Cystic alveolar adenoma: an unusual clinical presentation of a rare lung neoplasm

Pathology (January 2015) 47(1), pp. 71–95

CORRESPONDENCE Malignant extraneural soft tissue perineurioma with striking microvascular proliferation

Sir, Perineurioma is an uncommon variant of peripheral nerve sheath tumour composed predominantly of perineurial cells. Malignant perineurioma is exceedingly rare. A case of malignant extraneural soft tissue perineurioma is described which arose on the lower left leg of an 83-year-old woman. There was striking microvascular proliferation, which seems to be a novel feature in this case. The palpable mass of unclear duration, measured 53
36
7 mm at ultrasound and was determined to be solid, hypoechoeic, heterogeneous, and to reside in the deep subcutaneous layer, superficial to fascia and adjacent to the tibialis anterior muscle. The surrounding adipose tissue was reported to appear oedematous and hyperaemic. Diagnostic and attempted excision biopsy was performed and the laboratory received three nodular pieces of grey/white rubbery mucoid tissue and adipose tissue measuring 25
20
8 mm in aggregate dimension following fixation in 10% neutral buffered formalin. Histologically (Fig. 1) there were multiple fragments of a nodular, variably cellular myxoid tumour, which was mostly well defined, but focally intermingled with adipose tissue. The tumour was composed of spindle cells with elongated nuclei and delicate long thin biphasic cytoplasmic processes arranged in a fascicular and somewhat wavy and whorling lamellar fashion in a predominantly myxoid background. There was variable nuclear atypia ranging from mild through to marked with highly pleomorphic and multinucleated tumour cells,

many of the latter aggregating around prominent small proliferating vessels. Mitoses, including occasional atypical forms, numbered up to 11/10HPF. Component vessels were of various sizes, but there were no arcuate arrangements. No necrosis was seen. The component of the tumour that infiltrated adipose tissue showed marked pleomorphism. No accompanying nerve was seen. Immunohistochemical markers performed (Fig. 2) demonstrated that the delicate elongated tumour cells expressed epithelial membrane antigen (EMA) and CD34, although the pleomorphic perivascular cells were non-reactive with these immunomarkers. The MIB-1/Ki-67 determined proliferation index (PI) was approximately 70%. Other immunomarkers performed were non-reactive (S100, SOX-10, HMB45, Melan-A, cytokeratins AE1/AE3 and CK8/18, actin, smooth muscle actin, desmin, CD31 and mucin-4). An external laboratory kindly performed human erythrocyte glucose transporter 1 (GLUT1) and tight junction associated protein (Claudin-1) immunohistochemical markers. They were diffusely and focally expressed, respectively. In view of the histological appearance and immunoprofile, the tumour was believed to be of perineurial origin and because of the pleomorphism, mitotic activity and high proliferation index, was diagnosed as a malignant extraneural soft tissue perineurioma (of ordinary type), and of intermediate grade according to the FNCLCC grading system, and WHO grade II. Complete and wide re-excision was recommended. The patient was referred to a specialist sarcoma surgeon. Re-excision was performed 31/2 weeks following the initial biopsy. The specimen comprised skin with underlying subcutis, fascia and skeletal muscle and measured 115
30
30 mm deep. Residual ill-defined myxoid appearing tumour was noted up to 25 mm in size within the subcutis (Fig. 3A). Tissue was

Fig. 1 Histological images of malignant perineurioma (H&E). (A) Elongated spindle cells in a pale myxoid background. (B) Pleomorphic cells in a perivascular arrangement associated with microvascular proliferation. (C) Pleomorphism and mitoses including atypical forms. (D) Permeation of adipose tissue.

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.

72

CORRESPONDENCE

Pathology (2015), 47(1), January

Fig. 2 Immunohistochemistry of malignant perineurioma. (A) EMA expression by spindle cells with poor expression by pleomorphic cells. (B) CD34 expression by spindle cells (positive internal control in endothelial cells). (C) S100 not expressed (positive internal control in adipose tissue). (D) MIB-1/Ki-67 proliferation index 70%.

taken for electron microscopy and the remaining specimen was fixed in 10% neutral buffered formalin. Histologically (Fig. 3B), the tumour had a similar morphological and immunohistochemical profile as seen in the initial biopsy, with a striking microvascular proliferation and adjacent pleomorphic cells. Tumour cells permeated fat within a myxoid background and in areas closely mimicked myxofibrosarcoma. It extended to several margins. Electron microscopy was reported as being compatible with a malignant peripheral nerve sheath tumour, of no specific type, perhaps reflecting tissue sampling or the degree of tumour differentiation. A diagnosis of malignant soft tissue perineurioma was maintained although it was acknowledged that myxofibrosarcoma could easily be entertained as an alternative diagnosis, were there not such strong EMA (and CD34) expression. Further re-excision was performed and no further lesion was identified. The perineurium of peripheral nerve fascicles comprises perineurial cells with distinct ultrastructural features such as distinct, thin, non-branching cytoplasmic processes coated by an external lamina, and contain numerous pinocytic vescicles, actin and vimentin filaments, but few organelles. Their thin

cytoplasmic processes are joined at their ends by tight junctions.1 The formation of tight junctions is unique to perineurial cells and endothelial cells.2 They are immunoreactive for vimentin and particularly epithelial membrane antigen (EMA), but also consistently express GLUT1 and Claudin1.1–7 The origin of the perineurial cell is debated, but origin from fibroblasts, Schwann cells or arachnoid cells have all been postulated.1 Reactive or neoplastic perineurial cells have been demonstrated ultrastructurally and immunohistochemistically in a number of pathological processes including reactive lesions (Morton’s neuroma, Pacinian neuroma) and neoplastic lesions, as a component of neurofibroma, dermal nerve sheath myxoma/ neurothekeoma and perineurioma.1 The immunophenotype of perineurioma parallels that of the normal perineurial cell.2 Perineurioma is generally benign and divided into intraneural perineurioma (previously confused with localised hypertrophic neuropathy)8 and soft tissue forms, which are unrelated to nerves (extraneural soft tissue perineurioma, ESTP), each of which generally behaves in a benign fashion. Intraneural perineurioma is composed of perineurial cells within the

Fig. 3 Excision biopsy of remaining malignant perineurioma. (A) Macroscopic image of cut surface demonstrating a mucoid appearance. (B) The microscopic findings were similar to the diagnostic biopsy with variable pleomorphism, microvascular proliferation and mitotic activity (H&E).

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

14

15

16

Rosenberg et al.17 (2002)

Yamazaki et al.18 (2005) Suster et al.19 (2005)

31

32

Sato et al.11 (2008)

Adachi et al.20 (2010)

17–30

76 / F

13

Za´mecnı´k et al.16 (2002)

Hornick et al.6 (2005)

24 / M

12

Fukunaga et al.15 (2001)

7/F

60 / M

28–79 / Fx8, Mx6

70 / M

25 / F

82 / F

63 / M

11

Za´mecnı´k et al.14 (1999)

39 / M

76 / F

10

9

55 / F

8

Karaki et al.13 (1999)

Hirose et al. (1998)

12

26 / M

52 / M

11 / M

5

6

25 / F

4

7

69 / M

83 / M

2

Hirose et al.10 (1998)

52 / M

Age (years) / Sex

3

1

Case no.

Tongue / Relatively well circumscribed but muscle infiltration

Dorsum of left wrist / Adherent to muscle which it infiltrated

Right lateral thigh / Circumscribed but irregularly infiltrated skeletal muscle Extremeties and trunk / Infiltrative margins a feature

Left arm (D/SC) / Partially and infiltrative Tongue

Left groin (D/SC) / Well circumscribed

Pelvis / Ill defined, non-encapsulated

Left forearm / Well circumscribed

Pancreas / Ill defined posteriorly invading pancreatic parenchyma Right para-vertebral, T6 / UK

Right thigh / UK

Left back / UK

Left back / UK

Right face / UK

Mediastinum / UK

Retroperitoneum / UK

Left thigh / UK

Left para-vertebral (IM) / Well circumscribed

Tumour location / Circumscription

Yes, or mixed sometimes with abrupt transition / Scattered pleomorphic cells Mixed hyper and hypocellular / No (mild atypia only) Uniformly hypercellular / No

Haphazard fascicles / Yes (mild-moderate)

Low / No in primary; High / Yes in recurrence and mets Mixed cellular and myxoid / Yes, lipoblast-like cells present Spindle cell and cellular *round cell solid epithelioid components NS / Yes ‘larger atypical cells scattered throughout’ Mixed with myxoid stroma / Yes NS / Yes

‘Slight’

Low / No

High / Few

Low / No

Low / No

High / Few

High / Few

High but variable / relatively uniform, but occasional large bizarre cells High / Few

Cellularity / Pleomorphism

Clinicopathological features of atypical and malignant perineurioma in the literature

Hirose et al.9 (1989)

Reference

Table 1

18mm

28 mm

18–120 mm

65 mm

30 mm

30 mm

20 mm

120 mm (recurrence)

40 mm

60 mm

45 mm

70 mm

65 mm

110 mm

15 mm

50 mm

300 mm

30 mm

65 mm

Size

No

No (central infarction)

No (central infarction in some)

No

No

NS

No

Yes

No

No in primary; Yes in recurrence and mets

Yes, 50%

No

Yes

No

No

Yes

Yes

Yes

UK

Necrosis

‘Conspicuous’

5/10 HPF

0–7/30 HPF (mean 1)

2/10 HPF

‘Frequent’

7/10 HPF

1/20 HPF

3/10 HPF

‘Numerous and often atypical’

No in primary; ‘Occasional’ in recurrence and mets

‘Scattered throughout’

‘Low level’ in LG

‘Frequent’ in HG

1–85 (mean 22)

‘Numerous’

Mitoses

UK

18.4%

UK

UK

40%

UK

2%

UK

UK

UK

UK

6.9%

29.8%

9.8%

10.7%

9%

14.9%

15.9%

UK

Ki67/ MIB-1 PI

Atypical

Malignant / LG

Atypical
14

Malignant / LG

Malignant / NS

Malignant / NS

Atypical

Malignant / HG

Malignant / NS

Malignant / LG to HG transformation

Malignant / LG (?AT) Malignant / NS

Malignant / LG (?AT) Malignant / HG

Malignant / LG (?AT)

Malignant / HG

Malignant / HG

Malignant / HG

Malignant / NS

Classification / Grade

Resected

Resected

? all resected

Resection

Resection

Resection

Excision biopsy followed by autopsy at 13 mo Resection

Resection

Resection

Resection

Resection þ Chemo Resection þ Chemo þ RT Resection þ RT Resection þ Chemo þ RT Resection

Resection þ RT Resection

Resection, Chemo, cerebral RT

Treatment

ANED 12 mo

ANED 12 mo

Recurrence twice
1 then ANED 134 mo, remainder no follow-up or ANED 12–90 mo

NS

ANED >24 mo

ANED 8 mo

ANED 32 mo

Recurrence of both components. Metastasis by round cell component. DOD 13 mo

ANED 24 mo

Recurrence and multiple mets at 10 y, DOD 10 y

Recurrence
1, ANED 75 mo Cerebral, lung soft tissue mets, AWD 88 mo Recurrence
1, ANED 55 mo ANED 13 mo

Recurrence
1, ANED 98 mo Recurrence
1, DUC 28 mo Metastasis to trapezius muscle, AWD 68 mo ANED 56 mo

Lung and brain mets, AWD 20 mo

Clinical outcome

CORRESPONDENCE

73

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

74

Pathology (2015), 47(1), January * This component showed moderate atypia. ANED, alive no evidence of disease; AT, atypical; AWD, alive with disease; Chemo, chemotherapy; D, dermis; DOD, dead of disease; DUC, dead of unrelated cause; HG, high grade; HPF, high power field; IM, intramuscular; LG, low grade; M, malignant; Met, metastasis; Mo, months; NS, not stated; R, recurrence; RT, radiotherapy; SC, subcutaneous; UK, unknown.

