WWTR1 translocation

Pathology (June 2014) 46(4), pp. 348–372

CORRESPONDENCE Focal nodular hyperplasia and hepatocellular carcinoma: uncommon companions? Sir, Focal nodular hyperplasia (FNH) is a benign condition of the liver characterised by nodules of hyperplastic hepatocytes divided by fibrous septa radiating from a fibrous scar. The fibrous scar contains abnormal vessels and FNH are thought to develop from a hypertrophic response of hepatocytes to an increase in local blood flow.1 It is widely accepted that FNH has an indolent natural history with minimal risk of haemorrhage or rupture and no acknowledged potential for malignant transformation.2 However our own experience with two patients with hepatocellular carcinoma (HCC) arising within an FNH, and a total of 20 reported patients with hepatocellular carcinoma occurring in close association with FNH3–18 (Table 1), indicates that FNH and HCC can occur in the same hepatic lesion. A 34-year-old female was noted to have abnormal liver function tests. Clinical examination was normal. An abdominal ultrasound demonstrated a large mass in the left side of the liver and this was further investigated with an abdominal CT scan. Arterial phase images showed a hypervascular mass in segments 2 and 3 without a central scar and with partial washout in the portal venous phase. Other investigations demonstrated a normal full blood count, plasma urea and electrolytes. Hepatitis B and C serology was negative and tumour markers including a-fetoprotein were within the normal range. Her liver function tests were mildly deranged with an alkaline phosphatase (ALP) of 142 U/L (normal range 40–100U/L) and a gamma glutamyl transferase (gGT) of 282 U/L (normal range <50 U/L). A diagnosis of a large adenoma or possible fibrolamellar carcinoma was made and a left hepatectomy was Table 1

performed without complication. Her subsequent clinical course was uneventful. The patient is alive and well 5 years post-resection. Pathological analysis showed a lesion with the histological appearances of FNH with a small HCC contained within the substance of the FNH (Fig. 1A). The surrounding FNH showed numerous large, unpaired, abnormal blood vessels consistent with a large vascular abnormality, such as a haemangioma. Further staining with a specific marker for glutamine synthetase demonstrated perivenular uptake in normal liver (Fig. 1B) with a more widespread geographic distribution in FNH (Fig. 1C) and diffuse intense uptake with the HCC (Fig. 1D). Glutamine synthetase staining also highlighted an area of large cell change adjacent to the HCC (Fig. 1E). A 45-year-old Caucasian male was referred with an abdominal mass. This had been incidentally noted during routine assessment of an incisional hernia. At age of 12 years the patient had been found to have portal hypertension and a mesocaval shunt was performed. At assessment the patient denied any medication use and drank less than 1 unit of alcohol per week although admitted to heavy alcohol intake 10 years previously. Examination was unremarkable except for a palpable mass in the epigastrium. Investigations showed a normal full blood count, urea and electrolytes, and an elevated a-fetoprotein of 23170 mg/L (normal range <10 mg/L). Liver function tests showed an elevated ALP of 455 U/L, gGT of 910 U/L, alanine transferase 147 U/L (normal range 0–45 U/L), and aspartate transferase of 90 U/L (normal range 0–45 U/L). A magnetic resonance scan showed a large hypervascular mass involving segments 2 and 3 and a diagnosis of hepatocellular carcinoma was made. A left hepatectomy was performed with the procedure and recovery uncomplicated and the patient discharged on day 5. The a-fetoprotein decreased to <5 mg/L

Summary of reported patients with hepatocellular carcinoma (HCC) and focal nodular hyperplasia (FNH)

Reference

Age

Saul et al. 198718

19 17

Davidson et al. 1990 Saxena et al. 199415 Muguti et al. 199216 Chen et al. 200114 Coopersmith et al. 200213 Cucchetti et al. 200312 Zhang et al. 200411 Imkie et al. 200510 9

Kataoka et al. 2006

Langrehr et al. 20058 7

Petsas et al. 2006 Sotiropoulos et al. 20086 Morise et al. 20095 4

Haubert et al. 2010 Scheuermann et al. 20123

Sex

Risk factors HCC

Cirrhosis

aFP

F

Nil

Nil

Normal

16 M 14 F 3 patients, 2 65 F 43 F 55 56

F M

27 45 64

Tumour diameter, cm

Histotype HCC

Histotype FNH

FNH 9 cm; HCC Fibrolamellar NS smaller Nil Nil Normal FNH NS; HCC 6 cm Fibrolamellar NS Nil Nil NS NS Fibrolamellar NS ordinary HCC, 1 fibrolamellar. Age, sex, risk factors, tumour site not stated. aFP normal in all 3 patients Nil Nil Raised 20 cm tumour Ordinary NS Nil Nil Raised 7 cm tumour Ordinary NS; Diffuse

Lesions adjacent Yes Yes Yes Yes Yes

Nil Present

Normal Raised

7 cm tumour FNH 3 cm; HCC 2 cm

Ordinary Ordinary

NS NS

No Yes

F F M

Nil Hepatitis B carrier Nil Nil Nil

Nil Nil Nil

Normal Normal NS

FNH 20 cm; HCC 6 cm FNH 14 cm; HCC NS NS

F F F F

Nil Nil Nil Nil

Nil Nil Nil Nil

NS NS Normal Normal

FNH FNH FNH FNH

NS NS Regenerative hyperplasia NS NS NS NS

Yes Yes Yes

46 50 23 31

Fibrolamellar Fibrolamellar Ordinary, also lymphoma Ordinary Ordinary Ordinary Ordinary

59

M

Nil

Raised

FNH 2 cm; HCC 0.9 cm

Ordinary

NS

Yes

86 38

F F

Hepatitis B & C carrier Nil Nil

Nil Nil

Normal Normal

NS NS

Ordinary Fibrolamellar

NS NS

Yes Yes, absent portal vein

14 cm; HCC 5 cm 12 cm; HCC 3 cm 4.5 cm; HCC 1 cm 15 cm; HCC 7 cm

Yes Yes Yes Yes

aFP, a fetoprotein; NS, not stated.

Print ISSN 0031-3025/Online ISSN 1465-3931

#

2014 Royal College of Pathologists of Australasia

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

CORRESPONDENCE

349

Fig. 1 (A) H&E stained tissue from Patient 1 demonstrating multiple abnormal vascular channels (arrow), hyperplastic hepatic tissue (FNH) and a small HCC (HCC). (B) Glutamine synthetase staining in normal liver demonstrating uptake in perivenular hepatocytes (arrow). (C) Glutamine synthetase staining in hyperplastic tissue showing more widespread uptake (arrows). (D) Glutamine synthetase staining in HCC demonstrating intense diffuse staining (single arrow) and an adjacent area of dysplastic change with increased, diffuse staining compared to normal (double arrows). (E) Detail of the area of dysplastic change demonstrating multiple atypical nuclei (arrows).

1 month post-resection and has remained normal. The patient is alive and well with no signs of recurrent disease at 4 years postresection. Pathological analysis showed a large, lobulated lesion, which occupied most of the resected liver and contained a pale stellate scar. About one-third of the lesion was occupied by a cholestatic and haemorrhagic lesion which also bore a pale nodule (Fig. 2). Histology confirmed a large FNH associated with HCC and a similar immunohistochemical staining profile to Patient 1. In 1980 Berman et al.19 suggested that FNH was a precursor lesion to fibrolamellar HCC in the same way that hepatic adenomas can undergo malignant transformation into nonfibrolamellar HCC. Subsequent clonal analysis of FNH has shown that these lesions are largely polyclonal suggesting a hyperplastic rather than a neoplastic growth pattern while HCC are monoclonal consistent with malignant transformation.20 However, there are a number of reports in the literature of HCC arising in close association to FNH indicating that these lesions can co-exist in the same patient (Table 1). Co-existence of HCC and FNH is rare. Only 22 instances have been reported although there is a high incidence of HCC

Fig. 2 Macroscopic specimen from Patient 2 demonstrating a large nodular lesion with a stellate scar (thin arrows) with the histological features of focal nodular hyperplasia and the hepatocellular carcinoma arising in this (thick arrows). This shows haemorrhage and cholestasis.

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

350

Pathology (2014), 46(4), June

CORRESPONDENCE

and prevalence of FNH worldwide. Review of the reported patients highlights a number of trends. Of the 19 patients in which demographic information is available, 14 were female with a median age of 45 years (range 14–86 years). Only two patients had known risk factors for HCC5,11 with the case of Zhang et al. being the only patient with signs of cirrhosis.11 Serum a-fetoprotein was raised in four patients with nonfibrolamellar HCC5,11,13,14 although serum concentrations of a-fetoprotein were not raised in four of the remaining nonfibrolamellar HCC and not stated in the other three patients (Table 1). Seven of the 22 patients with HCC in association with FNH were fibrolamellar HCC while the remaining 15 HCC were all non-fibrolamellar. The histotype of FNH was not specified in any patient other than our own where it was classical. In the patients reported by Muguti et al.,16 the proximity of the FNH and HCC are not specified but in all other patients the lesions lie in close proximity (Table 1). In those patients in whom tumour size was specified the FNH tended to be large with the HCC smaller, possibly indicating that the FNH had been present for longer and the HCC had developed later. Although, in the patients of Zhang et al.11 and Morise et al.5 the FNHs were small (3 cm and 2 cm diameter, respectively) the majority of the other reported FNH were larger, varying between 4.5 cm diameter7 and 20 cm diameter.10 Of note, the HCC that occurred in association with the two smallest FNH5,11 also developed in the two patients with known risk factors for HCC. The possible reasons for the presence of HCC and FNH in the same specimen are not clear. FNH are common lesions and may simply occur incidentally in a proportion patients undergoing hepatectomy for HCC. It has been suggested that FNH develop in response to the hypervascularity accompanying a HCC.9,11,15 However Morise et al.5 have shown the development of an HCC within an FNH on sequential imaging, and our cases as well as others have demonstrated HCC lying within FNH, indicating that the development of HCC occurs after the growth of FNH. The presence of an additional vascular abnormality may also be significant. In both our patients the FNH was associated with a vascular anomaly (haemangioma in Case 1, mesocaval shunt with extrahepatic portal hypertension in Case 2) as was the case of Scheuermann et al.3 where congenital absence of the portal vein was noted. The association of congenital venovenous malformations, either extrahepatic (Abernethy malformation)21 or intrahepatic (Park malformation),22 with the development of FNH and HCC is now recognised and indicates that these a malformations are accompanied by proliferative effects on hepatocytes. Finally, the finding of large cell change adjacent to the HCC in Case 1 of the current report indicates that, at times, conditions within FNH may favour hepatocyte proliferation and abnormal growth. The role of large cell change in the pathogenesis of HCC is poorly understood, with the phenomenon often regarded as a degenerative feature. However, the demonstration of glutamine synthetase in Case 1, which is intermediate in distribution and intensity of staining between normal liver and HCC, would suggest that large cell change may be a precursor lesion to HCC. Clinically these observations emphasise that all FNH should be investigated and the diagnosis confirmed radiologically or with biopsy if considered necessary. FNH in patients with risk factors for HCC such as cirrhosis or venous malformations should be followed and FNH that are behaving in an atypical manner should be considered for resection.

Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Jonathan B. Koea1 Mee Ling Yeong2 1

The Department of Surgery, North Shore Hospital, Takapuna, Auckland, and 2Diagnostic Medlab, Ellerslie, Auckland, New Zealand Contact Associate Professor J. B. Koea. E-mail: [email protected] 1. Wanless IR, Mawdsley C, Adams R. On the pathogenesis of focal nodular hyperplasia of the liver. Hepatology 1985; 5: 1194–200. 2. Nahm C, Ng K, Lockie P, et al. Focal nodular hyperplasia – a review of myths and truths. J Gastrointestinal Surg 2011; 15: 2275–83. 3. Scheuermann U, Fotys D, Otto G. Focal nodular hyperplasia proceeds hepatocellular carcinoma in an adult with congenital absence of the portal vein. Transpl Int 2012; 25: e67–8. 4. Haubert L, Yearsley M, Bloomston M. Hepatocellular carcinoma arising within focal nodular hyperplasia. Am Surg 2010; 76: 335–6. 5. Morise Z, Sugioka A, Mizoguchi Y, et al. Stepwise carcinogenesis of hepatocellular carcinoma in a nodule surrounded by hyperplastic and hypervascular liver tissue. Can J Surg 2009; 52: E5–7. 6. Sotiropoulos GC, Bockhorn M, Molmenti EP, et al. Hepatocellular carcinoma as a coincidental finding in a patient undergoing surgery for focal nodular hyperplasia. Liver Int 2008; 28: 578–9. 7. Petsas T, Tsamandas A, Tsota I, et al. A case of hepatocellular carcinoma arising with large focal nodular hyperplasia with review of the literature. World J Gastroenterol 2006; 12: 6567–71. 8. Langrehr JM, Pfitzmann R, Hermann M, et al. Hepatocellular carcinoma in association with focal nodular hyperplasia. Acta Radiol 2006; 47: 340–4. 9. Kataoka TR, Tsukamoto Y, Kanazawa N, et al. Concomitant hepatocellular carcinoma and non-Hodgkins lymphoma in a patient with nodular regenerative hyperplasia. Pathology Int 2006; 56: 279–82. 10. Imkie M, Myers SA, Li Y, et al. Fibrolamellar hepatocellular carcinoma arising in a background of focal nodular hyperplasia. J Reprod Med 2005; 50: 633–7. 11. Zhang S-H, Cong W-M, Wu M-C. Focal nodular hyperplasia with concomitant hepatocellular carcinoma: a case report and clonal analysis. J Clin Pathol 2004; 57: 556–9. 12. Cucchetti A, Vivarelli M, De Ruvo N, et al. Simultanoeus presence of focal nodular hyperplasia and hepatocellular carcinoma: Case report and review of the literature. Tumori 2003; 89: 434–6. 13. Coopersmith CM, Lowell JA, Hassan A, et al. Hepatocellular carcinoma in a patient with focal nodular hyperplasia. HPB 2002; 4: 135–8. 14. Chen TC, Chou TB, Ng KF, et al. Hepatocellular carcinoma associated with focal nodular hyperplasia. Report of a case with clonal analysis. Virchows Arch 2001; 438: 408–11. 15. Saxena R, Humphreys S, Williams R, et al. Nodular hyperplasia surrounding fibrolamellar carcinoma: a zone of arterialized liver parenchyma. Histopathology 1994; 25: 275–8. 16. Muguti G, Tait N, Richardson A, et al. Hepatic focal nodular hyperplasia: A benign incidentaloma or a marker of serious hepatic disease. HPB 1992; 5: 171–80; 1992. 17. Davidson BR, Varsamidakis N, Scheuer P, et al. Fibrolamellar carcinoma arising in an abnormal liver. J Hepatol 1990; 11: 388. 18. Saul SH, Titelbaum DS, Gansler TS, et al. The fibrolamellar variant of hepatocellular carcinoma. Its association with focal nodular hyperplasia. Cancer 1987; 60: 3049–55. 19. Berman MM, Libbey NP, Foster JH. Hepatocellular carcinoma. Polygonal cell type with fibrous stroma - An atypical variant with favorable prognosis. Cancer 1980; 46: 1448–55. 20. Cai Y-R, Gong L, Teng X-Y, et al. Clonality and allelotype analysis of focal nodular hyperplasia compared with hepatocellular adenoma and carcinoma. World J Gastroenterol 2009; 15: 4695–708. 21. Lisovsky M, Konstas AA, Misdraji J. Congenital extrahepatic portosystemic shunts (Abernethy malformation): A histopathologic evaluation. Am J Surg Pathol 2011; 35: 1381–90. 22. Witters P, Maleux G, George C, et al. Congenital veno-venous malformations of the liver: Widely variable clinical presentations. J Gastroenterol Hepatol 2007; 23: E390–4.

