrument components with gel, including the sampling probe, instrument mixers, cuvette and cuvette wash station. In response, the sampling probe had to be replaced and gel manually removed from the other components, resulting in 3 h of ‘downtime’ for the analyser and substantial delays in processing other specimens. To investigate the biochemical effects of the iodinated contrast media contamination, an aliquot of the plasma sample was separated and tested (Table 1). All assays were performed on an Abbott Architect analyser except for osmolality which was measured using an Advanced Instruments Model 3250 Osmometer (Advanced Instruments, USA) and sodium, which was also measured by direct ion-selective electrode (ISE), on an ABL90 Flex analyser (Radiometer, Denmark). Plasma iodine was measured in the sample on an Agilent 7700 Series instrument (Agilent Technologies, USA) and was extremely elevated at 49,470 mg/L (reference interval 0.04–0.09 mg/L), confirming iodinated contrast media contamination. The estimated density of the plasma was also found to be abnormally high at 1.047 g/cm3, as measured on a refractometer (American Optical, USA; model number 10408)

Fig. 1 (A) An image of the contaminated sample showing the floating gel, with (B) a normal sample for comparison.


CORRESPONDENCE

187

Table 1 Biochemistry results showing the contaminated sample (B), with preceding (A) and subsequent (C) results on the same patient provided for comparison

Table 2 Biochemistry results from four blood samples, collected in PST II tubes (4.5 mL draw volume) from a laboratory volunteer, and spiked with increasing volumes of Visipaque

Sample

Volume of Visipaque added (mL)

Collection time

Analyte

Reference interval

Sodium, mmol/L Potassium, mmol/L Chloride, mmol/L Urea, mmol/L Creatinine, mmol/L eGFR, mL/1.73 m2 Sodium (direct*), mmol/L Glucose, mmol/L Osmolality, mmol/kg Osmolal gap,{mmol/kg

135–145 3.5–5.2 95–110 3.2–7.7 50–110 80–120 135–145 3.5–7.7 280–300 10

* {

A

B

C

Day 1 01:55

Day 1 11:30

Day 1 13:00

138 4.0 – 7.4 115 56 – 11.7 – –

123 5.4 95 10.3 116 55 131 9.1 313 31.6

134 5.5 – 11.6 130 48 – – – –

Direct ion-selective electrode (ISE) method used. Calculated using sodium measured by direct ISE.

using a previously reported equation relating specific gravity and protein concentration.4 The details of this case illustrate the hazards associated with iodinated contrast media contaminated specimens, including severe instrument malfunction (i.e., contamination with gel) and spurious laboratory results such as pseudohyponatraemia (123 mmol/L) and increased osmolal gap (31.6 mmol/Kg). The interference observed in iodine contrast media contaminated samples has not previously been described, although potential dilutional,6 aspiration3 or other analytical errors6 have been suggested. The pseudohyponatraemia in this case could be attributed mainly to an increase in solid phase particles, as it largely corrected with direct ISE. Some dilutional effect may also have been present, based on the mildly low direct ISE sodium concentration (but additional markers of this, such as haematocrit changes, were not available). The osmolal gap was increased, suggesting contamination with unmeasured osmotically active substances, as was the osmolality (and although the manufacturer indicates Visipaque is isotonic, this effect on osmolality has previously been demonstrated).5 The high plasma density (1.047 g/cm3), due to contamination with Visipaque which has a density of 1.356–1.369 g/cm3 (GE Healthcare), is consistent with previous studies showing floating gel occurring at plasma densities of 1.038–1.045 g/cm3 in PST II tubes.2,5 Other reported analytical interferences that can occur with iodinated contrast media contamination, include falsely elevated cTnI, using the Opus Magnum cTnI assay (Behring Diagnostics) and an abnormal peak on capillary zone electrophoresis of serum proteins.9 Despite being spurious, the results in this case were otherwise biologically plausible and might have been accepted if other events (i.e., instrument error alerts) had not flagged the sample problem to laboratory staff (who subsequently noted they were significantly discordant from previous results). The biochemical findings were later confirmed with in vitro experimental data obtained by spiking blood, collected in PST II tubes from a healthy laboratory volunteer, with Visipaque (after

Floating gel observed Sodium, mmol/L Sodium (direct*), mmol/L Osmolality, mmol/kg Osmolar gap,{ mmol/kg Density, g/cm3 Urea, mmol/L Creatinine, mmol/L Glucose, mmol/L * {

0

50

100

200

No 138 141 288 6 1.029 6.4 82 5.6

No 134 139 291 0.8 1.032 6.3 78 5.9

No 132 137 292 5.9 1.034 6.1 77 6.0

Yes 124 131 299 25.6 1.040 6.1 73 5.3

Direct ion-selective electrode (ISE) method used. Calculated using sodium measured by direct ISE.