Malignant / FNCLCC intermediate grade 70% 11/10 HPF No 25 mm initial excision Re-excision ¼ further 25 mm Variable / Yes (focally marked) Lower left leg (SC) / Infiltrated adipose tissue 35 Present case (2014)

83 / F

Recurrence Cellular / Yes, moderate to marked atypia, tumour giant cells 34 Rekhi et al.22 (2013)

45 / M

Resected

Recurrence and multiple metachronous ‘benign’ perineuriomas, ?mixed grade metastases; DOD at 160 mo Resected Malignant / HG 50% 5–6/10 HPF Yes

Recurrence at 9 mo, ANED 24 mo Resected
3 þ RT Malignant / HG 10% ‘Average’ 5/1/10 HPF Yes

152 mm (recurrence ¼ 180 mm) Recurrence 80 mm (original UK) High / ‘Monomorphous’

Buttock / Infiltrated around sciatic nerve and its branches Right thigh / UK 42 / F 33 Mitchell et al.21 (2012)

Case no. Reference

Table 1

(Continued )

Age (years) / Sex

Tumour location / Circumscription

Cellularity / Pleomorphism

Size

Necrosis

Mitoses

Ki67/ MIB-1 PI

Classification / Grade

Treatment

Clinical outcome

CORRESPONDENCE

confines of a nerve and has a characteristic pseudo-onion bulb pattern.7 ESTP is a peripheral nerve sheath neoplasm that most often presents as a painless solitary nodule or mass in the subcutaneous tissues of the trunk or limbs but may occur at other sites including viscera, and shows no age or sex predilection. The usual or ‘ordinary’ forms of ESTP are macroscopically white/ grey, well circumscribed but not encapsulated and have a firm consistency and wide size range from 0.3 to 20 cm. Histologically, they are composed of elongated tumour cells with wavy nuclei and cytoplasmic characteristics vary, but at least focally one finds characteristic spindle cells with tapering nuclei and very long, thin, widely separated bipolar cytoplasmic processes. Cells may be arranged in lamellar, perivascular whorling or storiform arrangements. The stromal content varies in amount and quality from collagenous to myxoid. Mitotic figures and tumour necrosis are not features of benign lesions. Less common variants of ESTP are the sclerosing perineurioma and the reticular or retiform perineurioma.6,7 The immunohistochemical profile of ESTP varies between the various types. The ordinary type classically expresses EMA, collagen IV, laminin, vimentin, GLUT1 and Claudin1, and lacks S100 expression. Sclerosing perineurioma, which contains prominent collagenous stroma, typically lacks CD34 expression, but often shows expression of actin and alpha smooth muscle actin and sometimes expresses keratins. Reticular perineurioma, which has a characteristic net-like growth pattern, also may show focal expression of CD34, desmin and cytokeratin.7 CD34 expression may be seen in up to two-thirds of cases of ESTP,6 the presence of which may reflect the occurrence of endoneurial fibroblasts in some tumours or transitional forms.5 Differing antibody types, immunohistochemical protocols and case selection are also suggested as explanations as to why reported rates of CD34 expression vary.7 Selection criteria may also account for the 5% rate of S100 expression in one series.6 However, the occurrence of cells intermediate between Schwann and perineurial cells may also explain this finding.9,10 The expression of immunomarkers GLUT1 and Claudin-1 is not unique to perineurioma and occurs in other peripheral nerve sheath tumours (within neurofibromas and predominantly capsular in schwannoma) reflecting the presence of perineurial cells in these tumours. However, those lesions typically show S100 positivity and generally lack EMA expression.2,4,5 Cytogenetically, intraneural and ESTP have both been shown to harbour chromosome 22 abnormalities. Deletions and point mutations of the NF1 gene are commonly found, although an association with NF1 is not typical. Chromosome 10 and 13 abnormalities have also been recorded.7 Loss of chromosome 17 was previously noted on interphase fluorescence in situ hybridisation in a low grade malignant ESTP.11 An extensive search of the English literature found 34 cases of perineuriomas with varying atypical features6–23 (Table 1). Some cases have been regarded as frankly malignant, termed ‘malignant perineurioma’, a variant of malignant peripheral nerve sheath tumour (MPNST) with features including pleomorphism or anaplasia, frequent mitoses, tumour necrosis, and/ or metastasis.9,10,12,15,17,22 Hirose et al.10 concluded that 4% of MPNST show perineurial differentiation. They stratified lesions into high or low grade based on tumour cellularity, often with whorled or storiform growth patterns, degree and extent of cytological atypia, presence of necrosis and mitotic activity. They stated that as experience was limited, they could

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

CORRESPONDENCE

not exclude that ‘low grade’ tumours were not better regarded as ‘cellular’ or ‘atypical’. Low grade or ‘atypical’ lesions may be of no clinical consequence.6,16 However, low to high grade malignant transformation has been recorded.13 The clinical behaviour of low grade or ‘atypical’ perineurioma remains unclear because they are uncommonly documented and reported follow-up is often short (Table 1). The differential diagnosis of perineurioma depends upon its type, intraneural or ESTP, and upon the degree of atypia.7 The current case is a malignant form of ordinary ESTP, the differential diagnosis of which can include DFSP, neurofibroma, solitary fibrous tumour, fibromatosis, smooth muscle tumours, low grade fibromyxoid sarcoma (LGFMS), low grade myxofibrosarcoma (LGMFS), and schwannoma.7 They are all generally EMA negative, but CD34 expression is seen in the first three. In difficult cases, GLUT1 and Claudin-1 are useful. EMA and Claudin-1 expression has previously been demonstrated in perineurioma-like LGFMS.23 However, MUC4 immunostaining and/or detection of a rearrangement of FUS is useful for the diagnosis of LGFMS. S100 expression would be expected in schwannoma. Cytogenetics and electron microscopy can also be utilised. In the current case the main differential diagnosis considered was of LGMFS in view of the myxoid stroma and cellular atypia in a lesion arising superficially on an extremity of an elderly individual. However, there were characteristic morphological features of perineurioma such as cells with slender bipolar cytoplasmic processes, diffuse EMA and GLUT1 expression, and focal Claudin-1 expression. There was striking microvascular proliferation, which may be a novel feature. Prognosis is unclear in the current case and the follow-up time is short. Although pleomorphism, multinucleation and infiltration of surrounding tissues are not necessarily adverse prognostic features,6 there were frequent mitoses including atypical forms and the MIB-1/Ki-67 PI was 70% so the lesion was highly proliferative and marked pleomorphism was seen, particularly in cells permeating adjacent adipose tissue. As such, these are disquieting features and prognosis must be guarded. Acknowledgements: Enormous thanks are owed to Peter Robbins for generously performing the GLUT1 and Claudin-1 immunohistochemistry. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. M. J. Wilsher1 A. Mahar2 R. A. Boyle3 S. F. Bonar1

75

3. Ariza A, Bilbao JM, Rosai J. Immunohistochemical detection of epithelial membrane antigen in normal perineurial cells and perineurioma. Am J Surg Pathol 1988; 12: 678–83. 4. Yamaguchi U, Hasegawa T, Hirose T, et al. Sclerosing perineurioma: a clinicopathological study of five cases and diagnostic utility of immunohistochemical staining for GLUT1. Virchows Arch 2003; 443: 159–63. 5. Hirose T, Tani T, Shimada T, Ishizawa K, Shimada S, Sano T. Immunohistochemical demonstration of EMA/GLUT1-positive perineurial cells and CD34-positive fibroblastic cells in peripheral nerve sheath tumors. Mod Pathol 2003; 16: 293–8. 6. Hornick JL, Fletcher CDM. Soft tissue perineurioma. Clinicopathologic analysis of 81 cases including those with atypical histologic features. Am J Surg Pathol 2005; 29: 845–58. 7. Macarenco RS, Ellinger F, Oliveria AM. Perineurioma. A distinctive and underrecognized peripheral nerve sheath neoplasm. Arch Pathol Lab Med 2007; 131: 625–36. 8. Tsang YW, Chan JKC, Chow LTC, Tse CCH. Perineurioma: an uncommon soft tissue neoplasm distinct from localized hypertrophic neuropathy and neurofibroma. Am J Surg Pathol 1992; 16: 756–63. 9. Hirose T, Kudo E, Teramae T, Higasa Y, Hizawa K. Malignant peripheral nerve sheath tumor (MPNST) showing perineurial cell differentiation. Am J Surg Pathol 1989; 13: 613–20. 10. Hirose T, Scheithauer BW, Sano T. Perineurial malignant peripheral nerve sheath tumor (MPNST): a clinicopathologic, immunohistochemical, and ultrastructural study of seven cases. Am J Surg Pathol 1998; 22: 1368–78. 11. Sato K, Ueda Y, Miwa S, Yokogawa A, Ozaki M, Katsuda S. Low-grade malignant soft-tissue perineurioma: interphase fluorescence in situ hybridization. Pathol Int 2008; 58: 718–22. 12. Hirose T, Maeda T, Furuya K, Kiyasu Y, Kawasaki H. Malignant peripheral nerve sheath tumor of the pancreas with perineurial cell differentiation. Ultrastruct Pathol 1998; 22: 227–31. 13. Karaki S, Mochida J, Lee YH, Nishimura K, Tsutsumi Y. Low-grade malignant perineurioma of the paravertebral column, transforming into a high-grade malignancy. Pathol Int 1999; 49: 820–5. 14. Za´mecnı´k M, Michal M. Malignant peripheral nerve sheath tumor with perineurial cell differentiation (malignant perineurioma). Pathol Int 1999; 49: 69–73. 15. Fukunaga M. Unusual malignant perineurioma of soft tissue. Virchows Arch 2001; 439: 212–4. 16. Zamecnik M, Koys F, Gomolcak P. Atypical cellular perineurioma. Histopathology 2002; 40: 296–9. 17. Rosenberg AS, Langee CL, Stevens GL, Morgan MB. Malignant peripheral nerve sheath tumor with perineurial differentiation: ‘malignant perineurioma’. J Cutan Pathol 2002; 29: 362–7. 18. Yamazaki H, Tsukinoki K, Shimamura K, Kaneko A. Malignant peripheral nerve sheath tumor with perineurial cell differentiation arising from the tongue. Oral Oncol Extra 2005; 41: 77–80. 19. Suster D, Plaza JA, Shen R. Low-grade malignant perineurioma (perineurial sarcoma) of soft tissue: a potential diagnostic pitfall on fine needle aspiration. Ann Diagn Pathol 2005; 9: 197–201. 20. Adachi S, Doi R, Mitani K, et al. Atypical soft tissue perineurioma in the tongue of a young girl. Pathol Int 2010; 60: 787–91. 21. Mitchell A, Scheithauer BW, Doyon J, Berthiaume MJ, Isler M. Malignant perineurioma (malignant peripheral nerve sheath tumor with perineural differentiation). Clin Neuropathol 2012; 31: 424–9. 22. Rekhi B. Perineurial malignant peripheral nerve sheath tumor in the setting of multiple soft tissue perineuriomas: a rare presentation of an uncommon tumor. J Cancer Res Ther 2013; 9: 131–4. 23. Thway K, Fisher C, Debiec-Rychter M, Calonje E. Claudin-1 is expressed in perineurioma-like low-grade fibromyxoid sarcoma. Hum Pathol 2009; 40: 1586–90.

DOI: 10.1097/PAT.0000000000000199

1

Douglass Hanly Moir Pathology, Macquarie Park, 2Royal Prince Alfred Hospital, Department of Pathology, Camperdown, and 3Boyle Orthopaedics, Mosman, NSW, Australia Contact Dr M. J. Wilsher. E-mail: [email protected] 1. Erlandson RA. The enigmatic perineurial cell and its participation in tumors and in tumor-like entities. Ultrastruct Pathol 1991; 15: 335–51. 2. Folpe AL, Billings SD, McKenney JK, Walsh SV, Nusrat A, Weiss SW. Expression of Claudin-1, a recently described tight junction-associated protein, distinguishes soft tissue perineurioma from potential mimics. Am J Surg Pathol 2002; 26: 1620–6.