DOI: 10.1097/PAT.0000000000000102

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

CORRESPONDENCE

351

Transformation of endometrioid carcinoma to carcinoma with trophoblastic differentiation: clinicopathological and whole genomic study Sir, Endometrioid carcinoma with trophoblastic differentiation is a rare entity with only a small number of previous cases reported.1 These gynaecological carcinomas secrete b-hCG, an observation that has also been found in rare tumours from diverse extragenital organs.2 Endometrial carcinomas with trophoblastic differentiation often follow an aggressive course with early metastatic spread and poor clinical prognosis; however, it is premature to suggest clinical outcomes due to the small number of cases described.1 We report a case of endometrioid carcinoma transformation to carcinoma with trophoblastic differentiation. A recent study suggested that there is clonal evolution from endometrioid carcinoma to trophoblastic tumour;1 therefore, whole genomic copy number analysis was performed in our case to investigate whether there was a clonal relationship between the trophoblastic and endometrioid components of the tumour. Consent for this study was obtained from the patient. The patient was a 54-year-old G4, P4, premenopausal woman who presented with a 1 year history of persistent vaginal bleeding. Pipelle biopsy of the endometrium showed FIGO 1 endometrioid carcinoma. A previous dilatation and curettage performed at 49 years for abnormal per vaginal bleeding had shown no evidence of malignancy or hyperplasia. Preoperative serum b-hCG (requested fortuitously) was 2300 IU/L (normal range 0 IU/L) and the CA125 was 88 IU/L (normal range <35 IU/L). At surgery, a large uterine tumour directly invaded the sigmoid colon. No other signs of tumour spread were found. A total hysterectomy and bilateral salpingo-oophorectomy with en-bloc resection of the sigmoid colon and protective loop colostomy were performed. The patient had an uneventful recovery from surgery. Macroscopically, the tumour measured 85
40 mm, filled the uterine cavity and invaded through the full thickness of the myometrium on the left side where it was adherent to the sigmoid colon. Microscopically, the tumour had three components: (i) the majority (70%) was grade 2 endometrioid carcinoma with squamous and mucinous metaplasia (Fig. 1A), (ii) approximately 30% was solid carcinoma with scattered multinucleated cells and confluent necrosis (Fig. 2A). The multinucleated cells tended to form incomplete rings around areas of confluent necrosis, (iii) the least common pattern (10%) was clear cell carcinoma. Immunoperoxidase for b-HCG showed only rare single positive cells in the grade 2 endometrioid and clear cell components (Fig. 1B), but confluent staining in the solid carcinoma where the multinucleated cells showed stronger staining than the mononucleate cells (Fig. 2B). The final histological diagnosis was combined grade 2 endometrioid carcinoma with trophoblastic differentiation and clear cell carcinoma. One week post-operatively, the b-hCG had dropped to 292 IU/L; however, 3 weeks later a positron emission tomography (PET) scan showed bilateral lung metastases and the serum b-hCG had risen to greater than 600 IU/L. It was hypothesised that the trophoblastic component of her disease had metastasised. She was commenced on EMA/CO with weekly monitoring of b-hCG. The b-hCG dropped again

Fig. 1 First area chosen for whole genomic study: FIGO grade 2 endometrioid carcinoma with squamous metaplasia arrowed (A, H&E) and b-hCG with positive cell arrowed (B, immunoperoxidase).

but it increased to 794 IU/L and a new metastatic lung nodule was found on computed tomography (CT). Her chemotherapy regime was changed to EP/EMA, and cisplatin was changed to carboplatin after 2 months as renal function declined. Chemotherapy was ceased due to poor toleration. She developed brain metastases and died 15 months after diagnosis. To further investigate the relationship between the endometrioid carcinoma and the region of carcinoma with trophoblastic differentiation, a whole genomic study was performed on two tumour samples, (Fig. 1 and 2), and one normal tissue sample by whole genome copy number analysis using Illumina CytoSNP Infinium HD FFPE assay (Illumina, USA). The two tumour components were distinct on routine histology. The tumours and normal tissue were sampled using a 2 mm punch biopsy of the block after areas of highest cell purity for each tissue type were identified on sections. Areas were chosen where there was minimal mixture of tumour cell types. The punch biopsies were deparaffinised, rehydrated, and digested overnight in Levi lysis buffer (Levi lysis buffer plus 0.5% Tween 20 and 0.45 mg/mL proteinase K). The lysate was extracted and ethanol precipitated to obtain a high DNA yield. Samples were processed according to manufacturer’s instructions and the data were analysed using GenomeStudio V2011.1 and Nexus 6.1 (Biodiscovery, USA) as previously described.3,4 The copy number variation profile of the normal tissue showed gains and losses across the genome which is consistent with our previous studies3,4 and the database of CNVs found in normal

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

352

CORRESPONDENCE

Pathology (2014), 46(4), June

Fig. 2 Second area chosen for whole genomic study: carcinoma with trophoblastic proliferation. H&E with syncytiotrophoblastic-like multinucleate cell arrowed adjacent to confluent necrosis (A, H&E) and b-hCG 3þ staining in multinucleate cell (arrowed) and 2þ in mononucleate cells (B, immunoperoxidase).

DNA as part of the Copy Number Variation (CNV) project, Wellcome Trust Sanger Institute (first reported in Redon et al. 2006).5 The copy number variation profiles also showed numerous losses and a small number of gains across the genome in the endometrioid carcinoma. A large number of the losses were also present in the carcinoma with trophoblastic differentiation; however, additional gains and losses were observed, specifically losses in chromosome 1, 4, 9 and 17 (Fig. 3). Two normal b-hCG alleles were observed in both types of tumour. The common genetic changes between the endometrioid carcinoma and the carcinoma with trophoblastic differentiation are evidence of clonal origin. The significant number of additional genetic changes in the carcinoma with trophoblastic differentiation is typical of a high grade malignancy and indicates marked neoplastic progression of the endometrioid carcinoma. The additional genetic changes support that the trophoblastic differentiation is a higher grade tumour, but do not explain how the morphology of the trophoblastic differentiation occurs. The purpose of our study was to investigate whether the trophoblastic component was derived from the more conventional tumour using a genetic technique. We would expect that the clear cell component would also show clonality with the other two tumour types; however, this was not the aim of the study.

Fig. 3 Normal tissue (A), FIGO grade 2 endometrioid carcinoma (B) and carcinoma with trophoblastic proliferation (C): whole genome analysis. The chromosome regions are labelled 1-Y. The chromosomal gains and losses are blue and red lines, respectively.

The trophoblastic tumour in our case does not show changes described in choriocarcinoma. Using comparative genomic hybridisation (CGH), Poaty et al. found that there were no copy number abnormalities (CNAs) in six of ten primary gestational choriocarcinomas.6 The other four cases exhibited only 20 CNAs between them, with 11 CNAs observed in one case. The choriocarcinomas showed far fewer changes than the trophoblastic tumour described herein. We compared the genetic changes of our case to gestational trophoblastic tumours as there are minimal data on the genetics of non-gestational trophoblastic tumours in the literature. We acknowledge that there may be differences between gestational and nongestational trophoblastic tumours. In another study, Ahmed et al. showed amplification of 7q (four cases) and loss of 8p in five of 12 cases of choriocarcinoma examined.7 These changes were not observed in our patient. In a study with some parallels to our own case, Liu et al. performed a CGH study on a rare primary gastric tumour, most of which resembled choriocarcinoma, but with a minor component of adenocarcinoma.8 Copy number gains

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

CORRESPONDENCE

of chromosomes 12, 17, 20, 22, and X, together with losses on 18q, were most commonly seen. Except for the gain of chromosome 12, which is known to be uncommon in primary gastric adenocarcinoma but may be seen in choriocarcinoma, the other copy number gains/losses are commonly reported in primary gastric adenocarcinoma. Only one molecular study of a combined endometrioid carcinoma with a trophoblastic tumour had been performed prior to our report. Olson et al. found a clonal origin of the two components with genetic progression in the trophoblastic component.1 Our findings support the conclusions of Olson and colleagues who propose a number of mechanisms underlying this transformation. The analysis of many similar cases is required before hypothesising which genes may be important in trophoblastic differentiation. We describe a case of an endometrioid carcinoma combined with a carcinoma showing trophoblastic differentiation. Whole genomic testing found evidence of clonality between the two components and neoplastic progression in the trophoblastic component. The genetic changes were different to those expected of choriocarcinoma, possibly providing an explanation as to why the tumour did not respond to anti-choriocarcinoma chemotherapy. In summary, our study provides further evidence that trophoblastic differentiation in endometrioid carcinomas is a poor prognostic sign, supports a previous study showing clonal tumour progression1 and, in a new finding, shows that the b-hCG gene is not amplified despite the production of this hormone by the tumour. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Katie A. Ashton1 James Scurry2 Arman Ouveysi3 Joseph Ebbs4 Kenneth Jaaback4 Nikola A. Bowden1 1

The Centre for Information-Based Medicine, Hunter Medical Research Institute, New Lambton Heights, and the Discipline of Medical Genetics, School of Biomedical Sciences and Pharmacy, Faculty of Health, University of Newcastle, Callaghan, 2Division of Anatomical Pathology, Hunter Area Pathology Service, John Hunter Hospital, New Lambton Heights, 3School of Medicine and Public Health, University of Newcastle, Callaghan, and 4Hunter Centre for Gynaecological Cancer, John Hunter Hospital, New Lambton Heights, NSW, Australia Contact Dr K. A. Ashton. E-mail: [email protected]; katie.a.ashton@gmail. com 1. Olson MT, Gocke CD, Giuntoli RL 2nd et al. Evolution of a trophoblastic tumor from an endometrioid carcinoma-a morphological and molecular analysis. Int J Gynecol Pathol 2011; 30: 117–20. 2. Bradley CS, Benjamin I, Wheeler JE, et al. Endometrial adenocarcinoma with trophoblastic differentiation. Gynecol Oncol 1998; 69: 74–7. 3. Ashton KA, Scurry J, Rutherford J, et al. Nodular prurigo of the vulva. Pathology 2012; 44: 565–7. 4. Young KMN, Scurry J, Jaaback K, et al. Bilateral dysgerminoma associated with gonadoblastoma and sex-cord stromal tumour with annular tubules in a 28 year old fertile woman with normal karyotype. Pathology 2012; 44: 257–60.

353

5. Redon R, Ishikawa S, Fitch K, et al. Global variation in copy number in the human genome. Nature 2006; 444: 444–54. 6. Poaty H, Coullin P, Peko JF, et al. Genome-wide high-resolution aCGH analysis of gestational choriocarcinomas. PLoS One 2012; 7: e29426. 7. Ahmed MN, Kim K, Haddad B, et al. Comparative genomic hybridization studies in hydatidiform moles and choriocarcinoma: Amplification of 7q21-q31 and loss of 8p12-p21 in choriocarcinoma. Cancer Genet Cytogenet 2000; 116: 10–5. 8. Liu AY, Chan WY, Ng EK, et al. Gastric choriocarcinoma shows characteristics of adenocarcinoma and gestational choriocarcinoma: a comparative genomic hybridization and fluorescence in situ hybridization study. Diagn Mol Pathol 2001; 10: 161–5.