discarding an equivalent volume of blood) and then centrifuging (Table 2). The analyser damage seen in this case, was more severe than previous floating gel cases, where only occlusion of sampling probes had occurred.1,5 Although the cause of this is unclear, one suggested explanation is a defect of the analyser’s sample detection mechanism (i.e., failing to differentiate between true sample and gel), resulting in continued sampling of the gel layer and more significant contamination. In this case however, the analyser did correctly detect the floating gel, prior to performing any assay. We speculate that gel from the sample may have adhered to the sampling probe and not been removed by the sampling probe wash station. The gel remaining on the probe could then have been transferred to the cuvette and other instrument components. Iodinated contrast media contamination has previously been described in specimens collected from arterial lines in patients undergoing coronary angiography and PCI.3,6 Collecting blood from indwelling lines is suboptimal and a potential source of contamination (or dilution).10 Lines flushed with interfering agents should preferably be avoided or, if used, should be flushed with saline and sufficient volume of blood discarded, prior to collection.6,10 In this case, although 5 mL of blood was discarded prior to collection, the line was not flushed after use of the contrast agent. Furthermore, the use of arterial blood samples for analytes other than ‘blood gases’ should be discouraged, as the concentration of some analytes (e.g., glucose) differs between collection sites.11 Delaying peripheral blood collection after administering iodinated contrast agents, for at least one distribution half-life (around 20 min)5,6 or one elimination half-life (e.g., 2 h for Visipaque)5 has also previously been recommended. In this case, repeat collection after 90 min showed resolution of the floating gel and biochemical abnormalities. Visual inspection of samples after centrifugation is another way to detect floating gel samples, but is impractical in laboratories with a high degree of automation1,3,4 and may be obscured by specimen labels.1 In conclusion, although uncommon, sample contamination with iodinated contrast media is an important cause of analyser malfunction and spurious results. If biochemical results are urgently required post use of a high volume of iodinated contrast media, clinical units should be made aware to use gel-free tubes and to take steps during collection to avoid contamination from an indwelling line.


188


CORRESPONDENCE

Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. J. B. Ryan S. J. Southby L. A. Stuart C. M. Florkowski P. M. George Canterbury Health Laboratories, Christchurch, New Zealand Contact Dr Joshua Ryan E-mail: [email protected] 1. van den Ouweland JMW, Church S. High total protein impairs appropriate gel barrier formation in BD Vacutainer blood collection tubes. Clin Chem 2007; 53: 364–5. 2. Fatas M, Franquelo P, Franquelo R. Anomalous floatation of separator gel: density or viscosity. Clin Chem 2008; 54: 771–2. 3. Daves M, Lippi G, Cosio G, et al. An unusual case of a primary blood collection tube with floating separator gel. J Clin Lab Anal 2012; 26: 246 – 7.

4. Faught RC, Marshall J, Bornhorst J. Solution densities and estimated total protein contents associated with inappropriate flotation of separator gel in different blood collection tubes. Arch Pathol Lab Med 2011; 135: 1081–4. 5. Spiritus T, Zaman Z, Desmet W. Iodinated contrast media interfere with gel barrier formation in plasma and serum separator tubes. Clin Chem 2003; 49: 1187–9. 6. Kaleta EJ, Jaffe AS, Baumann NA, Block DR. A case of floating gel. Clin Chem 2012; 58: 1604–5. 7. Srivastava R, Murphy MJ, Card J, Severn A, Fraser CG. The case of the floating gel. J Clin Pathol 2004; 57: 1333–4. 8. Gerin F, Ramazan DC, Baykan O, Sirikci O, Haklar G. Abnormal gel floatation in a patient with apparent pneumonia diagnosis: a case report. Biochemia Medica 2014; 24: 180–2. 9. Lippi G, Daves M, Mattiuzzi C. Interference of medical contrast media on laboratory testing. Biochemia Medica 2014; 24: 80–8. 10. CLSI. Procedures for the collection of diagnostic blood specimens by venipuncture; Approved standard, fifth edition. Document H3-A5. Wayne (PA): CLSI, 2003; 21. 11. Young DS, Bermes EW, Haverstick DM. Specimen collection and processing. In: Burtis CA, Ashwood ER, Bruns DE, editors. Tietz Textbook of Clinical Chemistry and Molecular Diagnostics. 4th ed. St Louis: Elsevier Saunders, 2006; 41–58.

DOI: 10.1097/PAT.0000000000000219


Primary hepatic peripheral T-cell lymphoma mimicking hepatocellular carcinoma: a case report.

Primary MALT lymphoma of the larynx.

[The primary pulmonary MALT lymphoma: a rare lung tumor].

Primary plasmacytic hepatic lymphoma mimicking hepatocellular tumor on MR imaging.

Diagnostic and therapeutic quandaries in a patient with primary hepatic lymphoma and concurrent hepatitis C infection.

Report of a primary lymphoma of the conjunctiva. A lymphoma of MALT origin?

Extrahepatic metastasis of hepatocellular carcinoma arising from a hepatic adenoma without concurrent intrahepatic recurrence.

Double primary hepatic cancer (sarcomatoid carcinoma and hepatocellular carcinoma): A case report.

Tonsil neuroendocrine carcinoma concurrent with hepatocellular carcinoma: A case report.

Primary MALT Lymphoma of the Breast Treated with Definitive Radiation.

[Pulmonary malt lymphoma].

Concurrent primary vitreoretinal and spinal cord lymphoma: a unique entity.

Double primary hepatic cancer (hepatocellular carcinoma and intrahepatic cholangiocarcinoma) in a single patient: A case report.

Primary hepatic lymphoma: a case report.

Primary hepatic lymphoma: a challenging diagnosis.

Primary Hepatic Lymphoma Mimicking Cholangiocarcinoma.

Hepatic resection for hepatocellular carcinoma.

Molecular Pathogenesis of MALT Lymphoma.

Perforated ileal phytobezoar revealed a MALT lymphoma.

Hepatic iron overload and hepatocellular carcinoma.

Helicobacter pylori Strains and Gastric MALT Lymphoma.

Gastric MALT lymphoma: old and new insights.

Helicobacter pylori gastritis and gastric MALT-lymphoma.

Primary hepatocellular carcinoma.