Osteosarcoma arising from metastatic low grade endometrial stromal sarcoma Sir, As a rare malignant mesenchymal neoplasm of the uterus, endometrial stromal sarcoma (ESS) constitutes only about 0.5% of all uterine malignancies and 10% of all uterine sarcoma, the median age of those affected being 52 years.1 ESS has traditionally been classified into high and low grade

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

76

CORRESPONDENCE

variety, based on tumour morphology, mitotic activity, cellularity and extent of necrosis. In general, high grade ESS behaves aggressively and the clinical outcome is poor. In contrast, the prognosis of low grade ESS is excellent; the 5- and 10-year actuarial survivals for patients with stage I tumours are 98% and 89%, respectively.2 However, it is characterised by late recurrences even in patients with stage I disease and thus long term follow-up is mandatory. Recurrences develop in about one-third of patients, most commonly in the pelvis and abdomen,3,4 and less frequently in the lung and vagina.2 It is also worth noting that in low grade ESS, local recurrences and distant metastases have an insidious growth, not infrequently occurring after long disease-free periods, even more than two or even approaching three decades after initial diagnosis and surgical resection.5 Bone metastases are very rare in stage 1 low grade ESS. The present report describes a patient presenting as secondary high grade osteosarcoma arising from the left humerus due to involvement by metastatic low grade ESS after a 22 year disease-free interval from resection of the stage 1 primary uterine tumour. A 54-year-old woman attended our orthopaedic outpatient clinic because of left shoulder pain associated with stiffness and limitation of movement for 3 months. There was no antecedent trauma to the region. She previously had low grade ESS confined to the uterus for which total hysterectomy had been performed 22 years ago, requiring no post-operative adjuvant treatment. She had been well since the operation and 5 years prior had been dismissed from follow-up on the belief that her disease was cured. Physical examination revealed no mass in the left shoulder but tenderness in the supraspinatus region and limitation of movement. There was no neurovascular deficit.

Pathology (2015), 47(1), January

Plain radiograph of the left shoulder revealed a densely osteoblastic lesion involving the proximal epiphysis and metaphysis of the left humerus with permeative infiltration into the diaphysis. There was invasion through the cortex with prominent periosteal reaction (Fig. 1A). Magnetic resonance imaging (MRI) showed widespread destruction of the left proximal humerus by the tumour which invaded through the cortex and infiltrated extensively into the surrounding soft tissue. The neurovascular bundle and shoulder joint appeared unaffected. Positron emission tomography – computed tomography (PETCT) depicted the left proximal humerus lesion as the only abnormality. Since the history of hysterectomy for low grade ESS had escaped the attention of the radiologist and orthopaedic surgeon alike, and in view of the solitary nature of the lesion, the presumptive diagnosis was primary malignant tumour of the proximal left humerus. Biopsy of the lesion showed sheets of closely packed small polygonal to plump spindle tumour cells resembling proliferative phase endometrial stromal cells, among which were rich networks of thin-walled arteriolar type vessels (Fig. 1B). The histology of the tumour was a morphological match for the previous hysterectomy specimen harbouring the ESS. It showed permeative infiltration into adjacent native bony trabeculae with extensive reactive bone formation. The tumour cells stained intensely for CD10 (Fig. 1C) while endothelial markers (Factor 8, CD31 and CD34) highlighted the endothelial cells of the arterioles. Many of the reactive bone fragments were in the form of small irregular globules and surrounded by rims of tumour cells (Fig. 1D). Nuclear pleomorphism was mild, mitoses less than 1/10 HPF and tumour necrosis was not seen. The tumour cells stained negatively for oestrogen and

Fig. 1 (A) Plain radiograph of the left shoulder showing a densely osteoblastic lesion involving the proximal epiphysis and metaphysis of the left humerus with permeative infiltration into the diaphysis. There is invasion through the cortex with prominent periosteal reaction. (B) Biopsy of the lesion shows closely packed small polygonal to plump spindle tumour cells resembling proliferative phase endometrial stromal cells. A rich network of thin-walled arteriolar type vessels is seen (H&E). (C) The tumour cells stain intensely while the endothelial cells are negative for CD10. (D) The tumour shows permeative infiltration among the native bony trabeculae with extensive reactive bone formation (H&E).

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

CORRESPONDENCE

progesterone receptors (previous hormone receptor status of the uterine pathology was not determined), S100, desmin, smooth muscle actin, melan A, epithelial membrane antigen and AE1/3. The features were those of metastatic low grade ESS with florid reactive bone formation which accounted for the densely osteoblastic radiographic features. As CT guided biopsy revealed histological features of metastatic low grade ESS, segmental resection of the left proximal humerus was performed. This showed a bony, hard tumour involving the entire epiphysis and metaphysis with permeative infiltration into the diaphysis (Fig. 2A). There was invasion through the cortex in the metaphyseal region with widespread infiltration into the surrounding soft tissue. The resection margins appeared uninvolved. The tumour exhibited mostly similar histological features as those in the biopsy, consisting of low grade ESS demonstrating permeative infiltration among the native marrow trabecular bone with exuberant irregular globular reactive bone formation. In the midst of florid reactive bone formation in the medial metaphyseal region at the periphery of the main tumour, however, was an area of high grade osteosarcomatous transformation (Fig. 2B). The tumour cells assumed marked nuclear pleomorphism and hyperchromatism (Fig. 2C). They showed direct deposition of lace-like osteoid and stained negatively for CD10. This high grade osteosarcomatous component accounted for about 25% of the entire tumour and demonstrated focal lymphovascular infiltration. The shoulder joint and distal resection margins were free from tumour involvement. The overall features were those of secondary high grade osteosarcoma arising from the left humerus due to involvement by metastatic low grade ESS. The patient was given post-operative chemotherapy. However, the patient suffered from severe epigastric pain and depression, and declined further chemotherapy. She remained well and her chest radiography was normal 6 months after surgery. Unfortunately, she developed left pleuritic chest pain 2 months afterwards and chest radiograph revealed a left lower zone nodule. She resumed chemotherapy with some initial response but finally succumbed from multiple lung and bone metastasis 24 months after operation. Bone metastases are distinctly uncommon in low grade ESS. To the best of our knowledge, excluding those bony involvements secondary to local infiltration and others without precise documentation, only five cases have been reported in the English literature. The metastases involve mainly the spine (4 cases)6–9 and the skull (1 case).10

77

Our patient presented with left shoulder pain and limitation of movement due to solitary left humerus metastasis 22 years after the initial operation. Although she was symptomatic for just 3 months, the extent of involvement by the similarly low grade ESS in the humerus, together with the florid reactive bone formation in the form of irregular spherules, attested to the chronicity of the lesion. This irregular spherule bone formation is typical of reactive bone formation in slow growing indolent lesions such as intraosseous lipoma.11 An area of high grade osteosarcomatous transformation arose in the midst of the exuberant reactive bone formation induced and surrounded by the tumour cells. Indeed, the patient experienced a hitherto undescribed phenomenon of developing a de novo high grade osteosarcoma of the left humerus secondary to involvement by metastatic low grade ESS. The process is distinct from histological progression to high grade tumour or de-differentiation, in which a longstanding low grade, well-differentiated mesenchymal neoplasm gives rise to or transforms into a high grade one of the same or different phenotype. It is recognised for example in welldifferentiated chondrosarcoma.12 Rather, it is malignant transformation of the persistently actively proliferating osteoblastic cells in reaction to stimulation by the metastatic low grade ESS, a mechanism exactly similar to that seen in secondary osteosarcoma complicating Paget’s disease.13 In fact, this process of de novo origination of a malignant osseous neoplasm secondary to another metastatic low grade malignancy is unique and has never been reported, although is understandable in view of the notoriously indolent growth of low grade ESS even in metastatic sites. The finding of secondary high grade osteosarcoma arising in the humerus due to involvement by metastatic low grade ESS extends beyond mere pathological interest and has important therapeutic implications. Even though this high grade component constitutes only about 25% of the entire tumour, in view of its permeation into lymphovascular vessels, the clinical course would be expected to follow that of a conventional high grade osteosarcoma. Furthermore, despite recent reports of the beneficial effects of hormonal therapies including progestins, aromatase inhibitors and gonadotropin-releasing hormone analogues in improving survival of patients with low grade ESS,14 they offered limited promise in our patient due to the negative immunohistochemical staining for oestrogen and progesterone receptors. Accordingly, our patient was offered a treatment regime for osteosarcoma but unfortunately she developed depression

Fig. 2 (A) Cut surface of the resected proximal humerus showing tumour involving the entire epiphysis and metaphysis with permeative infiltration into the diaphysis. There is invasion through the cortex in the metaphyseal region with widespread infiltration into the surrounding soft tissue. The joint appears uninvolved. (B) The tumour shows an area of osteosarcomatous transformation from a background of florid reactive bone formation. The high grade malignant tumour cells are surrounded by lacelike osteoid (H&E). (C) High power magnification to show the osteosarcomatous area. The tumour cells show marked nuclear pleomorphism and hyperchromatism, and are surrounded by lace-like osteoid (H&E).

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

78

CORRESPONDENCE

from the side effects of the chemotherapy which was prematurely terminated. As a result, our patient pursued a progressively downhill clinical course, typical of that of inadequately treated osteosarcoma. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Louis Tsun Cheung Chow Department of Anatomical and Cellular Pathology, Prince of Wales Hospital, Hong Kong Contact Louis T. C. Chow. E-mail: [email protected] 1. Chan JK, Kawar NM, Shin JY, et al. Endometrial stromal sarcoma: a population-based analysis. Br J Cancer 2008; 99: 1210–5. 2. Chang KL, Crabtree GS, Lim-Tan SK, et al. Primaryuterine endometrial stromal neoplasms. A clinicopathologic study of 117 cases. Am J Surg Pathol 1990; 14: 415–38. 3. Bakker IS, Hoven-Gondrie ML, Moll FC, de Haan HH. A very late recurrence of a formerly misdiagnosed low grade endometrial stromal sarcoma metastasized to the colon. Int J Surg Case Rep 2013; 4: 1113–6. 4. Tian W, Latour M, Epstein JI. Endometrial stromal sarcoma involving the urinary bladder: a study of 6 cases. Am J Surg Pathol 2014; 38: 982–9. 5. Styron SL, Burke TW, Linville WK. Low-grade endometrial stromal sarcoma recurring over three decades. Gynecol Oncol 1989; 35: 275–8. 6. Lehrner LM, Miles PA, Enck RE. Complete remission of widely metastatic endometrial stromal sarcoma following combination chemotherapy. Cancer 1979; 43: 1189–94. 7. Matsuura Y, Yasunaga K, Kuroki H, et al. Low-grade endometrial stromal sarcoma recurring with multiple bone and lung metastases: report of a case. Gynecol Oncol 2004; 92: 995–8. 8. Al-Salam S, El-Terifi H, Ghazal-Aswad S. Low-grade endometrial stromal sarcoma with sex cord-like differentiation metastatic to the thoracic spines. APMIS 2006; 114: 651–5. 9. Batista LM, Carvalho CH, Acioly MA, et al. Spinal metastasis of endometrial stromal sarcoma: Clinicopathological features and management. Surg Oncol 2011; 20: e78–83. 10. Date I, Yasunori Y, Bukeo T. Endometrial stromal sarcoma metastatic to the skull—case report. Neurol Med Chir (Tokyo) 1986; 26: 571–4. 11. Chow LT, Lee KC. Intraosseous lipoma. A clinicopathologic study of nine cases. Am J Surg Pathol 1992; 16: 401–10. 12. Bruns J, Fiedler W, Werner M, Delling G. Dedifferentiated chondrosarcoma—a fatal disease. J Cancer Res Clin Oncol 2005; 131: 333–9. 13. Dray MS, Miller MV. Paget’s osteosarcoma and post-radiation osteosarcoma: secondary osteosarcoma at Middlemore Hospital, New Zealand. Pathology 2008; 40: 604–10. 14. Ioffe YJ, Li AJ, Walsh CS, et al. Hormone receptor expression in uterine sarcomas: prognostic and therapeutic roles. Gynecol Oncol 2009; 115: 466–71.