DOI: 10.1097/PAT.0000000000000101

Carcinoma cuniculatum of the cervix Sir, Carcinoma cuniculatum (epithelioma cuniculatum, CC) is a rare form of well differentiated squamous cell carcinoma (SCC) with a distinctive burrowing growth pattern (Latin, cuniculus ¼ burrow). It was first described by Aird et al. in 1954 with reference to three cases on the plantar region.1 Subsequently it has also been reported on the hands, sacrum, penis, oral cavity, oropharynx, and oesophagus.2 Here we report a case of CC of the cervix. A 34-year-old non-vaccinated for quadrivalent HPV disease, non-smoking woman attended for a routine Pap smear which revealed a possible low grade squamous intra-epithelial lesion. At colposcopy she had a raised, inflamed, 10 mm lesion with yellow exudate on the posterior lip of the cervix. The vagina was normal and there were no palpable lymph nodes. A cervical biopsy showed CIN 3. A cold knife cone biopsy of the cervix was then performed. The cone biopsy specimen measured 30
25
28 mm. Histological examination revealed a 10 mm lesion consisting of surface keratinising cervical intraepithelial neoplasia (KCIN) with deep burrowing branched sinuses (Fig. 1A). The sinus tracts contained keratin debris and were lined with KCIN (Fig. 1B). In the deep region of the tracts the epithelial lining was completely eroded and the tracts were lined by polymorphs overlying a dense layer of lymphocytes and plasma cells. Sinuses lined by this inflammatory reaction involved the deep margin which was 10 mm below the surface. p16 stain was positive in the KCIN (Fig. 1C). Polymerase chain reaction (PCR) performed on homogenised tissue sections for human papillomavirus (HPV) DNA revealed HPV type 18. A diagnosis of CC of the cervix with regression involving the deep margin was made. A repeat cold knife cone biopsy measuring 20
15
10 mm was performed 4 weeks later. Histological examination of the second cone biopsy showed a channel involving the full thickness of the posterior lip and extending to the deep margin just distal to the ectocervical margin (Fig. 2A). The channel was lined by polymorphs overlying a dense layer of lymphocytes and plasma cells, resembling the deep part of the lesion seen on the original biopsy (Fig. 2B). No residual abnormal squamous epithelium was seen in the channels or elsewhere in the specimen. Immunohistochemistry staining for involucrin confirmed that there was no residual squamous epithelium or keratin in the channel. A diagnosis of regressed CC was made on the second cone biopsy.

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

354

CORRESPONDENCE

Pathology (2014), 46(4), June

Fig. 2 Second cone biopsy. (A) Whole slide mount showing tract (black arrow); (B) tract lined by inflammation.

Fig. 1 First cone biopsy. (A) Whole slide mount showing branching tract lined by keratinising squamous epithelium superficially (white arrow) and deeply by inflammation (black arrow); (B) keratinising CIN lining the tract; and (C) p16 positive in the keratinsing CIN lining the tract.

The defining feature of CC, as seen in our case, is its unique burrowing architecture consisting of tracts lined by keratinising neoplastic squamous epithelium. We considered differential diagnoses of cervical intraepithelial neoplasia (CIN) involving glands and verrucous carcinoma (VC). The lesion was too deep for CIN involving the glands. Although in the literature CC is often referred to as either synonymous with or as a subset of VC, this lesion is not VC because it lacks the defining feature of exophytic growth. CC is an extremely rare carcinoma and we are unaware of any previous case on the cervix. However, given the multitude of sites where CC has been described, it is probable that CC can occur at any site with a squamous epithelium.

An early study suggested a possible role for HPV 11 in the pathogenesis of CC.3 The only subsequent study to demonstrate the presence of HPV DNA in CC is a case of the fingers where HPV types 31 and 33 were identified.4 PCR in our case showed HPV 18. Since CC may arise with and without HPV and when with HPV of high or low risk HPV types, it would seem that the pathogenesis of CC is not always directly related to HPV. The replacement of neoplastic squamous epithelium by an inflammatory reaction, as occurred in this case, is interpreted as regression. Regression of CC has not previously been documented. However, in keratoacanthoma (KA), another well differentiated neoplasm of squamous cells, regression is the rule and is characterised by an inflammatory reaction.5 The importance of a diagnosis of CC is that it has a better prognosis compared to typical SCC. Unlike typical SCC, CC rarely metastasises, but can be locally aggressive and recur.6 Treatment of CC usually involves wide excision.7 In this case, the dilemma was that it was uncertain whether the tracts lined by inflammation were still at risk for recurrence, even though no abnormal squamous epithelium remained in the second biopsy. In consultation with the patient, and cognisant of the apparent spontaneous regression, a decision was made

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

CORRESPONDENCE

355

for active surveillance after the second cone biopsy. No recurrence has occurred after 3 months from the second cone biopsy. In summary we report a case of CC with several unique features: cervical involvement, HPV 18 positivity and spontaneous regression. Melanie Smyth1 Ken Jaaback2 Sepehr N. Tabrizi3 Suzanne M. Garland4 Hui Yin5 James Scurry5 1

Medical Student, School of Medicine and Public Health, University of Newcastle, Callaghan, NSW, 2Hunter Centre for Gynaecological Cancer, John Hunter Hospital, New Lambton Heights, NSW, 3Department of Obstetrics and Gynaecology, University of Melbourne, The Royal Women’s Hospital and Murdoch Childrens Research Institute, Parkville, Vic, 4Molecular Microbiology Laboratory, Department of Microbiology and Infectious Diseases, Laboratory Services, Parkville, Vic, and 5Division of Anatomical Pathology, Hunter Area Pathology Service, John Hunter Hospital, New Lambton Heights, NSW, Australia Contact Associate Professor J. Scurry. E-mail: [email protected] 1. Aird I, Johnson HD, Lennox B, Stansfeld AG. Epithelioma cuniculatum: a variety of squamous carcinoma peculiar to the foot. Br J Surg 1954; 42: 245– 50. 2. Barreto JE, Velazquez EF, Ayala E, et al. Carcinoma cuniculatum: a distinctive variant of penile squamous cell carcinoma: report of 7 cases. Am J Surg Pathol 2007; 31: 71–5. 3. Knobler RM, Schneider S, Neumann RA, et al. DNA dot-blot hybridization implicates human papillomavirus type 11-DNA in epithelioma cuniculatum. J Med Virol 1989; 29: 33–7. 4. Wastiaux H, Dreno B. Recurrent cuniculatum squamous cell carcinoma of the fingers and virus. J Eur Acad Dermatol Venereol 2008; 22: 627–8. 5. Weedon D. Weedon’s Skin Pathology. 3rd ed. London: Churchill Livingstone, 2010; 667–708. 6. Kruse AL, Graetz KW. Carcinoma cuniculatum: a rare entity in the oral cavity. J Craniofac Surg 2009; 20: 1270–2. 7. Sun Y, Kuyama K, Burkhardt A, Yamamoto H. Clinicopathological evaluation of carcinoma cuniculatum: a variant of oral squamous cell carcinoma. J Oral Pathol Med 2012; 41: 303–8.

DOI: 10.1097/PAT.0000000000000100

Epithelioid haemangioendothelioma of the uterine cervix with CAMTA1/WWTR1 translocation Sir, A 40-year-old Caucasian woman presented to her general practitioner for routine Pap smear. Pelvic examination showed firm, pale discolouration of the ectocervical surface prompting referral to a gynaecologist. A cervical biopsy and subsequently a Fischer cone excision were performed, both of which showed multiple nodules of tumour in the subepithelial stroma of the cervical transformation zone. The largest nodule measured 6 mm in diameter. Minimal infiltration into the adjacent stroma was evident at the deep aspect of the nodules (Fig. 1A). The

Fig. 1 (A) Low power view of the Fischer cone biopsy section showing a pale, well circumscribed, focally infiltrative solid tumour at the cervical transformation zone (H&E). (B) High power view of the tumour cells and stroma. Note the nuclei-indenting intracytoplasmic vacuoles some of which contain red blood cells (H&E).

tumour was composed of large cells with eosinophilic cytoplasm in a dense hyalinised stroma. The tumour cells showed minimal pleomorphism, vesicular nuclei and small nucleoli. Occasional intracytoplasmic vacuoles, some of which contained red blood cell fragments, were evident (Fig. 1B). Mitotic figures were not seen. There was no evidence of human papillomavirus cytopathic effect, squamous or glandular dysplasia of the overlying epithelium. The tumour focally involved both the ectocervical and deep soft tissue margins of the Fischer cone biopsy specimen and a subsequent total hysterectomy was carried out 2 weeks later. The entire cervix and lower uterine segment were embedded and no residual tumour was identified. Immunohistochemistry was similar in both the initial cervical biopsy and the Fischer cone excision. The tumour cells were positive for vascular markers: CD31, Factor VIII associated protein and CD34 (Fig. 2). Smooth muscle (desmin, SMA), trophoblastic (b-hCG, hPL and p63) and epithelial markers (MNF116, AE1/AE3, EMA, CEA, 34bE12) were negative. A diagnosis of epithelioid haemangioendothelioma was made based on the morphology and immunohistochemical profile. To confirm the presence of the fusion between the CAMTA1 and WWTR1 genes, using the SureFISH custom probe design service (Agilent Technologies, USA) we designed two probes spanning the known breakpoint region on each gene that results

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

356

CORRESPONDENCE

Pathology (2014), 46(4), June

Fig. 3 Tumour cells show at least one fusion signal (yellow) confirming the presence of CAMTA1-WWTR1 gene fusion.

Fig. 2 The tumour cells show variable cell membrane and cytoplasmic immunopositivity for (A) Factor VIII, (B) CD31 and (C) CD34.

in t(1;3)(p36.23;q25.1).1 More than half of the tumour cells had at least one fusion signal indicating the presence of a CAMTA1-WWTR1 fusion. Images of the FISH result were captured using ISIS FISH Capture and Analysis software (Metasystems, Germany) and a representative image can be seen in Fig. 3. The term ‘epithelioid haemangioendothelioma’ (EHE) in its current usage was first coined by Weiss and Enzinger in 1982.2 Previously the term haemangioendothelioma had been used by various authors sometimes to refer to what is now called angiosarcoma3 and sometimes to benign vascular tumour, now called haemangioma.4 In the initial series of 41 cases,

EHE most commonly occurred in superficial and deep soft tissues of the limbs and was associated with a vein in approximately two-thirds of the cases. Recently the tumour has also been described in the uterus,5 clitoris6 and vulva.7 To the best of our knowledge, this is the first description of primary EHE of the uterine cervix in the English medical literature. The most useful clue to the diagnosis is the presence of sheets of so-called ‘blister cells’: large epithelioid cells, some of which contain intracytoplasmic vacuoles with red blood cell fragments.2 The tumour cells form sheets, cords and/or nests in a hyalinised or fibromyxoid background stroma. The cells may have prominent nucleoli and mitotic activity is variable. In a recent reported series of 49 cases, a two-tiered sub-classification was proposed.8 Tumours with >3 mitotic figures/50 high power fields and size >3.0 cm were considered high grade or malignant EHE and had a 5-year survival rate of 59%. Low grade EHE had a survival rate of 100% in this series. In all, metastases occur in 20–30% of EHE cases. Immunohistochemically, the tumour is positive for endothelial and lymphatic cell markers such as CD31, CD34, Factor VIII, von Willenbrand factor, FLI-1 and podoplanin (D2-40). This profile differentiates the tumour from sheets or nodules of

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

CORRESPONDENCE

poorly differentiated carcinoma, for which it is most often mistaken.2 Ultrastructurally EHE is characterised by Weibel-Palade bodies, immature cell junctions, abundant intermediate filaments and intracytoplasmic lumina formation.9 It has been shown that the majority of EHE has a characteristic translocation [t(1;3)(p36;3q25)] causing fusion of WWTR1 to CAMTA1.1 Recently, TFE3 gene rearrangements have been demonstrated in a subset of EHE without the translocation and showing distinct morphological features.10 The main differential diagnosis of EHE in the uterine cervix is squamous cell carcinoma. Poorly differentiated carcinomas are distinguishable by their clinical features, presence of marked cytological atypia, variable squamous or glandular differentiation and lack of endothelial marker expression. Pseudovascular squamous cell carcinoma, which may closely mimic a high grade vascular tumour, has been described in the cervix in one case report.11 EHE is CK7 and CK18 positive in at least 50% of cases due to the presence of abundant intracytoplasmic filaments in the tumour cells. Diagnosticians need to be aware of this potential pitfall in interpretation of the immunohistochemical stains.12 As in our case, squamous and glandular dysplasia, koilocytic atypia and p16 positivity are not expected in EHE. Glassy cell carcinoma is a rare histological variant of adenosquamous carcinoma of the cervix which is characterised by sheets or nests of pleomorphic large tumour cells with distinct cytoplasmic membranes, abundant granular to ground-glass cytoplasm, large oval nuclei, prominent macronucleoli and high mitotic activity. Eosinophils and plasma cells that often accompany the poorly differentiated tumour cells are also useful clues to diagnosis of this aggressive tumour.13 Evidence of squamous and glandular differentiation (intracytoplasmic mucin) may be present on histological and/or immunohistochemical examination. Glassy cell carcinoma is associated with HPV type 18.14 Histologically EHE should also be distinguished from epithelioid trophoblastic tumour (ETT), a recently described, rare gestational trophoblastic neoplasm which involves the cervix and lower uterine segment in half of the cases.15 ETT is characterised by nodules of epithelioid intermediate trophoblastic cells with abundant eosinophilic or clear cytoplasm and medium to large, irregular nuclei with prominent nucleoli. The centre of the tumour nodules are often hyalinised and multinucleated tumour cells may be seen. Mitoses, necrosis and apoptosis are frequent. Immunohistochemically, ETT is positive for trophoblastic markers: p63, inhibin, b-hCG, hPL and cytokeratins such as CK18.16 Placental site nodule is considered a benign counterpart or precursor of ETT. Epithelioid angiosarcoma is a more clinically aggressive vascular tumour which has been reported in the uterus in case reports.17,18 The frankly malignant endothelial cells in this subtype have abundant eosinophilic cytoplasm, large vesicular nuclei and associated brisk mitoses. The myxohyaline stroma that characterises EHE is not seen in angiosarcomas. Low grade vascular tumours that have been reported in the uterine cervix and are to be distinguished from EHE include Kaposiform haemangioendothelioma, spindle cell haemangioma, cavernous haemangioma and Masson’s tumour or intravascular papillary endovascular hyperplasia. There are no data available on the management of EHE in the uterine cervix. However, based on its behaviour in other organ

357

systems, we suggest complete surgical excision with close follow-up to exclude recurrence and metastatic disease. Our patient had gynaecological oncologist’s follow-up for 2 years after the total hysterectomy and remains free of the disease 5 years later. Acknowledgements: The authors would like to thank Dr Geoffrey Michael Strutton of Pathology Queensland, Princess Alexandra Hospital, Brisbane, for peer review of the manuscript. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Admire Matsika1 Jasen K. Anderson2 Judith F. Bligh3 Ann L. Whitehouse3 1

Anatomical Pathology, Pathology Queensland and University of Queensland, Brisbane, 2Cytogenetics Department, and 3 Histopathology Department, Sullivan and Nicolaides Pathology, Taringa, Qld, Australia Contact Dr A. Matsika. E-mail: [email protected] 1. Errani C, Zhang L, Sung YS, et al. A novel WWTR1-CAMTA1 gene fusion is a consistent abnormality in epithelioid hemangioendothelioma of different anatomic sites. Genes Chromosomes Cancer 2011; 50: 644– 53. 2. Weiss SW, Enzinger FM. Epithelioid hemangioendothelioma: a vascular tumor often mistaken for a carcinoma. Cancer 1982; 50: 970–81. 3. Ehrmann RL, Griffiths CT. Malignant hemangioendothelioma of the uterus. Gynecol Oncol 1979; 8: 376–83. 4. Pearl GS, Takei Y, Tindall GT, et al. Benign hemangioendothelioma involving the central nervous system: ‘‘strawberry nevus’’ of the neuraxis. Neurosurgery 1980; 7: 249–56. 5. Koutsopoulos AV, Sivridis E, Tsikouras P, et al. Low-grade uterine epithelioid hemangioendothelioma presented as a submucosal leiomyoma during labor. Case Rep Pathol 2013; 423584: 20. 6. Strayer SA, Yum MN, Sutton GP. Epithelioid hemangioendothelioma of the clitoris: a case report with immunohistochemical and ultrastructural findings. Int J Gynecol Pathol 1992; 11: 234–9. 7. da Silva BB, Lopes-Costa PV, Furtado-Veloso AM, et al. Vulvar epithelioid hemangioendothelioma. Gynecol Oncol 2007; 105: 539–41. 8. Deyrup AT, Tighiouart M, Montag AG, et al. Epithelioid hemangioendothelioma of soft tissue: a proposal for risk stratification based on 49 cases. Am J Surg Pathol 2008; 32: 924–7. 9. Vasquez M, Ordonez NG, English GW, et al. Epithelioid hemangioendothelioma of soft tissue: report of a case with ultrastructural observations. Ultrastruct Pathol 1998; 22: 73–8. 10. Antonescu CR, Le Loarer F, Zhang L, et al. TFE3 gene rearrangements define a distinct subtype of epithelioid hemangioendothelioma (EHE). Lab Invest 2013; 93 (Suppl. 1): 8A-A. 11. Horie Y, Kato M. Pseudovascular squamous cell carcinoma of the uterine cervix: a lesion that may simulate an angiosarcoma. Pathol Int 1999; 49: 170–4. 12. Miettinen M, Fetsch JF. Distribution of keratins in normal endothelial cells and a spectrum of vascular tumors: implications in tumor diagnosis. Hum Pathol 2000; 31: 1062–7. 13. Silverberg SG, Ioffe OB. Pathology of cervical cancer. Cancer J 2003; 9: 335–47. 14. Kato N, Katayama Y, Kaimori M, et al. Glassy cell carcinoma of the uterine cervix: histochemical, immunohistochemical, and molecular genetic observations. Int J Gynecol Pathol 2002; 21: 134–40. 15. Fadare O, Parkash V, Carcangiu ML, et al. Epithelioid trophoblastic tumor: clinicopathological features with an emphasis on uterine cervical involvement. Mod Pathol 2006; 19: 75–82. 16. Jordan S, Randall LM, Karamurzin Y, et al. Differentiating squamous cell carcinoma of the cervix and epithelioid trophoblastic tumor. Int J Gynecol Cancer 2011; 21: 918–22.

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

358

Pathology (2014), 46(4), June

CORRESPONDENCE

17. Drachenberg CB, Faust FJ, Borkowski A, et al. Epithelioid angiosarcoma of the uterus arising in a leiomyoma with associated ovarian and tubal angiomatosis. Am J Clin Pathol 1994; 102: 388–9. 18. Tallini G, Price FV, Carcangiu ML. Epithelioid angiosarcoma arising in uterine leiomyomas. Am J Clin Pathol 1993; 100: 514–8.

DOI: 10.1097/PAT.0000000000000106

Polyacrylamide hydrogel (Aquamid) filler in vagina Sir, Commencing in the 1970s, injectable permanent and semipermanent filler substances, either polymer particle suspensions or viscous fluids, were developed for the treatment of facial wrinkles1 as well as more extensive injections of traumatic or developmental facial defects.2 Cosmetic augmentation of the lips, familiar to most pathologists, has expanded to include internal bulking of sites such as the cardiac sphincter in patients with gastro-oesophageal reflux,3 and the bladder neck4 and anal sphincter5 for incontinence. In the 1980s, they were extensively used for breast augmentation in China and the Ukraine and marketed as ‘Amazing Gel’.6 When such materials are surgically removed because of local complications, they will yield a perplexing histological appearance to those who have not encountered their use before. We recently encountered an unexpected example in a patient being treated for vaginal contracture and urinary urge incontinence. A 58-year-old woman had a modified Manchester repair in March 2000, which was followed almost immediately by a

cystocoele recurrence. An anterior colporrhaphy was performed 8 months later and involved augmentation with a bridge graft of de-epithelialised vaginal muscularis. This second surgery was followed immediately by urinary urgency and a constant feeling of bladder discomfort. In 2012, the patient developed another cystocoele recurrence. Clinical examination showed a concertina-like shortening of the anterior vaginal wall, to about 4 cm (i.e., a 50% shrinkage), with a diffuse ‘woody’ cicatrisation of the anterior vaginal wall. A preoperative diagnosis of polypropylene mesh-induced vaginal wall contracture was made, and the patient scheduled for mesh excision with site-specific cystocoele repair using a remodelling biological graft. Retrospective and quite specific enquiries of the patient concerning procedures between the original and most recent surgery did not elicit any satisfactory explanation for the pathological findings. The resected surgical specimen was a flattened piece of soft fibrous tissue 35
35
12 mm (Fig. 1), with a variably ‘mucoid’ cut surface. Histologically, acellular and somewhat hyalinised fibrocollagenous connective tissue was disrupted by sharply circumscribed lakes and pools of homogeneous, nonbirefringent, basophilic, Alcian blue positive and mucicarminophilic foreign material which did not dissolve out during processing (Fig. 2) and distributed in a fashion not dissimilar to that seen with epithelial mucin in pseudomyxoma. However, it was unassociated with any epithelial cells and had elicited no stromal reaction whatsoever, except only where it had been injected adjacent to a small intramural squamous

A

B Fig. 1 The surgical specimen with a vaguely lobulated appearance. Mucoid ‘pseudocysts’ are visible at the lower margin of the specimen.

Fig. 2 Low power of pools of Aquamid gel in soft tissue of the anterior vaginal wall (A, H&E; B, mucicarmine).

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

CORRESPONDENCE

Fig. 3 A florid foreign body giant cell reaction to degenerate squames lining portion of a vaginal cyst. Polyacrylamide gel was present in some of the giant cells but this might have been incidental (H&E).

epithelial-lined cyst, where a florid giant cell histiocytic response was observed to both degenerate squames as well as, possibly, the polyacrylamide hydrogel (Fig. 3). Vascular space invasion was also prominent in the immediate vicinity of the interstitial pools of gel but also away from them (Fig. 4).

359

The popular synthetic permanent soft tissue fillers fall into three main groups: (1) homogeneous polymer gels, (2) suspensions of insoluble microspheres or fragments in resorbable liquids and, (3) suspensions of slowly degradable polymer microspherules in a resorbable liquid.7 To these might be added (non-synthetic) autologous fat injections for completeness. A comprehensive review of their biology and clinical applications is given by Broder and Cohen.8 The present case falls into the first group of which the silicone gels (e.g., Dermagen) and non-toxic polyacrylamide hydrogels (e.g., Aquamid) are the dominant members. Aquamid is manufactured by Ferrosan A/S in Denmark and is a transparent injectable gel of 97.5% water and 2.5% hydrophilic cross-linked polyacrylamide. The injected gel rarely induces a foreign body reaction,7,9,10 as in this case, but is associated with the variable production of a fine fibrocellular capsule that thickens over 6–9 months.8,11–13 Depending on the volume of polyacrylamide injected, it has a tendency to slowly disperse through the tissues.12 No studies have indicated this is by lymphatic transport to regional lymph nodes as suggested by our observations of lymphovascular space invasion. It is approved for use in Europe, South America, the Middle East and Australia and will presumably present itself more frequently in years to come as a diagnostic trap for the unwitting pathologist. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Peter Russell1,2 Mark Wilsher1 Richard Reid3 1

GynaePath, Douglass Hanly Moir Pathology, Macquarie Park, Department of Obstetrics Gynaecology and Neonatology, University of Sydney, and 3private practice, Sydney, Australia 2

Contact Professor P. Russell. E-mail: [email protected]

A

B Fig. 4 Polyacrylamide gel within vascular spaces. (A) Large, thick-walled lymphatic. (B) Small vascular spaces well away from the interstitial pools of gel (H&E).

1. Kinney BM, Hughes CE 3rd. Soft tissue fillers: an overview. Aesthet Surg J 2001; 21: 469–71. 2. Cahill KV, Burns JA. Volume augmentation of the anophthalmic orbit with cross-linked collagen (Zyplast). Arch Ophthalmol 1989; 107: 1684–6. 3. Feretis C, Benakis P, Dimopoulos C, et al. Endoscopic implantation of Plexiglas (PMMA) microspheres for the treatment of GERD. Gastrointest Endosc 2001; 53: 423–6. 4. Bent AE, Foote J, Siegel S, et al. Collagen implant for treating stress urinary incontinence in women with urethral hypermobility. J Urol 2001; 166: 1354–7. 5. Whitehead WE, Wald A, Norton NJ. Treatment options for fecal incontinence. Dis Colon Rectum 2001; 44: 131–42. 6. Cheng NX, Wang YL, Wang JH, et al. Complications of breast augmentation with injected hydrophilic polyacrylamide gel. Aesth Plast Surg 2002; 26: 375–82. 7. Christensen L, Breiting V, Janssen M, et al. Adverse reactions to injectable soft tissue permanent fillers. Aesth Plast Surg 2005; 29: 34–48. 8. Broder KW, Cohen SR. An overview of permanent and semipermanent fillers. Plast Reconstruct Surg 2006; 118: 7S–14S. 9. Kawamura JY, Domaneschi C, Migliari DA, Sousa SO. Foreign body reaction due to skin filler: a case report. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2006; 101: 469–71. 10. Parada MB, Michalany NS, Hassun KM, et al. A histologic study of adverse effects of different cosmetic skin fillers. Skinmed 2005; 4: 345–9. 11. Christensen LH, Breiting VB, Aasted A, Jorgensen A, Kebuladze I. Longterm effects of polyacrylamide hydrogel on human breast tissue. Plast Reconstruct Surg 2003; 111: 1883–90.

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

360

Pathology (2014), 46(4), June

CORRESPONDENCE

12. Lemperle G, Morhenn V, Charrier U. Human histology and persistence of various injectable filler substances for soft tissue augmentation. Aesth Plast Surg 2003; 27: 354–66. 13. Zarini E, Supino R, Pratesi G, et al. Biocompatibility and tissue interactions of a new filler material for medical use. Plast Reconstruct Surg 2004; 114: 934– 42.