Pathology (2015), 47(1), January

this unusual clinical presentation is a unidirectional checkvalve mechanism, which is also supposed to cause cavitation in lung cancers.2 A 67-year-old man with diabetes and hypertension under regular medical control had suffered from progressive dysphagia and nausea for 2 months. Due to worsening of these symptoms, he visited our hospital. Endoscopy showed much food debris at the lower oesophagus and one large out-pouching hollow structure at the lower oesophagus near the oesophagogastric junction which was consistent with an oesophageal diverticulum. In addition to oesophageal wall thickening, one large unicystic lesion (3.9 cm) in the right lower lung (Fig. 1) was disclosed by the chest computed tomography (CT) and a pneumatocoele was considered. He received video-assisted thoracoscopic surgery for the oesophageal diverticulum and the lung cyst. During the operation, one oesophageal diverticulum 35 cm below the incisors was found and excised. Wedge resection was performed to remove the 4 cm lung cyst in the right lower lobe. The pathological specimen of the lung tissue showed a collapsed, thin-walled subpleural cyst with a well-demarcated border. Although grossly undetectable, microscopic examination disclosed a small nodular nest (measuring 2 mm in size) inside its cavity (Fig. 2A). The small nodular nest was composed of cuboidal epithelial cells arranged in a microcystic or follicular pattern with dense eosinophilic, colloid-like secretion, morphologically mimicking thyroid tissue (Fig. 2B). An interstitial component of bland spindle mesenchymal cells was also noted. The cyst was covered by a single layer of cuboidal epithelial cells and an interstitial component of bland spindle cells. In addition, tiny foci of proliferative epithelial cells with microcystic formation and eosinophilic secretion were randomly identified in the cystic wall (Fig. 2C). The cyst was not accompanied by fibrous tissue or inflammatory cells which typically present in a pulmonary bulla. Immunohistochemical stains confirmed both the small nodular nest and the cystic wall had the same cellular components: the epithelial cells were positive for TTF-1, cytokeratin, and epithelial membrane antigen (EMA), and negative for calretinin, thyroglobulin, synaptophysin pneumocytes, while the interstitial cells

DOI: 10.1097/PAT.0000000000000198

Cystic alveolar adenoma: an unusual clinical presentation of a rare lung neoplasm Sir, Alveolar adenoma is an exceedingly rare benign lung neoplasm and has a distinct histological feature composed of a network of spaces covered by pneumocytes and a background of spindle cell stroma.1 Alveolar adenoma typically presents as a peripherally located, well-circumscribed solid nodule. Here we report a case of alveolar adenoma with prominent cystic change, manifesting as a single large cyst that mimics a pneumatocoele. We believe the most possible explanation of

Fig. 1 The chest CT showed a pneumatocoele-like unicystic lesion (arrowhead) without obvious solid component in the right lower lobe of the lung.

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

CORRESPONDENCE

79

A

B C

B

C

Fig. 2 (A) In low power field there was one collapsed cyst (arrowhead) with a small nodular nest (arrow). The small nodular nest measured 2 mm in maximal dimension (H&E). (B) The nodular nest inside the cyst had TTF-1þ pneumocytes (inset) arranged in microcystic spaces containing dense eosinophilic secretion and separated by some mesenchymal cells (H&E). (C) The wall of the collapsed cyst also contained the same cellular component including TTF-1þ pneumocytes (inset), mesenchymal cells, and eosinophilic secretion (H&E).

were smooth muscle actin positive, and negative for desmin, TTF-1, cytokeratin, EMA. Since identical tumour tissue was identified inside the small nodular nest and throughout the cystic wall, an alveolar adenoma with prominent cystic change was diagnosed. Alveolar adenoma is a rare benign lung neoplasm initially reported in 1986.3 There are only 40 cases reported in the English literature.4 Most patients are asymptomatic and incidentally found to have a well circumscribed solitary mass on chest X-ray or CT, usually a peripheral or subpleural location.1,4–6 Under microscopic examination, alveolar adenoma has a welldefined border to adjacent lung parenchyma. Alveolar adenoma has a characteristic multicystic pattern with dilated spaces lined by type II pneumocytes and filled with eosinophilic material. The interstitium is composed of bland spindle mesenchymal cells. The differential diagnosis includes sclerosing haemangiomas which show similar clinical and radiological features as solitary, well circumscribed nodules. Histologically, sclerosing haemangiomas also contain mixed epithelial and stromal cells. However, sclerosing haemangiomas usually have complex structures including papillary, solid, and sclerotic patterns as opposed to uniform microcystic pattern in alveolar adenomas. These two benign lung neoplasms can be easily separated by immunostain since the interstitial cells in alveolar adenomas are not immunoreactive for TTF-1 while the stromal cells in sclerosing haemangiomas show TTF-1 expression.1,4,5 The microcysts in alveolar adenomas have varying sizes and infrequently alveolar adenomas can become large multicystic

lesions.4,6 However, manifesting as a single large cyst that mimics a pneumatocoele, as in this case, is extremely rare. This unusual clinical presentation has only been reported once in the literature.7 Because our case contains microscopic foci of tumour remnants throughout the cystic wall, we consider the whole cystic lesion as an alveolar adenoma with prominent cystic change, instead of a small alveolar adenoma incidentally found beside a pneumatocoele. Another interesting finding in our case is the presence of dense eosinophilic, colloid-like secretion within the uniform microcysts formed by cuboidal epithelial cells. As this morphologically mimics thyroid tissue with the same expression of TTF-1 stain, the lesion could be confused with a metastatic follicular carcinoma from the thyroid gland. In the literature, alveolar adenomas are described to have granular, proteinaceous eosinophilic material. There is no previous report mentioning such a thyroid-like appearance.1,3–7 Our case suggests that alveolar adenoma can infrequently resemble a metastatic thyroid cancer. The distinction can be easily made since alveolar adenoma is immunohistochemically negative for thyroglobulin stain. Solitary cystic lesions in chest CT scan can be developmental abnormalties or acquired lesions. Congenital abnormalities include cystic adenomatoid malformations or bronchogenic cysts, while acquired lesions include bullae, infections, pneumatocoeles or neoplasms.8 Among pulmonary neoplasms with cavitation, the majority (80%) of cases are squamous cell carcinomas, followed by adenocarcinomas and large cell

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

80

CORRESPONDENCE

carcinomas.8 The causes of cavitation in squamous cell carcinomas are tumour necrosis and keratinisation. In adenocarcinomas, cavitation often follows tumour necrosis in high grade lesions or results from a unidirectional check-valve mechanism.2 The mechanism begins when neoplastic cells proliferate in the alveolar wall and grow toward bronchiole. Due to lack of cartilage in the bronchiole, a unidirectional check-valve is subsequently formed, causing the accumulations of gases and the rupture of alveoli. The ruptured alveoli ultimately fuse into a solitary thin-walled cavity which could gradually increase in size with the continual increase in inner pressure.2 The unidirectional check-valve mechanism also applies to cavitated lung cancers other than adenocarcinomas.9 Since alveolar adenoma is a pulmonary neoplasm with proliferative alveolar epithelium and septal mesenchyma, we believe such a mechanism is the most likely cause of cyst formation in the tumour in our case. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Min-Shu Hsieh1* Yao-Hui Tseng2* Syue-Fong Hua1 Yueh-Hung Chou3 1

Department of Pathology, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, 2Department of Radiology, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, and 3Department of Anatomical Pathology, Far Eastern Memorial Hospital, New Taipei City, Taiwan; *these authors contributed equally to this work Contact Yueh-Hung Chou, MD. E-mail: [email protected] 1. Burke LM, Flieder DB. Alveolar adenoma. In: Travis WD, Brambilla E, Muller-Hermelink HK, Harris CC, editors. World Health Organization Classification of Tumours. Tumours of the Lung, Pleura, Thymus and Heart. Lyon: IARC Press, 2004; 82–3. 2. Xue X, Wang P, Xue Q, et al. Comparative study of solitary thin-walled cavity lung cancer with computed tomography and pathological findings. Lung Cancer 2012; 78: 45–50. 3. Yousem SA, Hochholzer L. Alveolar adenoma. Hum Pathol 1986; 17: 1066– 71. 4. Wang X, Li WQ, Yan HZ, et al. Alveolar adenoma combined with multifocal cysts: case report and literature review. J Int Med Res 2013; 41: 895–906. 5. Sak SD, Koseoglu RD, Demirag F, et al. Alveolar adenoma of the lung. Immunohistochemical and flow cytometric characteristics of two new cases and a review of the literature. APMIS 2007; 115: 1443–9. 6. Petrella F, Rizzo S, Pelosi G, et al. Giant alveolar adenoma causing severe dyspnoea. J Thorac Oncol 2010; 5: 1088–90. 7. Cavazza A, Paci M, De Marco L, et al. Alveolar adenoma of the lung: a clinicopathologic, immunohistochemical, and molecular study of an unusual case. Int J Surg Pathol 2004; 12: 155–9. 8. Vourtsi A, Gouliamos A, Moulopoulos L, et al. CT appearance of solitary and multiple cystic and cavitary lung lesions. Eur Radiol 2001; 11: 612–22. 9. Yamada S, Noguchi H, Nabeshima A, et al. Basaloid carcinoma of the lung associated with central cavitation: a unique surgical case focusing on cytological and immunohistochemical findings. Diagn Pathol 2012; 7: 175.

DOI: 10.1097/PAT.0000000000000201

Pathology (2015), 47(1), January

Combined haemosiderin and melanin pigmentation of prominent submucosal plexus ganglia and nerves within a pedunculated sigmoid tubular adenoma

Sir, Pigmentation within colonic mucosa is most often seen in the setting of lipofuscinosis (‘melanosis’) and haemosiderin deposition. The latter is frequently encountered in the setting of pedunculated mucosal polyps when they twist upon their pedicles. Herein reported is the presence of prominent and pigmented ganglia and nerves within the submucosa of a pedunculated sigmoid tubular adenoma, which histochemically was demonstrated to represent a combination of haemosiderin and melanin. A 50-year-old woman underwent colonoscopy following a positive faecal occult blood test as part of a national bowel cancer screening program. She was otherwise well and there was no specific history of use of anthraquinone laxatives or of known neurofibromatosis or multiple endocrine neoplasia. Four mucosal polyps were biopsied, one from the transverse colon and three from the sigmoid colon. The transverse colonic polyp and two of the sigmoid polyps were low grade tubular adenomata. The remaining sigmoid polyp was a hyperplastic polyp. The largest polyp was a pedunculated sigmoid tubular adenoma, which measured 11
9
7 mm with a 4 mm long stalk. It demonstrated fibromuscular hypertrophy, a degree of mucosal displacement, prominent golden brown pigment representing haemosiderin deposition within submucosal macrophages (siderophages) and active chronic inflammation, taken as evidence of secondary traumatic change following torsion of the polyp upon its pedicle. Additionally there were prominent ganglia and nerves within the submucosa of the stalk. Moreover, the ganglia and nerves contained finely granular, muddy brown pigment (Fig. 1A–C). A few tiny nerves extended into the lamina propria but no ganglioneuroma formation was present. Immunohistochemical markers (Fig. 1D,E) demonstrated that both nerves and macrophages contained pigment. Histochemical markers (Fig. 2) were performed to elucidate the nature of the pigment. Submucosal siderophages were highlighted by a Perl’s Prussian Blue stain, which also highlighted the nerves. Masson–Fontana and Schmorl’s stains almost exclusively also highlighted the nerves, but no staining was demonstrated following pretreatment with 10% hydrogen peroxidase bleach for the minimum standard treatment time of 10 min. The pigment within nerves did not appear to specifically stain with periodic acid–Schiff (PAS) or Long Ziehl–Neelsen stains, which would be more expected were it lipofuscin.1 Therefore, the pigment within the nerves was considered to represent a combination of haemosiderin and melanin. Pigmentation of a few less prominent nerves was also seen within a second, 7
5
4 mm, tubular adenoma. Mucosal lipofuscinosis was not seen. Repeated torsion is regarded as the explanation for the occurrence of submucosal haemorrhage with haemosiderin deposition and mucosal displacement through branching and hyperplastic muscularis mucosa in some pedunculated colorectal polyps. The frequency of this secondary change increases with sigmoid origin, and with increasing polyp size and length of the polyp stalk.2

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

CORRESPONDENCE

A

B

D

81

C

E

Fig. 1 Tubular adenoma (H&E). (A) Overview of portion of polyp. (B) Golden brown pigment within macrophages and predominantly muddy brown pigmentation of submucosal plexus and nerves. (C) Submucosal plexus ganglion with pigmentation. (D) S100 (red chromagen) immunomarkers demonstrating a prominent ganglionated nerve. (E) CD163 (red chromagen) highlighting macrophages (siderophages) whilst a pigmented nerve is not highlighted (arrow).