DOI: 10.1097/PAT.0000000000000105

Olmesartan induced enterocolitis Sir, Olmesartan is an angiotensin II receptor antagonist used for the management of hypertension. It has been available for prescription in Australia since February 2007.1 Recently, RubioTapia et al. reported a series of 22 cases of severe sprue-like enteropathy associated with olmesartan use.2 We report a case with biopsy findings of sprue-like enteropathy with associated collagenous ileitis and colitis that showed complete resolution of histological and clinical symptoms following withdrawal of the medication. The patient was a 78-year-old woman with a clinical history of hypertension, gastroesophageal reflux disease, hypercholesterolaemia and osteoarthritis, who had been prescribed olmesartan for 4 years. Other regular medications were lercanidipine, amitriptyline, atenolol and atorvastatin. There was no history of recent use of a non-steroidal anti-inflammatory medication. Over the past 4 months she had experienced severe watery diarrhoea which resulted in three hospital admissions, including an ICU admission for acute renal failure secondary to dehydration. Upper endoscopy and colonoscopy were performed to investigate the severe symptoms. Biopsies revealed mild villous blunting in the proximal small intestine with intraepithelial lymphocytosis and lamina propria inflammation (Fig. 1A). In the terminal ileum, and colonic biopsies, there was

thickening of the subepithelial basement membrane, intraepithelial lymphocytosis and lamina propria chronic inflammation with eosinophil infiltration up to 50 per high power field (HPF) (
400; Olympus BX 41 microscope; Olympus, Japan) (Fig. 1B,C). Apoptosis was readily seen within the colonic crypts with up to 10 apoptotic bodies/HPF (
400; Olympus BX 41 microscope) (Fig. 1D). Gastric biopsies were not taken. An initial diagnosis of coeliac-like enteropathy, collagenous ileitis and collagenous colitis was made. The differential diagnosis for this constellation of findings included coeliac disease, autoimmune enteropathy, chronic variable immunodeficiency (CVID)3 and medication reaction. Of the patient’s regular medications, olmesartan2 and atorvastatin, a statin family medication,4 have been reported with microscopic colitis; however, only olmesartan has been associated with a sprue-like enteropathy and hence it was considered the most likely drug culprit. Serum tissue transglutaminase antibodies were negative and the clinical setting did not support an autoimmune enteropathy or immunodeficiency syndrome. Withdrawal of olmesartan and atorvastatin with implementation of total parenteral nutrition and oral budesonide produced resolution of diarrhoea. On selectively recommencing only olmesartan and oral intake, the diarrhoea returned. Her antihypertensive medication was subsequently changed to ramipril and the diarrhoea again resolved. A subsequent colonoscopy, performed 4 months after the initial biopsies, showed microscopically normal appearing small and large intestine with complete resolution of enteropathy-like changes and thickened collagen band (Fig. 2A,B). Sprue-like enteropathy associated with olmesartan therapy was recently reported in a series of 22 cases.2 In those with available detailed medical data, a mean interval between commencing olmesartan and developing diarrhoea was 3.1 years (range 0.5–7 years).2 Baseline intestinal biopsies in all patients showed villous atrophy with seven cases displaying features of collagenous sprue. Three patients also exhibited collagenous colitis. All patients had negative serum

A

B

C

D

Fig. 1 (A) Proximal small intestinal biopsy displaying mild villous blunting and lamina propria inflammation. (B) Terminal ileal biopsy displaying mild thickening of the subepithelial collagen layer and lamina propria inflammation (‘collagenous ileitis’). (C) Collagenous colitis. (D) Prominent crypt apoptosis in the colonic biopsies.

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

CORRESPONDENCE

A

361

B

Fig. 2 Improvement in the appearance of histological features in (A) proximal small intestine and (B) colon following withdrawal of the drug.

tissue transglutaminase antibodies, and the symptoms and histological findings resolved completely after ceasing olmesartan. Twenty patients had been pre-treated with budesonide/ corticosteroids prior to entering the study. To our knowledge, since this publication there have been 20 additional reported cases of suspected olmesartan associated enteropathy with either histological resolution, clinical resolution or both.2,5–9 This includes 16 identified in a retrospective review of cases originally believed to represent coeliac disease, accounting for 84% of all mediation related duodenal villous atrophy.7 The mechanism associated with olmesartan related enteropathy is unclear. The typically long delay between commencing olmesartan and development of diarrhoea suggests cell mediated immune damage rather than a type 1 hypersensitivity reaction. Angiotensin II receptor antagonist inhibition of transforming growth factor beta (TGFb) action, a molecule important for gut immune homeostasis,2,3,5 and/or a cell mediated immune reaction that damages the small intestinal brush border6 may be responsible. There are marked similarities between our case and the series of olmesartan related enteropathy reported by Rubio-Tapia et al. While a medication reaction is always difficult to prove, the resolution of symptoms and histological findings in the absence of other medical conditions, suggest that this association is unlikely a result of chance. Furthermore, the conspicuous finding of apoptosis and the prominent eosinophilia in our case are features that can indicate a medication reaction.10 Olmesartan is a relatively new anti-hypertensive medication and it is possible that more cases of enteropathy will be seen in the future. Thus, it is important to consider olmesartan induced enteropathy in patients with histological sprue-like findings, with or without colonic inflammation, in the absence of other coeliac disease or other medical condition. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Carrie Gallivan1 Ian Brown1,2 1

Anatomical Pathology, Pathology Queensland, Brisbane and Women’s Hospital, Herston, and Specialist Pathologists, Herston, Qld, Australia Contact Dr I. S. Brown. E-mail: [email protected]

2

Royal Envoi

1. Australian Register of Therapeutic Goods (ARTG). Approved Product Information: Olmetec. Canberra: ARTG, 13 Sept 2012. 2. Rubio-Tapia A, Herman ML, Ludvigsson JF, et al. Severe spruelike enteropathy associated with olmesartan. Mayo Clin Proc 2012; 87: 732–8. 3. Daniels JA, Lederman HM, Maitra A, Montgomery EA. Gastrointestinal tract pathology in patients with common variable immunodeficiency (CVID): a clinicopathologic study and review. Am J Surg Pathol 2007; 31: 1800–12. 4. Pascua MF, Kedia P, Weiner MG, Holmes J, Ellenberg J, Lewis JD. Microscopic colitis and medication use. Clin Med Insights Gastroenterol 2010:11–9. 5. Talbot G. Small bowel histologic findings suggestive of coeliac disease in an asymptomatic patient receiving olmesartan. Mayo Clin Proc 2012; 87: 1231–2. 6. Dreifuss S, Tomizawa Y, Farber N, Davison J, Sohnen A. Spruelike enteropathy associated with olmesatan: an unusual case of severe diarrhea. Case Rep Gastrointest Med 2013; 2013: 618071. 7. DeGaetani M, Tennyson CA, Lebwohl B, et al. Villous atrophy and negative celiac serology: a diagnostic and therapeutic dilemma. Am J Gastroenterol 2013; 108: 647–53. 8. Nielsen JA, Steephen A, Lewin M. Angiotensin-II inhibitor (olmesartan)induced collagenous sprue with resolution following discontinuation of drug. World J Gastroenterol 2013; 19: 6928–30. 9. Stanich PP, Yearsley M, Meyer MM. Olmesartan-associated sprue-like enteropathy. J Clin Gastroenterol 2013; 47: 894–5. 10. Villanacci V, Casella G, Bassotti G. The spectrum of drug-related colitides: important entities, though frequently overlooked. Dig Liver Dis 2011; 43: 523–8.

DOI: 10.1097/PAT.0000000000000104

High takeoff of the left main coronary artery at autopsy after sudden unexpected death in a male

Sir, Coronary artery anomaly is a rare but important cause of sudden unexpected death, probably because of myocardial ischaemia.1,2 One of these anomalies is high takeoff of the coronary artery, in which the ostium of the coronary artery is located above the sinotubular junction of the aortic valve.3 Clinical and pathological reports of sudden death in patients with high takeoff of the coronary artery are extremely rare,2,4 and the importance of this anomaly as a cause of myocardial ischaemia or sudden unexpected death is controversial. Here, we report the autopsy findings in a patient with sudden unexpected death in whom there was high takeoff of the left

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

362

CORRESPONDENCE

main coronary artery (LMCA) and who died when in a state of emotional arousal. A 38-year-old male collapsed suddenly at home during a heated argument with his family. He was immediately transferred to hospital, but a heartbeat could not be re-established. The man had had an approximately 20-year history of schizophrenia; he had not worked or participated in sports. Neither he nor his family had a history of cardiac problems. At autopsy there was no injury to the body. Toxicological screening of sampled blood was negative. The heart weighed 400 g; serial horizontal sections showed mild hypertrophy of the left ventricle without myocardial necrosis (Fig. 1A). The ostium of the LMCA was anomalously located about 1.5 cm above the sinotubular junction (Fig. 1B). The ostium did not have an obvious valve-like ridge, but serial horizontal sections of the LMCA showed that proximally the artery ran intramurally in the aorta for a short distance (Fig. 1C). The angle between the aorta and the exiting LMCA was acute (less than 458). In the posterior wall of the left ventricle, because the posterior descending artery branched off from the right coronary artery, the coronary system was considered to be right dominant. Serial 3 mm sections of the coronary tree showed no marked atherosclerotic stenosis (i.e., beyond 75%). Microscopically, parts of the walls of the intramural LMCA and the aorta shared a common media. There was no interpositioned adventitia (Fig. 2A). Immunohistochemistry using CD34 and a-smooth muscle actin showed a lack of smooth muscle cells in the media of the intramural LMCA (Fig. 2B), and endothelial cells continuously lined the luminal surface. No coagulative necrosis of myocytes of either ventricle was found, but severe interstitial fibrosis was present in the subendocardium of the anterior wall and the ventricular septum, which were the territories of the left descending coronary artery (Fig. 2C). The incidence of high takeoff of the LMCA is extremely low. This anomaly has been found in five of 126,000 serial coronary

Pathology (2014), 46(4), June

angiographies,5 and a comparable rate was obtained recently in a number of 64-slice coronary computed tomography scan series.6 However, the definition of high takeoff of the coronary artery remains controversial. If it includes all cases in which the coronary ostium is located above the sinotubular junction, even by only a small distance, then many such cases must be considered clinically to be normal variants. For this reason, various researchers have proposed defining significantly high takeoff as more than 5 mm,4 1 cm,7 or 2 cm8 above the sinotubular junction. Nevertheless, it may still be difficult to decide whether a high takeoff is of significance by simply measuring the distance between the coronary ostium and the sinotubular junction. The pathological significance of high takeoff of the coronary artery is also controversial. This anomaly is believed to be linked to myocardial ischaemia and sudden death due to decreased coronary perfusion by displacement of the coronary ostium from coronary sinus,2,4,8 but some researchers still consider high takeoff to be a ‘benign’ anomaly.5 In a literature search, we were unable to find any autopsy or clinical cases in which high takeoff of the LMCA was considered to be the main cause of ischaemia or sudden death. In contrast, there have been a few reports of high takeoff of the right coronary artery as a possible cause of myocardial ischaemic events and sudden death.2,4 Origination of a coronary artery from the wrong coronary sinus is the most common coronary artery anomaly.1 These anomalies have been reported as morphological findings that have clinical importance and could cause sudden cardiac death from myocardial ischaemia.1,9 However, they are also found incidentally at autopsy and therefore are not considered to be uniformly fatal. According to Taylor et al.1 a left coronary artery originating from the right coronary sinus accounts for 57% of sudden deaths in patients with isolated coronary artery abnormalities, whereas a right coronary artery originating from the left coronary sinus accounts for 25%. Some researchers

Fig. 1 Gross appearance of the heart. (A) Horizontal section of the ventricle. (B) View of the aortic root, showing high takeoff of the left main coronary artery (arrow). (C) Serial sections of the left main coronary artery, showing its intramural course in the aorta over a short distance (arrows).

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

CORRESPONDENCE

363

Fig. 2 Microscopic appearance of the heart. (A) Intramural course of the left main coronary artery (Elastica–Masson). (B) Immunohistochemistry for a-smooth muscle actin, showing the absence of smooth muscle cells in the outer media (M) of the intramural part of the artery. Arrow shows positive staining of capillary in the periadventitial adipose tissue (A). (C) Interstitial fibrosis in the ventricular septum (Elastica–Masson).

have noted anatomical or pathological coronary disorders other than dislocation of the ostium that could explain the variable prognosis in patients with these anomalies. These lesions, which include an intra-arterial course of the proximal coronary artery (sandwiched between the aorta and the pulmonary artery), an intramural course of the proximal coronary artery (within the aortic wall), acute takeoff of the coronary artery along the aortic wall, and the presence of an ostial valve-like ridge, are considered to further diminish blood flow in the coronary artery.1,9 When a coronary ostium originates from the wrong coronary sinus, dislocation of the coronary ostium occurs mainly in a horizontal direction, whereas high takeoff of a coronary ostium yields dislocation in a vertical direction. In our subject, careful examination revealed that the acute angle of takeoff and the intramural course of the LMCA were associated with upward dislocation of the left coronary ostium. Thus high takeoff of the LMCA may also be an important cause of sudden death, at least when additional anatomical changes such as intramural course and acute takeoff angle result from dislocation of the ostium and potentially further diminish blood flow in the coronary artery. The importance of the association between high takeoff of the LMCA and sudden death is reflected in the previously proposed association between the coronary arteries originating from wrong coronary sinus and sudden death.1,2,9 Of course, acquired coronary artery changes such as atherosclerotic narrowing may also be important factors in ischaemia in high takeoff of the coronary artery, as is the case with other coronary anomalies.1,10 Our subject was in an excitable state just before his collapse. It is well known that in some cases obvious triggering events have occurred before sudden cardiac death in patients with coronary artery abnormality; exercise is believed to be the most important of these events.11 In addition, our case suggests that emotional stressors such as anger are important risk factors for sudden death. Examination of healthy populations has shown

that emotional stress increases systolic blood pressure and sympathetic tone and thus vasoconstriction.12 In patients with coronary anomalies, exercise not only increases the heart’s oxygen demand but also suddenly dilates the aortic root by increasing systolic blood pressure. The dilated aortic root may have pressed on the intramural artery of our subject, thus decreasing blood flow in the coronary artery. Our subject had low everyday levels of physical activity; his sudden excitation before death may have suddenly decreased the flow in the intramural part of the LMCA through increased pressure from the aorta. This shows the importance of emotional stress as a trigger of sudden death in cases of anomalous dislocation of a coronary ostium with intramural course. In conclusion, the present case shows that high takeoff of the LMCA may not be a harmless anomaly, especially when other anatomical or pathological findings are associated with dislocation of a coronary ostium. In addition, an understanding of the state of patients with coronary artery anomaly just before death, and thus an understanding of triggering events, may help to determine the cause of death at autopsy. Conflicts of interest and sources of funding: This work was supported in part by a KAKENHI grant from JSPS, Japan, to YH (24590852). The authors state that there are no conflicts of interest to disclose. Naoki Nishida Yukiko Hata Koshi Kinoshita Department of Legal Medicine, Graduate School of Medicine and Pharmaceutical Sciences, University of Toyama, Toyama, Japan Contact Professor N. Nishida. E-mail: [email protected]