Nerve density within the colon in the setting of normal or lesional mucosa has previously been documented in detail. Within the normal colon, the muscularis propria is richly and diffusely innervated by Auerbach’s plexus, which in turn gives rise to the submucosal or Meissner’s plexus. Each in turn gives rise to fine nerve fibres that innervate in submucosa and lamina propria. The stalks of pedunculated tubular and villous adenomata have been shown to contain Meissner’s plexus ganglia and large diameter nerve fibres. No comment regarding any torsion related changes was made.3 Common pigment granules seen during the histological assessment of tissues include lipofuscin (lipochrome/‘wear and tear pigment’), melanin and haemosiderin. Lipofuscin granules are residual bodies with undigested and/or oxidised lipids, which are thought to be the consequence of the processes of sequestration and enzymatic degradation of cell organelles within lysosomes. Melanin is produced within the melanosome as a result of the oxidation of tyrosine to dopa and ultimately to melanin, whereas haemosiderin develops in residual bodies that result from phagocytosis by macrophages of erythrocytes and their breakdown products.4 Melanosis of intestines simply refers to brown pigmentation. Melanosis coli is a term that refers to the brown or black

A

B

pigmentation of colonic mucosa (lamina propria) found on endoscopy or microscopic examination and often seen incidentally. It is caused by mucosal deposition of lipofuscin pigment rather than melanin, which was previously incorrectly thought to be the case. Hence it may also be referred to as lipofuscinosis coli. Melanosis develops in over 70% of persons who use anthraquinone laxatives (e.g., senna) because they damage colonic epithelial cells resulting in their increased apoptosis, which are then either shed into the colonic lumen, or their damaged organelles are sequestrated within macrophage autolysosomes and digested to form residual lipofuscin bodies.4,5 Intestinal polyps may also show lipofuscinosis, again within the mucosa (lamina propria).6 Within the central nervous system (CNS), types of pigment deposition vary depending upon cell and tissue type. This includes neuromelanin and lipofuscin within neurons and cranial nerves, and occasionally melanin within meninges. Neuromelanin and lipofuscin share histochemical features and experimentally have been shown to be interconvertable. Neuromelanin may be formed within lysosomes via several routes. It can be formed as a product of auto-oxidation of catecholamine precursors. However, it has also been suggested that it may arise following melanisation of lipofuscin.7 Melanin

C

Fig. 2 Histochemical stains characterising pigment type. (A) Perl’s Prussian blue stain for iron highlighting siderophages as well as a submucosal ganglia and nerves. (B,C) Masson–Fontana and Schmorl’s stains, respectively, highlighting pigment within nerves. No staining was seen in either following pretreatment with potassium permanganate for 10 min.

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

82

Pathology (2015), 47(1), January

CORRESPONDENCE

also shares some histochemical staining characteristics with neuromelanin and lipofuscin. Schmorl’s and Masson–Fontana stains highlight melanin and neuromelanin, but lipofuscin in its oxidised form also may stain. The positive staining of melanin and neuromelanin is eliminated by pre-treatment with potassium permanganate for the standard 10 min, whereas the elimination of positive staining by lipofuscin requires much longer pretreatment. Unlike melanin, neuromelanin in common with lipofuscin may show positive staining with PAS and Long Ziehl–Neelsen stains.8 The constituents of neuromelanin include hydrolytic enzyme activity, lipofuscin, melanin and endogenous metals, and are characteristic of lysosomal residual bodies. Pigment composition varies between specific cell types and in normal and disease conditions depending upon the properties of the lysosomally sequestered cytoplasmic material.7 Lipofuscin accumulation within enteric neurons has been observed for over a century, its accumulation seen as an inevitable consequence of age. Similar to the central nervous system, its accumulation and patterns of pigmentation within enteric neurons may be type-specific. Similar pigment in disease conditions may represent very similar ceroid pigment.9 Melanin granules have been demonstrated in Schwann cells at various sites in normal and disease conditions. Schwannian differentiation is also observed by melanocytes and neuroid transformation with maturation is a normal feature in melanocytic naevi. Schwann cells and melanocytes have a common embryonic origin in the neural crest. Moreover, both cell types share an intermediate pluripotential precursor.10 The observation of pigmentation within enteric nerves may most often be interpreted as due to lipofuscin accumulation. However, in the current instance, not only was there haemosiderin deposition, but the non-ferrous pigmentation within the prominent submucosal plexus and nerves displayed histochemical staining characteristics of melanin. Because lipofuscin is expected within nerves, the melanin demonstrated in this case could in fact represent the equivalent of neuromelanin (melanised lipofuscin), which is traditionally described within the CNS. The submucosal plexus and nerves were prominent as has been described in peduncluated adenomatous polyps. Cell damage due to local ischaemia as a result of twisting of the polyp upon its pedicle, along with inflammation and haemorrhage may be an encompassing plausible explanation for the striking yet incidental combined accumulation of haemosiderin and melanin (possibly via lipofuscin) within submucosal ganglia and nerves in this instance. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Mark J. Wilsher Douglass Hanly Moir Pathology, Macquarie Park, NSW, Australia Contact Dr Mark Wilsher. E-mail: [email protected] 1. Chan ACL, Ho LC, Yip WWL, et al. Pigmented ependymoma with lipofuscin and neuromelanin production. Arch Pathol Lab Med 2003; 127: 872–5. 2. Muto T, Bussey HJR, Morson BC. Pseudo-carcinomatous invasion in adenomatous polyps of the colon and rectum. J Clin Pathol 1973; 26: 25–31.

3. Tomita T. Localization of nerve fibers in colonic polyps, adenomas, and adenocarcinomas by immunocytochemical staining for PGP 9.5. Dig Dis Sci 2012; 57: 364–70. 4. Freeman HJ. ‘‘Melanosis’’ in the small and large intestine. World J Gastroenterol 2008; 14: 4296–9. 5. Kew ST. Melanosis coli. IeJSME 2012; 6 (Suppl 1): S53–60. 6. Coyne JD. Melanosis coli in hyperplastic polyps and adenomas. Int J Surg Pathol 2013; 21: 261–3. 7. Barden H. The histochemical relationship of neuromelanin and lipofuscin. J Neuropathol Exp Neurol 1969; 28: 419–41. 8. Chan AKL, Ho LC, Yip WWL, et al. Pigmented ependymoma with lipofuscin and neuromelanin production. Arch Pathol Lab Med 2003; 127: 872–5. 9. Brehmer A, Blaser B, Seitz G, et al. Pattern of lipofuscin pigmentation in nitrergic and non-nitrergic, neurofilament immunoreactive myenteric neuron types of human small intestine. Histochem Cell Biol 2004; 121: 13–20. 10. Carbonell AL, Boya J, Garcı´a-Maurin˜o JE. Presence of melanin in normal human Schwann cells. Histol Histopathol 1992; 7: 329–32.

DOI: 10.1097/PAT.0000000000000205

Plasmacytoma with crystal-storing histiocytosis exhibiting FGFR3 and IgH translocation

Sir, Crystal-storing histiocytosis (CSH) is a rare disorder generally associated with plasma cell neoplasms or B-cell lymphomas.1,2 CSH is characterised by massive accumulation of crystallised immunoglobulins (Igs) in histiocytes,3 but has also been reported in the absence of this process, and its pathogenesis is not clearly understood. Because of its rarity, the cytological features of CSH are not well understood, and the diagnosis is difficult and often overlooked.2 Here, we present for the first time a case of CSH occurring in a vertebral plasmacytoma with a reciprocal translocation of t(4;14)(p16;q32). A 46-year-old Chinese man presented to Xijing Hospital, Xi’an, China, reporting back pain of 2 months duration and for the previous 10 days paralysis of both lower limbs that was particularly pronounced in the left leg. The patient also reported gatism. Anterior and posterior planar 99mTc-MDP bone scans demonstrated an osseous lesion in the right front side of 11th –12th thoracic vertebra (Fig. 1A,B). Computed tomography examination showed a hyperdense lesion in the right front side of 11th –12th thoracic vertebra with slight destruction of bone. The lesion was also visible in the right side of the spinal canal through the intervertebral foramen (Fig. 1C). During the surgery, decompression and fixation of 11th –12th thoracic vertebra were undertaken, and the mass was subsequently excised. The specimen consisted of a 0.8
0.5
0.3 cm fragmented, poorly demarcated soft tissue mass that varied in cross section from grey-white to yellow-tan. It was submitted in its entirety for microscopic evaluation (Fig. 2). Haematoxylin and eosin (H&E) stained sections exhibited diffuse sheets of immature plasma cells which were positive for CD138 (Fig. 2C), MUM1 (Fig. 2D) and lambda light chain (Fig. 2E) restriction, with numerous Dutcher bodies and a few mutinucleate giant cells which were positive for CD68. In addition, non-polarisable crystal-like striations positive for lambda light chain staining were observed in the cytoplasm of the mutinucleate giant cells (Fig. 2A), raising the possibility of CSH. The histiocytes were negative for amyloid

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

CORRESPONDENCE

83

B

A

C

Fig. 1 (A,B) Anterior and posterior planar 99mTc-MDP bone scans demonstrated a multifocal osseous metastasis in the right front side of 11th –12th thoracic vertebra (white arrow). (C) CT shows a hyperdense lesion in the right front side of 11th –12th thoracic vertebra with slight destruction of bone (red arrow), and the lesion is visible in the right side of the spinal canal through the intervertebral foramen.

P, CD20, CD3, CD38, and EBER. The Ki-67 proliferation index was greater than 60%. The IgH/FGFR3 DF FISH Probe Kit (Vysis, USA) detected the t(4;14)(p16;q32) reciprocal translocation, involving the FGFR3 and IgH gene regions. In neoplastic plasma cells of

A

the present case, however, two red/green fusion signals were observed, indicating a reciprocal t(4;14) translocation (Fig. 2B). Finally, because the tumours involved only a single site, with no positive findings in bone marrow aspirates, and no evidence of systemic multiple myeloma (MM), a diagnosis of

B

C

D

E

Fig. 2 (A) A few mutinucleate giant cells were observed. Non-polarisable elongated and rhomboid shaped cytoplasmic crystal-like striations were observed in the cytoplasm of mutinucleate giant cells. (B) Fluorescence in situ hybridisation analysis demonstrates that the tumour cells have two fusion signal patterns. The tumour cells are positive for (C) CD138 and (D) MUM1. (E) The neoplastic plasma cells and the crystals are positive for lambda light chain staining.

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

84

Pathology (2015), 47(1), January

CORRESPONDENCE

‘plasmacytoma with crystal-storing histiocytosis CSH showing t(4;14)(p16;q32) translocation involving the FGFR3 and IgH gene regions’ was made. Recognition of CSH is very important due to its association with occult but potentially treatable lymphoplasmcytic disorders.1,2 Nonetheless, the limited number of cases of plasmacytoma with CSH indicates that systemic disease may not be involved, as has been previously reported in the kidney, head, neck, eye, colon, bone marrow, skin and lung.4 Although CSH is typically associated with a plasma cell neoplasm,3,5 it has never to date been reported in association with a t(4;14)(p16;q32) translocation involving the FGFR3 and IgH gene regions. CSH occurs primarily in association with MM,6 which is often characterised by translocations involving Ig-encoding genes, IgH translocations.7 These translocations are thought to result from errors in isotype switch recombination and involve a number of partner chromosomes, mostly frequently 11q13 and 4p16. Moreover, it is thought that t(4;14)(pl6:q32) genomic location breakpoints are located almost exclusively in or near switch regions of the IgH loci.7 The oncongenic event resulting from the (4;14)(p16;q32) translocation has been shown to be the up-regulation of FGFR3 expression through juxtaposition of the FGFR3 and IgH sequences. This in turn results in increased amounts of FGFR3 protein, FGFR3 activation.8 CSH is an uncommon phenomenon associated with plasma cell dyscrasias, and characterised by the expression of monoclonal Igs. Initially associated with B-cell neoplasms in 1917,1,9 CSH has also been linked to multiple myeloma, light chain disease, solitary plasmacytoma, lymphoplasmacytoid lymphoma, and more rarely, to cases of low grade MALT lymphoma.10 Deposition of crystalline material usually occurs in histiocytes and, less frequently, in epithelial cells or connective tissue stroma of organs and tissues. In other words, the term CSH should be applied when the crystals are noted in the histiocytes, but not in the epithelial cells of organs and tissues. Because of the rare incidence of CSH to date, there is no clear basis for differentiating prognoses between the patients with or without the concomitant symptom of CSH. Although expression of monoclonal IgG is often associated with CSH, the mechanisms by which Ig crystals are formed, and their storage in histiocytes, have not been well understood.7,11 One possible mechanism is overproduction of kappa light chain, which is much more frequently reported in other cases than lambda light chain overproduction,11 which was identified in the present case. The combination of overproduction and conformational alterations in the Ig light chains may lead to crystallisation, impaired histiocytic enzymatic degradation and crystal accumulation within these cells and other tissues. In addition, the light chain crystals in CSH are most likely formed either within the endoplasmic reticulum of plasma cells, prior to extrusion into the extracellular space, or within the lysosomal compartment of macrophages during lysosomal digestion and degradation of ingested Igs. Accumulation of monoclonal Ig light chain within histiocytes, observed as crystal deposition, is usually localised at the site of the lymphoplasma cellular proliferation. The multitude of factors involved complicate determining why these crystalline inclusions accumulate and persist in some patients but not in others. Similarly, the t(4;14) of tumour cells found in our present CSH case may not be the cause of CSH formation. Rich macrophage infiltrates with cytoplasmic inclusions may complicate the diagnosis of CSH. The large histiocytes with

ample fibrillary cytoplasm, for example, can be misidentified as storage diseases, inhaled crystalline materials, mycobacterial or fungal infections, Langerhans cell histiocytosis, or as neoplasms with a histiocytoid appearance, such as melanoma, leukaemia, oncocytic neoplasms, rhabdomyoma, or carcinoid.11,12 Accordingly, the use of immunological methods to confirm that the histiocytes contain crystals that are reactive for Ig light chains, would be useful to support the diagnosis of CSH. In conclusion, we report a case of plasmacytoma with a unique intersection of two unusual situations, namely, CSH and t(4;14)(p16;q32) translocation involving the FGFR3 and IgH gene regions. The neoplastic plasma cells are positive for CD138, the mutinucleate giant cells are positive for CD68 and the crystal is positive for lambda light chain staining, while most other cases are positive for kappa light chain staining, which is a clear difference from the present study. To our knowledge, this is the first report of plasmacytoma with CSH containing a t(4;14) translocation, and we suggest that CSH should be borne in mind as a less common situation of plasmacytic tumours. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Yang Lv* Yixiong Liu* Xia Li Qingguo Yan Zhe Wang State Key Laboratory of Cancer Biology, Department of Pathology, Xijing Hospital, Fourth Military Medical University, Xi’an, Shaanxi Province, P. R. China, *these authors contributed equally and are co-first authors Contact Dr Zhe Wang or Dr Qingguo Yan. E-mail: [email protected]; [email protected] 1. Radhakrishnan S, Maneksha V, Adulkar N. Crystal-storing histiocytosis masquerading ocular adnexal lymphoma: a case report and review of literature. Ophthal Plast Reconstr Surg 2014; 30: e67–9. 2. Rossi G, De Rosa N, Cavazza A, et al. Localized pleuropulmonary crystalstoring histiocytosis: 5 cases of a rare histiocytic disorder with variable clinicoradiologic features. Am J Surg Pathol 2013; 37: 906–12. 3. Johnson M, Mazariegos J, Lewis PJ, Pomakova D. Crystal storing histiocytosis presenting as a temporal lobe mass lesion. Surg Neurol Int 2013; 4: 112. 4. Qureshi A, Kashif M. Crystal-storing histiocytosis. Blood 2010; 115: 2568. 5. Zioni F, Giovanardi P, Bozzoli M, et al. Massive bone marrow crystalstoring histiocytosis in a patient with IgA-lambda multiple myeloma and extensive extramedullary disease. A case report. Tumori 2004; 90: 348–51. 6. Yamamoto T, Hishida A, Honda N, et al. Crystal-storing histiocytosis and crystalline tissue deposition in multiple myeloma. Arch Pathol Lab Med 1991; 115: 351–4. 7. Lebeau A, Zeindl-Eberhart E, Mu¨ller EC, et al. Generalized crystal-storing histiocytosis associated with monoclonal gammopathy: molecular analysis of a disorder with rapid clinical course and review of the literature. Blood 2002; 100: 1817–27. 8. Stewart JP, Thompson A, Santra M, et al. Correlation of TACC3, FGFR3, MMSET and p21 expression with the t(4;14)(p16.3;q32) in multiple myeloma. Br J Haematol 2004; 126: 72–6. 9. Takahashi K, Naito M, Takatsuki K, et al. Multiple myeloma, IgA kappa type, accompanying crystal-storing histiocytosis and amyloidosis. Acta Pathol Jpn 1987; 37: 141–54. 10. Llobet M, Castro P, Barcelo´ C, et al. Massive crystal-storing histiocytosis associated with low-grade malignant B-cell lymphoma of MALT-type of the parotid gland. Diagn Cytopathol 1997; 17: 148–52.

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

CORRESPONDENCE

11. Jones D, Bhatia VK, Krausz T, Pinkus GS. Crystal-storing histiocytosis: a disorder occurring in plasmacytic tumors expressing immunoglobulin kappa light chain. Hum Pathol 1999; 30: 1441–8. 12. Harada M, Shimada M, Fukayama M. Crystal-storing histiocytosis associated with lymphoplasmacytic lymphoma mimicking Weber-Christian disease: immunohistochemical, ultrastructural, and gene-rearrangement studies. Hum Pathol 1996; 27: 84–7.

DOI: 10.1097/PAT.0000000000000200

Primary pancreatic ALK negative anaplastic large cell lymphoma Sir, Malignant tumours of the pancreas constitute about 15% of cancer patients, with adenocarcinoma being the most common malignancy. Primary pancreatic lymphomas (PPL) are extremely rare and constitute less than 0.5% of all pancreatic malignancies and less than 2% of extranodal lymphomas.

PPL are commonly intermediate to high grade non-Hodgkin lymphoma with the most common being diffuse large B-cell lymphoma (60%).1 Very few cases of anaplastic lymphoma kinase negative (ALK–) anaplastic large cell lymphoma (ALCL) have been reported in the literature. Herein, we describe an interesting case of primary pancreatic ALK– ALCL which was diagnosed after a Whipple procedure. A 40-year-old woman was admitted to our gastroenterology department with complaints of jaundice and fatigue. On further evaluation she had total bilirubin of 10.2 mg/dL with direct bilirubin being 8.9 mg/dL, albumin: globulin was reversed to 0.6 with elevated SGOT 532 IU/L, SGPT 256 IU/L, GGT 1257 IU/L, alkaline phosphatase 1870 IU/L and lactate dehydrogenase 659 U/L. Computed tomography scan of the abdomen showed a diffuse peri-ampullary mass. Whipple procedure was performed. Haematoxylin and eosin (H&E) stained sections from the peri-ampullary region showed a tumour composed of large neoplastic cells arranged in diffuse sheets having round to oval nucleus with prominent nucleoli and scanty cytoplasm with high mitotic count. Many interspersed bizarre multinucleated cells were also seen (Fig. 1A,B).

A

LCA

CD2

85

B

CD30

ALK

EMA

Ki67

Fig. 1 (A,B) Biopsy shows large neoplastic cells arranged in diffuse sheets along with many interspersed bizarre multinucleated cells (H&E). Immunohistochemistry of ALCL showing expression of LCA, CD30, CD2, EMA, Ki67:80% but no expression of ALK.

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

86

Pathology (2015), 47(1), January

CORRESPONDENCE

Immunohistochemistry was performed for further categorisation. These large cells were diffusely positive for LCA (2B11þ PD7/26; Dako, Denmark), CD3 (IS503; Dako), CD30 (Ber-H2; Dako), CD2 (AB75; Dako), EMA (E20; Dako) and negative for ALK (CD246-ALK1; Dako), CD20 (L26; Dako), CD5 (SP-19; Dako), CD7 (CBC 37; Dako), and CK (AE1 and AE3; Biogenix), with Ki-67 (MiB-1; Dako) being 80% (Fig. 1). A final diagnosis of ALK–ALCL of T lineage was made. Five regional lymph nodes were isolated and all were free of malignancy. Bone marrow biopsy showed no evidence of lymphoma infiltration. Diagnostic criteria for primary pancreatic lymphoma as suggested by Dawson et al. includes: (1) no significant lymphadenopathy; (2) a normal total leukocyte count; (3) no hepatic or splenic involvement; and (4) tumour mass in the pancreas and spread to peri-pancreatic region only.2 The majority of cases reported in the literature are lymphoma of B-cell origin and only a few cases of T-cell primary pancreatic lymphoma have been described. The main differential considered is a poorly differentiated adenocarcinoma. ALK– ALCL is a provisional entity, and is defined as a CD30þ T cell neoplasm that is not reproducibly distinguishable on morphological grounds from ALKþ ALCL, but lacks ALK protein.3 The current WHO classification considers ALK– ALCL to be a provisional entity within the spectrum of mature T-cell lymphomas, distinct from both ALKþ ALCL and peripheral T-cell lymphoma not otherwise specified (PTCL, NOS). The peak incidence of ALK– ALCL is in adults 40–65 years with a modest male preponderance. Cases occurring in the gastrointestinal tract must be distinguished from CD30þ enteropathy associated T-cell lymphoma. Most patients present with advanced stage III or IV disease with peripheral and/or abdominal lymphadenopathy and B symptoms. Histologically, these cells grow in a cohesive pattern that may mimic adenocarcinoma with interspersed multinucleated including wreath-like cells and ‘hallmark’ cells with eccentric, horse shoe shaped nuclei. Neoplastic cells in ALK– ALCL tend to be larger and more pleomorphic and have a higher N/C ratio than ALKþ ALCL. Emphasis should be placed on CD30 expression. If a large cell lymphoma of non-B-cell phenotype shows strong homogeneous CD30 expression, especially if this is strongest in the golgi and membrane regions and morphological features are consistent with ALCL, it should be classified as ALK– ALCL. Correct diagnosis is facilitated by a careful immunophenotypic analysis and is essential for delivering the appropriate therapy, avoiding an unnecessary gastrointestinal resection and opening the possibility for lymphomadirected chemotherapy.4 Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Pankaj Mishra Manavi Dang Smeeta Gajendra Lipika Lipi Ritesh Sachdev Department of Pathology and Laboratory Medicine, MedantaThe Medicity, Gurgaon, Haryana, India Contact Dr Ritesh Sachdev. E-mail: [email protected]

1. Savopoulos CG, Tsesmeli NE, Kaiafa GD, et al. Primary pancreatic anaplastic large cell lymphoma, ALK negative: a case report. World J Gastroenterol 2005; 11: 6221–4. 2. Saif MW. Primary pancreatic lymphomas. JOP 2006; 7: 262–73. 3. Galani K, Gurudu SR, Chen L, Kelemen K. Anaplastic large cell lymphoma, ALK-negative presenting in the rectum: a case report and review of the literature. OJ Pathol 2013; 3: 37–40. 4. Mason DY, Harris NL, Delsol G, et al. Anaplastic large cell lymphoma, ALK-negative. In: Swerdlow S, Campo E, Harris NL, editors. WHO Classification of Tumors of Hematopoietic and Lymphoid Tissues. Lyon: IARC, 2008; 317–9.