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

364

Pathology (2014), 46(4), June

CORRESPONDENCE

1. Taylor A, Rogan KM, Virmani R, et al. Sudden cardiac death associated congenital coronary anomalies. J Am Coll Cardiol 1992; 20: 640–7. 2. Frescura C, Basso C, Thiene G, et al. Anomalous origin of coronary arteries and risk of sudden death: A study based on an autopsy population of congenital heart disease. Hum Pathol 1998; 29: 689–95. 3. Angelini P. Normal and anomalous coronary arteries: definition and classification. Am Heart J 1989; 117: 418–34. 4. Menke DM, Waller BF, Pless JE. Hypoplastic coronary arteries and high takeoff position of the right coronary ostium. Chest 1985; 88: 299–301. 5. Yamanaka O, Hobbs RE. Coronary artery anomalies in 126,595 patients undergoing coronary angiography. Cathet Cardiovasc Diagn 1990; 21: 28– 40. 6. Montaudon M, Latrabe V, Iriart X, et al. Congenital coronary anomalies: review of the literature and multidetector computed tomography (MDCT)appearance. Surg Radiol Anat 2007; 29: 343–55. 7. Kosar P, Ergun E, Ozturk C, et al. Anatomic variations and anomalics of the coronary arteries: 64-slice CT angiographic appearance. Diagn Interv Radiol 2009; 15: 275–83. 8. Thakur R, Dwivedi SK, Puri VK, et al. Unusual ‘high take off’ of right coronary artery from ascending aorta. Int J Cardiol 1990; 26: 369–71. 9. Roberts W, Siegel R, Zipes DP. Origin of the right coronary artery from the left sinus of Valsalva and its functional consequence: analysis of 10 necropsy cases. Am J Cardiol 1982; 49: 863–8. 10. Nishida N, Chiba T, Ohtani M, et al. Two adult cases of congenital atresia of the left coronary ostium- comparison of a sudden death case with a longterm survival case. Virchows Arch 2005; 447: 742–6. 11. Basso C, Maron BJ, Corrado D, et al. Clinical profile of congenital coronary artery anomalies with origin from wrong aortic sinus leading to sudden death in young competitive athletes. J Am Coll Cardiol 2000; 35: 1493–501. 12. Leor J, Poole WK, Kloner RA. Sudden cardiac death triggered by an earthquake. N Engl J Med 1996; 334: 413–9.

DOI: 10.1097/PAT.0000000000000099

First report of adrenal cortical endothelial (vascular) cyst mimicking phaeochromocytoma (pseudophaeochromocytoma) Sir, Phaeochromocytomas are catecholamine-secreting tumours of the adrenal medulla that typically present with symptoms

including headache, diaphoresis, and tachycardia, signs of hypertension, and elevated catecholamines. Rarely, adrenal cortical tumours and other conditions may present with similarly elevated catecholamines, and are thus referred to as pseudophaeochromocytomas. We present the first case of an adrenal cortical endothelial cyst mimicking signs, symptoms, and catecholamine elevation characteristic of phaeochromocytoma, representing a new aetiology of pseudophaeochromocytoma. A 36-year-old female had a right adrenal mass incidentally found during work-up after a fall. Abdominal computed tomography (CT) scan demonstrated a 5.6
2.6 cm hypoattenuating cystic-type lesion in the right adrenal (Fig. 1A). Four years later, the patient complained of headaches and palpitations and was noted to be hypertensive (144/98 mm Hg). Repeat imaging demonstrated that the mass was unchanged. Plasma norepinephrine (893 pg/mL) and dopamine (34 pg/mL) were elevated and cortisol (2.00 pg/mL) and aldosterone (4.7 pg/mL) were decreased. A laparoscopic right-sided adrenalectomy was performed for a presumptive phaeochromocytoma with cystic change. Follow-up cortisol, ACTH, and catecholamines were normal, although post-operative blood pressure was still mildly elevated (127/69 mm Hg). The 18.3 g adrenalectomy specimen contained a multiloculated cyst measuring 4.4
2.5
1.5 cm (Fig. 1B). Normal appearing adrenal tissue was present adjacent to the cyst. Microscopic examination revealed a cyst lined by flattened, bland endothelial cells. At the periphery of the cyst, anastomosing channels were present (Fig. 1C). Immunohistochemical stains demonstrated that the cyst lining was positive for vascular markers including CD31, ERG, and D2-40 (Fig. 1D). SMA was also positive. The cyst did not express AE1/3 or CK5/6. There are multiple entities which can produce elevated catecholamines and phaeochromocytoma-like symptoms. Adrenal cysts are rare and can occur at any age but most often present in the 5th to 6th decade of life and are three times more prevalent in women. Adrenal cysts are classified into four major types: endothelial (vascular), pseudocysts, parasitic, and epithelial.1 Endothelial cysts are the most common type

A

B

C

D

Fig. 1 (A) Abdominal CT scan demonstrating a hypoattenuating lesion in right adrenal gland. (B) Gross adrenalectomy specimen showing a multiloculated cystic lesion with adjacent remaining adrenal cortex. (C) The well-circumscribed cystic lesion was comprised of flattened cells with anastomosing channels. (D) CD31 immunohistochemical stain showed strong reactivity in the endothelial cyst lining. Adrenal cortical cells adjacent to the cyst were negative.

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

CORRESPONDENCE

of cyst within the adrenal gland. These cysts are lined by flat to cuboidal endothelial cells that may communicate with the surrounding microvasculature. The cyst may be haemorrhagic with fibrosis, calcification, or metaplastic bone or fat. We document for the first time that an adrenal endothelial cyst can produce phaeochromocytoma-like symptoms and catecholamine elevation. Several lesions and dysfunctions of the adrenal gland, such as non-functioning adrenal cortical adenoma/carcinoma and adrenomedullary response to venography, have been noted as rare causes of pseudophaeochromocytoma.2 Regarding adrenal cysts, in addition to the patient described in this report, one case of a hydatid cyst with phaeochromocytoma-like findings has been published. The patient presented with flank pain, headaches, palpitations, hypertension, and elevated vanillylmandelic acid, initially suggesting phaeochromocytoma.3 The mechanism by which adrenal lesions produce elevated levels of catecholamines is not known, but may be mediated by mass effect of the lesion pushing against the adrenal medulla.4 In conclusion, physicians should be aware that an adrenal endothelial cyst is an addition to the continually growing list of non-neoplastic causes of pseudophaeochromocytoma. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Meredith O. Morse1 Frank L. Schwartz2 Debra L. Zynger1 1 Department of Pathology, The Ohio State University Medical Center, Columbus, and 2Department of Specialty Medicine, Ohio University Heritage College of Osteopathic Medicine, Athens, OH, USA

Contact Dr D. Zynger. E-mail: [email protected] 1. Lack EE. AFIP Atlas of Tumor Pathology. Tumors of the Adrenal Glands and Extraadrenal Paraganglia. Series 4. Washington DC: American Registry of Pathology, 2007; 185–8. 2. Kuchel O. Pseudopheochromocytoma. Hypertension 1985; 7: 151–8. 3. Horchani A, Nouira Y, Nouira K, et al. Hydatid cyst of the adrenal gland: a clinical study of six cases. Scientific World J 2006; 6: 2420–5. 4. Perrino CM, Prall DN, Calomeni EP, et al. Ultrastructural findings in adrenal cortical adenomas clinically mimicking pheochromocytoma: a comparison with other adrenal tumors and tissue preparation techniques. Ultrastruct Pathol 2012; 36: 287–93.

DOI: 10.1097/PAT.0000000000000109

Cysticercosis in the thyroglossal duct cyst: rare site for a common problem Sir, Cysticercosis is caused by the larval stage of Taenia solium. It is a major public health problem in India.1 In a community survey in India, neurocysticercosis (NCC) was detected in 28.4% of the patients who had epilepsy.2 Cysticercosis is considered as a biological marker of the social and economic development of a

365

community.1 Even though NCC is the most common, other sites such as muscles and subcutaneous tissues are also affected.1 We present a case of cysticercosis of the thyroglossal duct cyst (TGD) which has not been reported previously in the literature. Institutional ethical committee approval was obtained for publication. A 7-year-old girl from a poor socioeconomic background presented to us with an asymptomatic swelling in the upper neck at the infrahyoid region which was of 7 months duration. Clinically the diagnosis was TGD (Fig. 1A). Ultrasound confirmed a hypoechoeic lesion in the subhyoid region. Fine needle aspiration cytology was not performed since the clinical and radiological diagnosis was TGD. Erythrocyte sedimentation rate (ESR) was 20 mm/hr. At surgery a 2.5
1.5
1 cm cystic lesion was detected in an infrahyoid location which was extending superiorly to the hyoid bone with the track ending in the hyoid (Fig. 1B). Sistrunk’s operation was performed. The histopathology revealed cysticercosis in the TGD (Fig. 2A,B). Preoperative screening was negative for HIV. There was no history of epilepsy which is often seen in neurocysticercosis. Humans are the definite hosts and pigs the intermediate host in the life cycle of T. solium. The adult tape worm is lodged in humans and the eggs are passed in faeces. Once the pigs ingest the eggs they develop as cysticerci in their internal organs such as muscle and brain. In humans, ingestion of undercooked pork meat results in development of adult worm infection in the small intestines.1 Humans can become the intermediate host if they happen to ingest the eggs and subsequently cysticerci develop in various organs due to spread via the blood stream. The most common area involved is the central nervous system, although involvement of skeletal muscle, diaphragm, heart, pleura and peritoneum have been described.3 Thyroid involvement is rare; however, it is possible as part of disseminated cysticercosis.4 Our case was unique because this patient had neither disseminated cysticercosis nor was she immunocompromised. This case was thus an isolated TGD involvement which has not been reported previously. Considering the modes of development of cysticercosis it can be assumed that the cysticercosis developed in TGD by blood stream spread. During excision of this cyst, dense perilesional adhesions were noted, unlike usual TGD cysts. Histopathology revealed intense inflammation even though the child did not have any pain. Living parasites generally do not elicit inflammatory response, however dead and dying parasites are said to have the maximum potential to elicit local inflammation.1 Management in a case of isolated TGD cysticercosis is not described in the literature. We recommend screening for involvement at other sites including neurocysticercosis (NCC). As NCC may be asymptomatic, brain computed tomography/magnetic resonance imaging is indicated regardless of the presence or absence of symptoms. Ocular cysticercosis should be excluded by an ophthalmological examination in all patients.5 Our patient had a normal CT of the brain. For presence of active disease serological tests are described.1 Western blot technique has sensitivity of 98% if there are three or more cysticerci and 65% with lesser numbers. Serological evaluation of antigen TS14 present in the lentil lectin-bound fraction of cyst glycoproteins has been demonstrated to have utility in diagnosis of cysticercosis. This antigen was found in the sera of the majority of patients with cysticercosis but not in the sera of patients with other helminthic infections.6 In cases where there is difficulty in identifying cysticercosis in histopathological

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

366

Pathology (2014), 46(4), June

CORRESPONDENCE

Excised hyoid

Excised TGD cyst

A

B

Fig. 1 (A) Clinical appearance of the cystic swelling in the upper neck. (B) Excised thyroglossal duct cyst with central hyoid.

A

B

Fig. 2 (A) Scolex of the cysticercus with cuticle surrounded by mixed inflammatory infiltrate; (B) scolex of the cysticercus with cuticle (H&E).

specimens, polymerase chain reaction (PCR) based molecular diagnosis can be used. During this process, extracted DNA samples from the paraffin embedded tissue are taken and cytochrome c oxidase subunit 1 gene and cytochrome b gene are amplified for the detection of taeniid DNA.7 Frequent false negative results are reported with patients who have a single cysticercus.8 In our case it was not performed because the test was not available. Even though controversial, NCC is managed with systemic administration of praziquantel/albendazole.1 It is not clear whether a patient with apparently isolated cysticercosis should be treated with systemic antiparasitic agents. We planned a 1 week course of albendazole at 15 mg/kg body weight in divided doses, however such treatment should be performed with caution since anaphylactic reactions are described. To conclude, isolated single site cysticercosis does occur. The further management in such cases is not clearly defined and should include screening for multiple site involvement and a course of systemic therapy. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. P. V. Pradeep1 Vissa Shanthi2 Kesapuram KarunaKumar Reddy3 Kuppili Venkata Murali Mohan2

1

Department of Endocrine Surgery, Baby Memorial Hospital, Kozhikode, Kerala, 2Department of Pathology, and 3 Department of Endocrine Surgery, Narayana Medical College & Super Speciality Hospital, Chinthareddypalem, Nellore, Andhra Pradesh, India Contact Dr P. V. Pradeep. E-mail: [email protected] 1. Prasad KN, Prasad A, Verma A, Singh AK. Human cysticercosis and Indian scenario: a review. J Biosci 2008; 33: 571–82. 2. Rajshekhar VM, Raghava V, Prabhakaran V, et al. Active epilepsy as an index of burden of neurocysticercosis in Vellore district, India. Neurology 2006; 67: 2135–9. 3. Sankar SK, Suryanarayana V, Vasantha S, et al. Biology of neurocysticercosis-parasite related factors modulating host response. Med J Armed Forces India 1994; 50: 79–88. 4. Bhalla A, Sood A, Sachdev A, Varma V. Disseminated cysticercosis: a case report and review of the literature. J Med Case Rep 2008; 2: 137. 5. Prasad A, Gupta RK, Nath K, et al. What triggers seizures in neurocysticercosis? A MRI based study in pig farming community from a district of North India. Parasitol Int 2008; 55: 166–71. 6. Greene RM, Hancock K, Wilkins PP, Tsang VCW. Taenia solium: Molecular cloning and serological evaluation of 14- and 18-kDa related diagnostic antigens. J Parasitol 2000; 86: 1001–7. 7. Yamasaki H, Nakao M, Sako Y, Nakaya K, Ito A. Molecular identification of Taenia solium cysticercus genotype in the histopathological specimens. Southeast Asian J Trop Med Public Health 2005; 36 (Suppl 4): 131–4. 8. Coyle CM, Tanowitz HB. Diagnosis and treatment of neurocysticercosis. Interdiscip Perspect Infect Dis 2009; 2009: 180742.