DOI: 10.1097/PAT.0000000000000203

A case of ALK negative anaplastic large cell lymphoma with leukaemic manifestation, transformed from CD4 positive T-cell large granular lymphocytic leukaemia Sir, T-cell large granular lymphocytic leukaemia (T-LGL) is a heterogeneous disorder characterised by the expansion of a discrete or monoclonal population of large granular lymphocytes in the peripheral blood.1 T-LGL usually expresses CD3 and CD8.2,3 In contrast, CD4þ T-LGL has been reported only sporadically.4 The transformation of a T-LGL into a high grade large T-cell lymphoma has rarely been reported,5,6 but no case of CD4þ T-LGL transformation into ALK– anaplastic large cell lymphoma (ALCL) has been recorded. We here report a case of ALK– anaplastic large cell lymphoma transformed from a CD4þ T-LGL. A 59-year-old man with skin rash that developed 6 months previously was admitted to our hospital for the evaluation of skin lesions. A peripheral blood examination indicated leukocytosis (4.5
1010 cells/L) with 34% of neoplastic lymphoid cells displaying large granules in the cytoplasm (Fig. 1A). Flow cytometric analysis revealed surface and cytoplasmic CD3þ, CD4þ, CD8– and CD56–. Mild hepatosplenomegaly and multiple small enlarged lymph nodes were noted on computed tomography. Bone marrow analysis revealed a hypercellular marrow with 2.4% neoplastic lymphoid cells. Immunohistochemical (IHC) stains revealed the neoplastic cells to be CD3þ, CD4þ, CD8–, TCR bF1þ, granzyme Bþ and TIA1þ. The left arm skin biopsy result showed atypical T-cells infiltration, consistent with CD4þ T-LGL, and IHC stains revealed the atypical T-cells to be CD3þ, CD4þ, CD8–, TCR bF1þ, granzyme Bþ, TIA1þ, and Ki-67 labelling index (2%). The right inguinal lymph node biopsy also revealed the involvement of CD4þ T-LGL. Cytogenetic analysis of the marrow sample revealed the abnormal clonal karyotype 46,XY,inv(3)(p21q27), t(12;17)(q24.1;q21),del(13)(q14q22)[2]/46,XY[28]. The patient was diagnosed with CD4þ T-LGL, and received chemotherapy (10.0 mg methotrexate/week for 4 months). After 3 months of this treatment, his skin lesions improved. However, at 1 year after the diagnosis, his skin lesions became aggravated, and fever, arthralgia and several lymph node enlargements in the chest were also noted. Peripheral blood examination revealed leukocytosis (9.8
1010 cells/L) with 68% of the neoplastic lymphoid cells displaying a large anaplastic morphology with irregular shaped nuclei. This appearance was completely different from the large granular lymphocytes that presented at diagnosis (Fig. 1B). A

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

CORRESPONDENCE

A

87

B

Fig. 1 Neoplastic lymphoid cells in the peripheral blood (Wright–Giemsa stain). (A) Neoplastic lymphoid cells at diagnosis in the patient displaying large granules. (B) Neoplastic lymphoid cells at transformation displaying irregular shaped nuclei.

bone marrow study revealed a hypercellular marrow with 49.6% neoplastic lymphoid cells. Flow cytometry revealed surface and cytoplasmic CD3þ, CD4þ and CD2þ. IHC on bone marrow biopsy material revealed the neoplastic cells to be CD30þ, CD3þ, CD4þ, granzyme Bþ, TIA1þ, EMAþ, ALK– and CD56–. Cytogenetic analysis of a marrow sample revealed the abnormal clonal karyotype 47,XY,inv(3)(p21q27),del(6) (q21q23),add(9)(p22),add(10)(p11.2),t(12;17)(q24.1;q21),del (13)(q14q22),þ19[17]/48,idem,þ19[2]/49,idem,þdel(13) (q14q22)[1]. Right axillary lymph node biopsy findings were consistent with ALK– ALCL. IHC testing of a lymph node biopsy disclosed CD30þ, CD4þ, ALK– and CD56–. The patient was diagnosed with leukaemic manifestation of ALK– ALCL transformed from CD4þ T-LGL, but did not survive despite salvage chemotherapy. Clear clinical differences exist between CD4þ T-LGL and CD8þ T-LGL, particularly with regard to the absence of neutropenia, anaemia, splenomegaly, rheumatoid arthritis and the higher incidence of association with malignant diseases in CD4þ T-LGL.4,7 In our current case, neutropenia, anaemia and rheumatoid arthritis were absent. Only one case of CD4þ T-LGL displaying skin lesions has been reported to date.4 In our present case, the skin lesions demonstrated atypical T-cell infiltration on skin biopsy. Generally, T-LGL involves the peripheral blood, bone marrow, liver and spleen, but our present CD4þ T-LGL case involved the skin. ALK– ALCL involves the lymph nodes and extranodal tissues such as bone, soft tissue and skin. A genetic evolution to a more complex karyotype was noted from our present case of CD4þ T-LGL transformation to ALK– ALCL. Our current report likely describes the first case of leukaemic manifestation and skin involvement of ALK– ALCL transformed from CD4þ T-LGL with skin lesions.

Contact Chan-Jeoung Park. E-mail: [email protected]

Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose.

Sir, Laboratory confirmed shigellosis is a national notifiable infection in Australia. While most cases are associated with recent overseas travel, locally acquired cases have been reported in men who have sex with men (MSM).1 Antibiotic therapy is recommended in all cases, both to treat severe disease, and to reduce the period of asymptomatic carriage.2 Increasingly, azithromycin is being used empirically for the treatment of bacterial gastroenteritis because of increasing antibiotic resistance in other gastrointestinal pathogens such as Salmonella species and Campylobacter species.3 While there are no official breakpoints for azithromycin against Shigella species, a

Jaewook Kim1 Chan-Jeoung Park1 Eul-Ju Seo1 Cheolwon Suh2 Departments of 1Laboratory Medicine, and 2Internal Medicine, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Korea

1. Semenzato G, Zambello R, Starkebaum G, et al. The lymphoproliferative disease of granular lymphocytes: updated criteria for diagnosis. Blood 1997; 89: 256–60. 2. Morice WG, Jevremovic D, Hanson CA. The expression of the novel cytotoxic protein granzyme M by large granular lymphocytic leukaemias of both T-cell and NK-cell lineage: an unexpected finding with implications regarding the pathobiology of these disorders. Br J Haematol 2007; 137: 237–9. 3. Morice WG, Kurtin PJ, Tefferi A, Hanson CA. Distinct bone marrow findings in T-cell granular lymphocytic leukemia revealed by paraffin section immunoperoxidase stains for CD8, TIA-1, and granzyme B. Blood 2002; 99: 268–74. 4. Lima M, Almeida J, Dos Anjos Teixeira M, et al. TCRalphabetaþ/CD4þ large granular lymphocytosis: a new clonal T-cell lymphoproliferative disorder. Am J Pathol 2003; 163: 763–71. 5. Matutes E, Wotherspoon AC, Parker NE, et al. Transformation of T-cell large granular lymphocyte leukaemia into a high-grade large T-cell lymphoma. Br J Haematol 2001; 115: 801–6. 6. Tagawa S, Mizuki M, Onoi U, et al. Transformation of large granular lymphocytic leukemia during the course of a reactivated human herpesvirus-6 infection. Leukemia 1992; 6: 465–9. 7. Garrido P, Ruiz-Cabello F, Barcena P, et al. Monoclonal TCRVbeta13.1þ/CD4þ/NKaþ/CD8-/þdim T-LGL lymphocytosis: evidence for an antigen-driven chronic T-cell stimulation origin. Blood 2007; 109: 4890–8.

DOI: 10.1097/PAT.0000000000000204

Outbreak of locally acquired azithromycinresistant Shigella flexneri infection in men who have sex with men

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

88

Pathology (2015), 47(1), January

CORRESPONDENCE

Number of cases

30

Serotype 1b 2a

20

2b 3a

surveillance of emerging antibiotic resistance in Shigella species is warranted, and clinical laboratories should be aware of possible azithromycin-resistant Shigella species locally acquired in Australia when providing advice to clinicians regarding empiric therapy.

6

10

Other

20 0 20 0 0 20 1 0 20 2 0 20 3 0 20 4 0 20 5 0 20 6 0 20 7 0 20 8 0 20 9 1 20 0 1 20 1 1 20 2 1 20 3 14 *

0

Year Fig. 1 Number of cases of Shigella flexneri serotypes in Victoria tested at MDU PHL, from 2000. *Data for 2014 is January to June inclusive.

minimum inhibitory concentration (MIC) 16 mg/mL is considered susceptible in the normal wild-type distribution.3,4 The Microbiological Diagnostic Unit Public Health Laboratory (MDU PHL) routinely performs identification, serotyping and antimicrobial susceptibility testing of Shigella isolates from clinical cases in Victoria and selected isolates from NSW.5 Recent reports of increasing antibiotic resistance in Shigella species, including a cases of infection where high-level resistance to azithromycin was demonstrated,3,6 and an increase in local laboratory confirmed cases of Shigella flexneri, prompted our investigation for local azithromycin resistance. Between 1 May 2013 and 30 June 2014, the number of Shigella flexneri 3a cases reported in Melbourne and Sydney was 29 and 16, respectively, a significant increase compared to previous years, with the major increase seen in 2014. Figure 1 summarises the serotype prevalence data for cases notified in Victoria from 2000 and where isolates were characterised at MDU PHL, demonstrating the significant increase in Shigella flexneri 3a cases in the first half of 2014. The number of infections caused by other Shigella flexneri serotypes was unchanged over the same time period. Thirty-eight of 45 isolates (14 from 2013, 12 males; 24 from 2014, all males) were submitted for azithromycin susceptibility testing, using azithromycin Etest according to the manufacturer’s instructions (bioMerieux, France). Susceptibility to azithromycin (MIC 16 mg/mL) was only seen in four 2013 isolates from two males and two females. Of the 34 remaining azithromycinresistant isolates (89% of isolates tested), 33 were resistant to azithromycin (MIC >256 mg/mL), ampicillin, and tetracycline, while one isolate was resistant to azithromycin (MIC 128 mg/ mL), ampicillin, tetracycline and exhibited decreased susceptibility to ciprofloxacin (0.25 mg/mL). Approval to include de-identified demographic data summarising risk factors for Shigella infection was granted by Department of Health, Victoria. The median age for the azithromycin resistant cases was 40 years (range 21–65 years), with most cases reported as occurring in MSM without overseas travel, indicating local acquisition in the MSM population in Melbourne and Sydney. These data are the result of passive surveillance of notifiable infections undertaken during this period, and not active surveillance in this risk group. Emerging azithromycin resistance in Shigella species further highlights the growing burden of antibiotic resistant bacterial infections, and has significant implications for the empiric treatment of severe bacterial gastrointestinal disease. Particularly in the MSM community, possible cases of shigellosis should not be empirically treated with azithromycin. Active

Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Mary Valcanis1 Jeremy D. Brown2 Briony Hazelton2 Matthew V. O’Sullivan2,3 Alex Kuzevski1 Courtney R. Lane4,5 Benjamin P. Howden1,6 1

Microbiological Diagnostic Unit Public Health Laboratory, Department of Microbiology and Immunology, The University of Melbourne at The Doherty Institute for Infection and Immunity, Melbourne, Vic, 2Centre for Infectious Disease and Microbiological Laboratory Services, Institute for Clinical Pathology and Medical Research, Westmead Hospital, Sydney, NSW, 3Marie Bashir Institute for Infectious Diseases and Biosecurity, University of Sydney, Sydney, NSW, 4Victorian Department of Health, Communicable Disease Epidemiology and Surveillance, Health Protection Branch, Melbourne, Vic, 5National Centre for Epidemiology and Population Health, Australian National University, Canberra, ACT, and 6Infectious Diseases Department, Austin Health, Heidelberg, Vic, Australia Contact Professor Benjamin Howden. E-mail: [email protected] 1. Rowe SL, Radwan S, Lalor K, et al. An outbreak of shigellosis among men who have sex with men, Victoria, 2008. Vic Infect Dis Bull 2010; 13: 119–23. 2. Gastrointestinal Expert Group. Therapeutic Guidelines: Gastrointestinal. Version 5. Melbourne: Therapeutic Guidelines Limited, 2011. 3. Hassing RJ, Melles DC, Goessens WHF, Rijnders BJA. Case of Shigella flexneri infection with treatment failure due to azithromycin resistance in an HIV-positive patient. Infection 2014; 42: 789–90. 4. European Committee on Antimicrobial Susceptibility Testing. Breakpoint Tables for Interpretation of MICs and Zone Diameters, Version 4, 2014. Cited 20 Jul 2014. http://www.eucast.org/clinical_breakpoints 5. Valcanis M. Laboratory testing of Shigella. Vic Infect Dis Bull 2010; 13: 114–8. 6. Heiman KE, Karlsson M, Grass J, et al. Notes from the field: Shigella with decreased susceptibility to azithromycin among men who have sex with men – United States, 2002-2013. MMWR Morb Mortal Wkly Rep 2014; 63: 132–3.