DOI: 10.1097/PAT.0000000000000107

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

CORRESPONDENCE

The utility of syphilis point of care testing in remote Queensland communities

Sir, Over the past 2 years the rate of syphilis diagnosis in the North Queensland region has been steadily increasing, with large foci of infection identified in North-West communities. The remoteness of these locations, in conjunction with the particular socio-cultural characteristics of the population, poses unique challenges to the traditional diagnostic and treatment paradigms for syphilis. Point of care testing (PoCT) offers an opportunity to provide on-the-spot diagnosis and treatment in a setting where there is significant loss to follow-up. The Townsville Sexual Health Service (TSHS) adopted the SD Bioline Syphilis 3.0 (Alere, USA), a rapid PoCT utilising treponemal antibodies in an immunochromatographic card format, as an adjunct to traditional methods of syphilis diagnosis during a Youth Health Screen in Mount Isa. Data from that screening program and central laboratory validation have been collated and assessed in order to determine the applicability of this test to a unique Australian situation. A total of 240 participants were screened during TSHS deployment to Mount Isa in May 2012. Serum was obtained on-site with a portable centrifuge, in order to provide comparison between testing in the field and in a laboratory setting. The SD Bioline assay was performed according to instructions provided by the manufacturer, and read by a combination of clinical and non-clinical staff in the field. Results were then entered into a central database. Each specimen was tested in parallel by the standard testing regimen at the Townsville laboratory (Fig. 1) to provide a comparison of PoCT to gold standard. The sera were tested again by PoCT in a clinical laboratory setting to ensure there was no discrepancy in results between field testing and controlled conditions. Discrepant results were managed with a third performance of the test in the laboratory. All specimens were de-identified to ensure interpretation was non-biased and any re-testing was performed alongside known

367

negatives. Two specimens were removed from analysis as no results were entered from initial field use. Twenty-two of the 238 specimens tested were positive by PoCT, while 27 were positive by the reference standard (Table 1). Analysis of these results demonstrates a sensitivity of 81.5% [95% confidence interval (CI) 76.6%–86.4%], and a specificity of 100%. Of note, only five of the confirmed cases of syphilis were new diagnoses; the other 22 participants had previously received treatment for syphilis. Importantly, one of these five new diagnoses was not detected by the PoCT, suggesting that the test may suffer from a lack of sensitivity in early stages of the disease. The sensitivity is considerably lower than that achieved through the manufacturer’s testing (100% sensitivity and specificity) and independent, peer-reviewed testing (90.2–100% sensitivity and 94.9–100% specificity).1–3 There were six instances of discrepant results, with three of these representing a discrepancy between the PoCT and the reference standard, one of these representing a discrepancy between results recorded in the field and results recorded in the laboratory, and two results affected by both discrepancies (Table 2). One result was recorded as ‘?positive’ by staff in the field (Case 6) due to an inability to interpret the result. It should be noted that all discrepant results positive by enzyme immunoassay (EIA) are on a background of past diagnosis and treatment for syphilis, reducing the likelihood of these results representing a false positive. The current algorithm for diagnosis of syphilis in Pathology Queensland uses two treponemal tests [EIA (Abbott, USA) and Treponema pallidum particle agglutination (TPPA; Fujirebio, Japan)] and one non-treponemal antibody test [rapid plasma reagin (RPR; BD, USA)]. While treponemal tests have the advantage of being specific, their positivity can be persistent and may not reflect an active infection. The RPR, though nonspecific, is a marker of disease activity and helps to distinguish between active and inactive disease.4 The SD Bioline Syphilis 3.0, like most other PoCT for syphilis, relies upon detection of treponemal-specific antibody. Thus, while it is more specific than a non-treponemal antibody test like RPR, the lifelong persistence of these antibodies means

Syphilis diagnosis algorithm Serum

EIA POS

NEG

Known positive? YES

Perform RPR

No further testing NO

Perform RPR and confirmatory TPPA TPPA NEG

Confirmed positive

Apply comment reporting possible false positive result and suggest retesting in 2–4 weeks

Fig. 1 Townsville laboratory syphilis diagnosis algorithm (reference standard). EIA, IgG enzyme immunoassay (Abbott, USA); RPR, rapid plasma reagin (BD, USA); TPPA, Treponema pallidum particle agglutination (Fujirebio, Japan).

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

368

Pathology (2014), 46(4), June

CORRESPONDENCE

Table 1 Results of the SD Bioline Syphilis 3.0 in comparison to the laboratory reference method Syphilis status via reference method

Syphilis status via PoCT Pos Neg Total

Pos

Neg

Total

22 5 27

0 211 211

22 216 238

Neg, negative; PoCT, point-of-care test; Pos, positive

it cannot distinguish between active and treated/latent disease. The risk in using this test alone for diagnosis is that patients without disease will receive unnecessary treatment. However, the argument for the use of PoCT in syphilis is that the risk of serious outcomes, such as congenital syphilis and the likelihood of continued spread of disease, outweighs the risk of adverse events related to treatment. Certainly this argument has been made in studies evaluating the performance of syphilis PoCT in large scale antenatal screening programs,5 but whether those arguments apply to mass screening of the general population in a low prevalence population is not clear. Unfortunately, the test performed poorly in this trial. While there has recently been a significant increase in the diagnosis of syphilis in this region,6 the Youth Health Screen did not reflect this. It is likely that reduced sensitivity at the lower limit of detection accounts for the poor performance of this test. Indeed, further laboratory evaluation of the test using 88 stored known positive and negative sera demonstrated an improved sensitivity and specificity (data not shown). The other disadvantage of this study was the imperative to use serum for diagnosis, as reliable interpretation of duplicate testing at the reference laboratory 900 km away necessitated obtaining serum at the point of care. Use of serum rather than whole blood delays provision of results to patients. However, independent testing demonstrates a loss of sensitivity and specificity with the latter.1,3

Table 2

Discrepant results

Case no.

PoCT (Mt Isa)

PoCT (TSV)

Re-test PoCT (TSV)

EIA S/CO

TPPA

RPR

1 2 3 4 5 6

Neg Neg ND Pos Neg ?Pos

Neg Neg Neg Neg Neg Neg

Neg Neg Neg Pos Pos Neg

8.64 3.37 1.1 10.81 7.66 2.83

ND ND Pos Pos Pos ND

Neg Neg Neg 1:1 1:1 Neg

Cases 1–3, POC-EIA discrepancy; Case 4, POC-POC discrepancy; Cases 5–6, discrepancy in both. Reactive EIA S/CO  1.0. EIA, enzyme immunoassay; ND, not done; Neg, negative; PoCT, point-of-care test; Pos, positive; RPR, rapid plasma reagin; S/CO, signal to cut-off; TPPA, Treponema pallidum particle agglutination; TSV, Townsville.

As with other PoCT, use of the SD Bioline Syphilis 3.0 relies upon training of operators to ensure appropriate performance and interpretation of the test. The authors were not involved in training of the field staff and it is not clear if formal training was documented. While these results do not expose significant errors in field interpretation of PoCT results, it is prudent to bear in mind that PoCT is only as reliable as the training delivered to the operators. To the authors’ knowledge, this is the first time a PoCT for the diagnosis of syphilis has been evaluated in remote communities in Australia. Given the results of this study, it is not possible to recommend the use of the SD Bioline Syphilis 3.0 for screening of a general population where disease prevalence is low. In addition, recent improvements in PoCT for syphilis, including the development of an assay that incorporates treponemal and non-treponemal antibodies for the simultaneous detection of both specific antibody presence and activity of disease,7 may render this particular assay obsolete. Furthermore, an assessment of the risks and benefits of using benzathine penicillin in patients with previously treated disease against the possibility of increased detection of disease is beyond the scope of this study, but warrants further analysis if this test is to be adopted among similar populations. Conflicts of interest and sources of funding: Funding was provided courtesy of the Royal College of Pathologists of Australasia (RCPA) Specialist Training Program (STP) grant. The authors state that there are no conflicts of interest to disclose. Gemma Robertson Glenda Gilmore Robert Norton Pathology Queensland, Qld, Australia Contact Dr G. Robertson. E-mail: [email protected] 1. Li J, Zheng HY, Wang LN, et al. Clinical evaluation of four recombinant Treponema pallidum antigen-based rapid diagnostic tests for syphilis. J Eur Acad Dermatol Venereol 2009; 23: 648–50. 2. Herring AJ, Ballard RC, Pope V, et al. A multi-centre evaluation of nine rapid, point-of-care syphilis tests using archived sera. Sex Transm Infect 2006; 82 (Suppl 5): v7–12. 3. Mabey D, Peeling RW, Ballard R, et al. Prospective, multi-centre clinicbased evaluation of four rapid diagnostic tests for syphilis. Sex Transm Infect 2006; 82 (Suppl 5): v13–6. 4. Lautenschlager S. Diagnosis of syphilis: clinical and laboratory problems. J Dtsch Dermatol Ges 2006; 4: 1058–75. 5. Jafari Y, Peeling RW, Shivkumar S, et al. Are Treponema pallidum specific rapid and point-of-care tests for syphilis accurate enough for screening in resource limited settings? Evidence from a meta-analysis. PLoS One 2013; 8: e54695. 6. The Kirby Institute. HIV, Viral Hepatitis and Sexually Transmissible Infections in Australia. Annual Surveillance Report 2013. Sydney: The Kirby Institute, University of New South Wales, 2013. 7. Castro AR, Esfandiari J, Kumar S, et al. Novel point-of-care test for simultaneous detection of nontreponemal and treponemal antibodies in patients with syphilis. J Clin Microbiol 2010; 48: 4615–9.

DOI: 10.1097/PAT.0000000000000112

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

CORRESPONDENCE

Pure erythroid leukaemia diagnosed on liver biopsy with concurrent haemophagocytic lymphohistiocytosis

Sir, Acute erythroid leukaemia comprises less than 5% of all acute myeloid leukaemias (AML).1,2 Pure erythroid leukaemia (PEL; AML-M6b in the FAB classification) is a rarer subtype of acute erythroid leukaemia. Haemophagocytic lymphohistiocytosis (HLH) is characterised by generalised histiocytic proliferation and a constellation of clinical symptoms and laboratory findings. The two subtypes of HLH are primary (familial/ hereditary) and secondary (acquired) forms that result from the activation of the immune system due to malignancy, infection or autoimmune disorders.3 Pure erythroid leukaemia with concurrent HLH has only recently been addressed in the literature.4,5 Here, we report a unique case of therapy-related pure erythroid leukaemia (t-PEL) with concurrent HLH diagnosed on liver biopsy. A 69-year-old male with a history of follicular lymphoma in 2009 treated with eight cycles of R-CVP (rituximab-cyclophosphamide, vincristine, prednisolone) and prostate cancer in 2011 managed with radiotherapy, developed therapy-related myelodysplastic syndrome (t-MDS) in 2012, which was treated with azacitadine. The patient presented with a syncopal episode and hypotension. On admission, he was noted to have fever, hepatosplenomegaly, new onset pancytopenia (WBC 0.7 K/mL, haemoglobin 8.3 g/dL, platelets 13 K/mL, with no circulating blasts on peripheral blood smear), elevated LDH of 1106 IU/L (normal 98–192 IU/L), and elevated ferritin level of 11960 ng/ mL (normal 22–322 ng/mL). Computed tomography (CT) scan showed multiple subcentimeter hypodense lesions in the

369

liver and spleen. Additional laboratory studies showed total bilirubin 17.0 (normal 0.2–1.4 mg/dL), direct bilirubin 9.8 (normal 0.0–0.3 mg/dL), SGPT 46 (normal 10–40 IU/L), total protein 5.3 (normal 6.5–8.3 g/dL), albumin 2.5 (normal 3.6– 5.0 g/dL), normal levels of SGOT/triglycerides/fibrinogen, and markedly elevated soluble interleukin 2 receptor (sCD25) of 5196 U/mL (normal 406–1100 U/mL). Based on the laboratory studies and abnormal CT scan findings in the liver and spleen, there was clinical concern for hepatic dysfunction due to possible infection or recurrence of follicular lymphoma. A liver biopsy was performed and showed sinusoidal infiltration by medium to large-sized cells with open chromatin and one to multiple nucleoli. A transformation of follicular lymphoma to diffuse large B-cell lymphoma (DLBCL) was the first differential diagnostic possibility; however, the immunohistochemical stains showed that neoplastic cells were negative for CD20. Additional stains CD3, CD5, CD34, myeloperoxidase (MPO), and AE1/AE3 were also negative. The neoplastic cells were positive for LCA (CD45), haemoglobin A and E-cadherin (Fig. 1). A bone marrow aspirate and biopsy with flow cytometry and cytogenetic studies were immediately performed. The bone marrow core biopsy was hypercellular for age (90% cellular) with architectural effacement by a leukaemic infiltrate consisting of medium to large sized cells. The aspirate smears showed a markedly increased blast count (82%). The large blasts were characterised by fine nuclear chromatin, prominent nucleoli, and basophilic cytoplasm with vacuoles consistent with proerythroblasts (Fig. 2A). Multinucleated erythroid precursors with nuclear and cytoplasmic changes similar to proerythroblasts were also seen (Fig. 2B). The very few maturing erythroid precursors present showed dysplastic changes such as, abnormal nuclear contours, nuclear budding, and nuclear/cytoplasmic asynchrony (Fig. 2C). Megakaryocytes and myeloid

A

B

C

D

Fig. 1 (A) Liver biopsy with diffuse hepatic sinusoidal infiltrate (H&E). (B) Blasts with prominent nucleoli within hepatic sinusoids (H&E). (C) Blasts expressing E-cadherin. (D) Blasts expressing hemoglobin A.

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

370

Pathology (2014), 46(4), June

CORRESPONDENCE

A

B

C

D

E

F

G

H

103

11

12

I

101

100

K1

K2

K3

K4

102 CD117 PC5

CD235aPE

102

103

14

13

100

101

100

101 CD71 FITC

102

103

100

101 CD71 FITC

102

103

Fig. 2 (A) Proerythroblasts with cytoplasmic basophilia, vacuoles and multinucleation (B). (C) Dyserythropoiesis was noted. (D–G) Multiple haemophagocytic histiocytes (Wright-Giemsa). Flow cytometric analysis show blasts expressing CD71, dim CD235a (H) and CD117 (I).

cells were markedly decreased. Numerous haemophagocytic histiocytes (showing phagocytosis of nucleated marrow cells) were identified on the aspirate smears (Fig. 2D–G). Flow cytometric analysis of the bone marrow showed increased blasts expressing very dim CD45, CD117 (c-Kit), CD71 (transferrin receptor), and dim CD235a (Glycophorin A) (Fig. 2H–I). Blasts were negative for other B, T, and myeloid markers. Immunohistochemical stains showed that blasts were negative for CD20, CD3, CD5, MPO, pankeratin, CD34 and were positive for LCA. Iron stain showed increased storage iron, and no ringed sideroblasts. Based on these data, a diagnosis of therapy-related AML, subclassified as PEL, (arising out of previous t-MDS) was made. The bone marrow cytogenetic studies showed a complex karyotype: 52, XY, þ1, þ2, þ6, þ8, 9, þ14, add(19)(p13.3), 20, þ21, þ2mar/51, XY, þ1, þ2, þ6, þ8, 9, þ11, add(19)(p13.3), 20, þ21, þmar, 1dmin. Since the patient met the clinical and laboratory diagnostic criteria (fever, splenomegaly, cytopenias, haemophagocytosis in the bone marrow, elevated soluble IL-2 receptor and elevated ferritin),6 a diagnosis of concurrent secondary HLH was also rendered. In spite of making a quick diagnosis due to rapid clinical deterioration, the patient expired 24 h later. Acute erythroid leukaemia is of two types. The more common variant is acute erythroid leukaemia (erythroid/myeloid), where more than 50% marrow cells are erythroid precursors and myeloblasts comprise 20% or more of non-erythroid cells in the marrow. The other, rarer subtype of acute erythroid leukaemia is PEL, showing proliferation of immature (undifferentiated/proerythroblastic-appearing) cells of erythroid lineage (more than 80% of marrow cells) without a significant

myeloblast component.2 It is associated with very poor response and survival with currently available therapeutic modalities.7 Upon application of the current 2008 World Health Organization (WHO) criteria for subclassification of AML, a de novo pure erythroid leukaemia is becoming a very rare diagnosis. Most of the cases that meet morphological criteria for the diagnosis of PEL can be assigned to other AML subcategories like AML with myelodysplasia-related changes or therapyrelated AML.8 In our case, due to previous history of chemo and radiation therapies, the case was classified as therapyrelated AML, although the morphological criteria for PEL are met. Therapy-related PEL has been reported following therapy for a variety of conditions like hypopharyngeal cancer,5 acute lymphoblastic leukaemia,9 chronic lymphocytic leukaemia,10 thyroid papillary carcinoma,11 etc. Of note, t-PEL potentially can arise following chemo and radiation therapy for any condition. Pathologists need to be aware of this rare differential diagnostic possibility, especially on a tissue (e.g., liver) biopsy where DLBCL transformation of follicular lymphoma was a more likely diagnosis. In a case like ours, appropriate work-up helped to make a correct diagnosis. HLH is a rare clinical condition and frequently secondary HLH can be associated with underlying conditions like infection (viral/bacterial/fungal) or malignancy that stimulate the immune system.12 Secondary HLH potentially can be fatal unless the underlying condition is treated. HLH associated with lymphoma and carcinoma has been reported, but HLH associated with AML (especially t-PEL) is extremely rare. To the best of our knowledge this is only the third reported case of HLH associated with t-PEL.4 The association of PEL with HLH and

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

CORRESPONDENCE

liver infiltration is extremely rare and has been reported only once before in the literature.5 In that case report, the liver infiltration was discovered only on autopsy. Our case represents the first case of concurrent HLH and t-PEL where the diagnosis of PEL was made on a liver biopsy. Although infection and DLBCL were more likely clinical possibilities, our case illustrates the need to be aware of the other rarer differential diagnostic possibilities and the need for a complete immunohistochemical and flow cytometric work-up to reach a correct diagnosis in a timely manner. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Woodlyne Roquiz Ameet R. Kini Milind M. Velankar Department of Pathology, Loyola University Medical Center, Maywood, IL, USA Contact Dr M. M. Velankar. E-mail: [email protected] 1. Arber DA, Brunning RD, Orazi A, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissue. Lyon: IARC Press, 2008; Chapter 6, Acute myeloid leukaemia and related precursor neoplasms. 2. Mihova D, Zhang L. Acute erythroid leukemia: a review. N A J Med Sci 2012; 5: 110–8. 3. Yamazaki S, Nakamura F, Nasu R, et al. Haemophagocytic lymphohistiocytosis is a recurrent and specific complication of acute erythroid leukaemia. Br J Haematol 2011; 153: 669–72. 4. Kitagawa J, Hara T, Tsurumi H, et al. Pure erythroid leukemia and hemophagocytosis. Inter Med 2009; 48: 1695–8. 5. Funakoshi Y, Matsuoka H, Yamamoto K, et al. Therapy-related pure erythroid leukemia with hepatic infiltration and hemophagocytic syndrome. Intern Med 2011; 50: 3031–5. 6. Henter JI, Horne AC, Arico M, et al. HLH-2004: Diagnostic and therapeutic guidelines for hemophagocytic lymphohistiocytosis. Pediatr Blood Cancer 2007; 48: 124–31. 7. Santos FP, Bueso-Ramos CE, Ravandi F. Acute erythroleukemia: diagnosis and management. Expert Rev Hematol 2010; 3: 705–18. 8. Liu W, Hasserjian RP, Hu Y, et al. Pure erythroid leukemia: a reassessment of the entity using the 2008 World Health Organization classification. Mod Pathol 2011; 24: 375–83. 9. Xu M, Finn LS, Tsuchiya KD, et al. Pure erythroid leukemia following precursor B-cell lymphoblastic leukemia. Pediatr Dev Pathol 2012; 15: 76–8. 10. Sadrzadeh H, Hasserjian R, Fathi AT. Pure erythroid leukemia evolving from a therapy-related myelodysplastic syndrome secondary to treatment for chronic lymphocytic leukemia. Am J Hematol 2013; 88: 240–1. 11. Trikalinos NA, Chen Q, Ning Y, et al. Unbalanced 11;18 translocation in an acute erythroid leukemia after radioactive iodine therapy. Cancer Genet 2013; 206: 252–5. 12. Higa B, Velankar MM. Familial haemophagocytic lymphohistiocytosis in twin infants. Pathology 2013; 45: 83–5.

DOI: 10.1097/PAT.0000000000000108

Diagnosis of diabetes in anaemias and haemoglobinopathies: salutary lessons for improved reporting and patient empowerment Sir, We bring to attention an issue raised by a lay reader at Lab Tests Online AU,1 a website (funded by the Australian

371

Government Department of Health Quality Use of Pathology Program) that provides information about pathology testing for the Australian general public. The issue is the potential for and consequences of errors in the interpretation of HbA1c due to conditions that can result in lower than expected HbA1c values because of shortened red cell lifespan. Our reader had a diagnosis of diabetes made after he developed symptoms and had elevated blood glucose values. He was initially started on medication, which was later stopped because his HbA1c values were not elevated. He continued home blood glucose monitoring, which showed high blood glucose values over many years. On repeatedly reporting this to a series of doctors (both general practitioners and specialists) the invariable response was to not take at face value these high home blood glucose results, as repeated laboratory HbA1c measurements were always low. He would then be reassured that he had excellent control of his diabetes. It took 25 years of illness before it was finally discovered that he also had a low grade haemolytic anaemia, which explains his falsely low HbA1c values. Fructosamine levels confirmed poor diabetic control. The patient described his experiences in patient support newsletters2,3 but these have so far failed to reach the major medical literature. New national guidelines have recently been published, permitting the use of HbA1c for diagnosing diabetes.4 These guidelines state that ‘caution is needed in interpreting HbA1c results in the presence of conditions affecting red blood cells or their survival times, such as haemoglobinopathies or anaemia’. Based on our informant’s personal experience, it seems that these limitations of the HbA1c test and their implications are not universally understood. The simplest solution would be for the laboratory to calculate and report estimated average glucose (eAG) whenever they report HbA1c.5 In keeping with recommended practice,4 the reported eAG would not be intended for patient management per se; instead, the report should state that the eAG is solely for comparison with the average of actual glucose measurements (preferably from the patient’s own home blood glucose monitoring) as a check that the HbA1c is truly reflecting the patient’s diabetes control. Our patient’s HbA1c was between 20–30 mmol/mol (implying an eAG of 4–5 mmol/L) despite home blood glucose levels greater than 10 mmol/L. A more sophisticated approach would be for the pathologist’s or the doctor’s information system to implement an algorithm that detects large discrepancies between blood glucose values and contemporaneous HbA1c measurements, as the HbA1c may be incorrect due to the presence of various anaemias (especially haemolytic anaemia), haemoglobinopathy or significant liver or renal disease. Advanced systems could also check the patient’s laboratory reports for haematology and chemistry results that might indicate a possible cause for the discrepancy. Such a system would probably have detected this patient’s problem as he sometimes had marginally low haemoglobin and marginally high bilirubin values, and he was subsequently also shown to have low haptoglobin and an elevated reticulocyte count. However, as far as we have been able to determine, none of this puzzle was put together by our patient’s medical attendants over some 20 years, most probably due to the separate reporting of results relevant to each discipline, chemistry results with chemistry, and haematology with haematology. ‘Smart reporting’ has the potential to add value to the pathology

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

372

Pathology (2014), 46(4), June

CORRESPONDENCE

consultation through integration of all available informatics resources. Whichever approach is chosen by the profession, ‘doing nothing’ is unlikely to be a satisfactory solution, as evidenced by this patient’s 25 year delay in having his diagnosis made. Finally, well-informed patients can form powerful partnerships with their doctors. Patients are becoming better informed because of internet resources like Lab Tests Online1 providing information aimed at the general public. A disempowered patient places the sole burden of diagnosis on the doctor, and a wrong or delayed diagnosis due to information overload (similar to our present case) can only increase the medico-legal risk for the doctor, as well as lessen the quality of care the doctor can offer the patient. In contrast, a knowledgeable and motivated patient competent in home glucose measurements can facilitate the monitoring efforts required for optimal patient care. Acknowledgements: We thank the patient who provided his valuable feedback and observations outlined in this correspondence, and Peter Graham who assisted with investigation of past pathology results.

Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Bruce Campbell1 Leslie Burnett1,2,3 Lab Tests Online AU, Sydney, 2PaLMS Pathology North, NSW Health, Royal North Shore Hospital, St Leonards, and 3 Sydney Medical School, Sydney, NSW, Australia 1

Contact Professor L. Burnett. E-mail: [email protected] 1. Lab Tests Online (Australasia). Cited Jul 2013. http://www.labtestsonline. org.au/ 2. Diabetes Australia. Living With Diabetes. December 2008. 3. Australian Diabetes Council. Diabetes Now. April 2011. 4. d’Emden MC, Shaw JE, Colman PG, et al. The role of HbA1c in the diagnosis of diabetes mellitus in Australia. Med J Aust 2012; 197: 220–1. 5. Nathan DM, Kuenen J, Borg R, Zheng H, Schoenfeld D, Heine RJ. the A1cDerived Average Glucose (ADAG) Study GroupTranslating the A1C assay into estimated average glucose values. Diabetes Care 2008; 31: 1473–8.

DOI: 10.1097/PAT.0000000000000103

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