DOI: 10.1097/PAT.0000000000000207

Direct identification of bacteria from positive BD-Bactec blood culture bottles on the Vitek MS Sir, Timely and the appropriate choice of antimicrobial therapy is associated with lower mortality in patients with bacterial sepsis.1 Unfortunately there is often a delay of days before

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

CORRESPONDENCE

bacteria are isolated from blood cultures, identified and antimicrobial susceptibilities are determined, during which time patients may remain on inappropriate empirical antimicrobial therapy. The recent widespread implementation of matrix assisted laser desorption and ionisation time of flight mass spectrometry (MALDI-TOF MS) into diagnostic laboratories offers a rapid and reliable alternative method for direct identification of isolates from blood cultures. Identification of the isolate may allow tailoring of antibiotic therapy based on local sensitivity data. It may also be used in combination with molecular techniques such as the Xpert MRSA/SA BC2 (Cepheid, USA), or inoculation of appropriate antimicrobial susceptibility cards on automated systems3 such as the Vitek-2 (bioMe´rieux, France). In this study we evaluated the performance of a modified extraction protocol previously reported for the BD-Bruker Maldi Biotyper4 (Becton Dickinson, USA), on the Vitek-MS (bioMe´rieux), including the relative contribution of protocol steps to yield. LabPlus serves as the diagnostic laboratory for a 1000 bed university-affiliated tertiary care hospital. Blood cultures are processed 24 h a day. All positive blood cultures have a Gram stain performed; the results of which determine the media that the broth is subcultured onto. Bacterial isolates are identified using the Vitek MS. Direct identification using the MS was performed on positive blood cultures prior to growth on subculture. Fisher’s exact test was used for comparing unmatched, and McNemar’s test for comparing matched data. For the basic protocol extraction method: 3.5 mL of broth was removed from the blood culture bottles and injected into a 3.5 mL serum separator tube (SST) (Becton Dickinson). This was centrifuged at 3300 rpm for 10 min. The supernatant was removed, and pellet re-suspended in 1.0 mL of sterile water. This was manually agitated prior to transfer to a 2 mL tube. After centrifugation at 13,000 rpm for 1 min, the supernatant was removed and resulting bacterial pellet spotted in quadruplicate onto a 48-well Vitek MS-DS disposable target slide. Sample wells were overlaid with 1 mL alpha-cyano-4-hydroxycinnamic acid (HCCA) matrix (bioMe´rieux), and run on the Vitek MS IVD system. For the additional centrifugation step, following centrifugation of the SST, the pellet was re-suspended in 1.5 mL of sterile water, transferred to a 2 mL tube, and centrifuged at 2100 rpm for 1 min. Then 1 mL of the supernatant was transferred to a 2 mL tube and centrifuged at 13,000 rpm for 1 min. The pellet was processed as for the basic protocol. For evaluation of formic acid extraction, a subset of samples were overlaid with 0.5 mL formic acid (bioMe´rieux) on two sample wells of the target slide. These were then overlaid with matrix and run as per the basic protocol. A total of 128 BD Bactec Plus Aerobic/F, Plus Anaerobic/F and Peds Plus (Becton Dickinson) blood culture bottles from 86 patients (73 adults and 13 children) collected between January and October 2013, were processed using either the basic method, or with the additional centrifugation step. Of these, 105 (82%) bottles were positive by Gram stain and had a single isolate grown on subculture. Fifteen (12.5%) bottles had growth of two or more isolates on subculture. Two bottles contained yeasts, and six bottles had flagged positive but had negative Gram stains (false positive).

89

For blood cultures with a single isolate, the overall correct bacterial identification to species level by Vitek MS software was achieved for 57 of 105 (54%) of isolates; 47 of 61 (77%) Gram negative organisms and 10 of 44 (23%) Gram positive organisms using either method. No identification was given for 40 of 105 (38%) of isolates. Incorrect identification occurred in eight of 105 (7.6%) single isolate cultures; two of 61 (3.3%) Gram negative organisms, and six of 44 (14%) Gram positive organisms (Table 1). For polymicrobial samples, correct identification for one organism was achieved in 13 of 15 (87%) of samples. In one of the six blood culture bottles which had flagged falsely positive, direct identification gave an erroneous identification of Vibrio fluvialis with high confidence (98.7%). The blood cultures remained negative after 5 days incubation. The method did not identify yeasts. Using the basic method, 15 of 24 (63%) of Gram negative organisms, six of 26 (23%) of Gram positive organisms, and 21 of 50 (42%) of all isolates were correctly identified. With the additional centrifugation step, 32 of 37 (86%) of Gram negative organisms, four of 18 (22%) of Gram positive organisms, and 36 of 55 (65%) of all bacteria were correctly identified. Formic acid extraction gave correct identification for 31 of 52 (60%) isolates compared with 30 of 52 (58%) ( p ¼ 0.262) of those processed without this step. We found this method to be reliable for the identification of Gram negatives, and were able to improve on an initial correct identification rate of 63% to 86% ( p ¼ 0.059) for Gram negatives (42% to 65%, p ¼ 0.019 for all bacteria) with an additional centrifugation step to remove remaining cellular debris, which may interfere with MS spectral recognition.5 Our results compare favourably both with the findings of Moussaoui et al. (91%),4 and those of other recent reports for the identification of Gram negatives.6–8 Identification of Gram positive organisms (23%) by this method was poor, compared with that reported by Moussaoui et al. (89%).4 The yield did not substantially improve with an additional centrifugation step. A significant (14%) risk of false positive identifications contributes to rendering this method unsuitable for testing Gram positive organisms on the Vitek MS. The performance of individual organisms using either the basic method, or with the additional centrifugation step is presented in Table 1. Issues with identification of Salmonella, Campylobacter, Rhizobium, and Fusobacterium have been reported elsewhere.4,5 The relatively small number of samples in this study restricted our ability to assess the performance of other less commonly isolated organisms. The sensitivity of this method for identification of Staphylococcus aureus was particularly poor (18%), and inferior to the sensitivity of the two hour tube coagulase test (68%) performed at our laboratory. Rapid identification of S. aureus isolates when paired with genotypic identification of methicillin resistance2 can be used to guide antibiotic choice. Coagulase negative staphylococci, which represented the largest group of Gram positive organisms, were also infrequently identified to species level (21%). Reliable identification of these organisms may provide valuable information that a positive blood culture is likely to be a contaminant, avoiding unnecessary escalation of antibiotics. The differences observed for identification of Gram positive organisms compared with Moussaoui et al. may be accounted for in part by differences in the performance of the Vitek MS,

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

90

Pathology (2015), 47(1), January

CORRESPONDENCE

Table 1

Direct MALDI-TOF identification compared with conventional subculture MALDI-TOF identification result for all isolates tested Correct identification Species level

Organism Gram negatives E. coli Enterobacter spp Klebsiella spp P. mirabilis Salmonella spp S. marcescens Y. enterocolitica C. jejuni H. influenzae K. kingae P. aeruginosa R. radiobacter Bacteroides spp F. necrophorum Subtotal (%) Gram positives CONS S. aureus Micrococcus spp Gemella sputi S. agalactiae Viridans streptococci Enterococcus spp L. monocytogenes Rhodococcus equi C. perfringens Subtotal (%) Total

All samples

Monomicrobial

Polymicrobial

35/38 4/5 6/9 0/1 3/8 3/3 1/1 0/1 1/1 0/1 1/1 0/1 3/3 0/2 60/75 (80%)

27/30 3/4 6/6 0/0 3/8 3/3 1/1 0/1 1/1 0/1 1/1 0/1 2/2 0/2 47/61 (77%)

8/8 1/1 0/3 0/1

7/34 2/12 0/1 0/1 1/2 0/3 2/3 1/1 0/1 1/2 14/60 (23%) 71/135 (53%)

6/24 2/11 0/1 0/1 1/2 0/3 0/0 1/1 0/1 0/0 10/44 (23%) 57/105 (54%)

Genus

Incorrect

4/8

1/1 1/1 1/1 10/14 (71%) 1/10 0/1

2/61 (3%) 10/34 4/12

1/2 2/3 1/1 1/2 4/16 (25%) 14/30 (47%)

6/44 (14%) 8/105 (8%)

CONS, coagulase negative staphylococcus.

and BD-Bruker MS for direct blood culture identification for this group of organisms.9 Others have also reported difficulties with Gram positive organism identification using centrifugation based methods on the BD-Bruker MS,6 possibly due to loss of bacteria in multiple centrifugation steps, as the number of bacteria present in positively flagged cultures with Gram positive organisms is generally lower than for Gram negative organisms.10 Therefore, a different approach for Gram positive organisms may be required on the Vitek MS platform. Two alternative methods, both evaluated on the Vitek MS have reported higher Gram positive organism identification rates, using lysis-centrifugation8 (92%) and lysis-filtration7 (75%). A known limitation of direct identification is the inability of Vitek MS software to detect polymicrobial infections, which mandates that the MS result be interpreted in conjunction with the Gram stain. The minimum time to results for samples was 35 min, with a median of 110 min in routine circumstances. This fitted well with hospital laboratory and clinical routines. We found that formic acid extraction made no difference to performance, allowing a time saving of 5–10 min. Direct identification was usually achieved 24 h before conventional methods, although we did not assess the use of the MS for identification of early growth from subcultures outside routine laboratory workflow, which may offer a labour saving compromise to this technique.11 The reagent cost for this method was NZ$1.16, comparing favourably with the Foster method (NZ$2.69) and the

Sepsityper (Becton Dickinson) (NZ$6.20). No additional equipment is required, and steps were simple to perform. The impact of additional labour time on laboratory functioning was not assessed, a factor likely to be a major limitation on the implementation of direct identification in laboratories. Several studies have shown that the routine identification of bacteria directly from blood cultures using MALDI-TOF MS has a positive impact on outcomes of patients with bacterial sepsis, increasing early institution of appropriate antibiotics and reducing cost and length of hospital stay,6,12 although the contribution of direct identification alone to these improved outcomes has been difficult to determine. This method is rapid, sensitive and reliable for Gram negative organism identification, simple to perform and requires only readily available and inexpensive consumables. Although it performs poorly for Gram positive organisms, laboratories may find routine application of this protocol useful for Gram negative organisms, as improvements in the early institution of appropriate antimicrobial therapy have been reported with these bacteria alone.12 Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Gary N. McAuliffe Mary Bilkey Sally A. Roberts

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

CORRESPONDENCE

Department of Microbiology, LabPlus, Auckland, New Zealand Contact Dr Gary McAuliffe. E-mail: [email protected] 1. Ferrer R, Martin-Loeches I, Phillips G, et al. Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guideline-based performance improvement program. Crit Care Med 2014; 42: 1749–55. 2. Clerc O, Prod’hom G, Senn L, et al. Matrix-assisted laser desorption ionization time-of-flight mass spectrometry and PCR-based rapid diagnosis of Staphylococcus aureus bacteraemia. Clin Microbiol Infect 2014; 20: 355–60. 3. Machen A, Drake T, Wang YF. Same day identification and full panel antimicrobial susceptibility testing of bacteria from positive blood culture bottles made possible by a combined lysis-filtration method with MALDITOF VITEK mass spectrometry and the VITEK2 system. PLoS One 2014; 9: e87870. 4. Moussaoui W, Jaulhac B, Hoffmann AM, et al. Matrix-assisted laser desorption ionization time-of-flight mass spectrometry identifies 90% of bacteria directly from blood culture vials. Clin Microbiol Infect 2010; 16: 1631–8. 5. Stevenson LG, Drake SK, Murray PR. Rapid identification of bacteria in positive blood culture broths by matrix-assisted laser desorption ionizationtime of flight mass spectrometry. J Clin Microbiol 2010; 48: 444–7. 6. Vlek AL, Bonten MJ, Boel CH. Direct matrix-assisted laser desorption ionization time-of-flight mass spectrometry improves appropriateness of antibiotic treatment of bacteremia. PLoS One 2012; 7: e32589. 7. Fothergill A, Kasinathan V, Hyman J, Walsh J, Drake T, Wang YF. Rapid identification of bacteria and yeasts from positive-blood-culture bottles by using a lysis-filtration method and matrix-assisted laser desorption ionization-time of flight mass spectrum analysis with the SARAMIS database. J Clin Microbiol 2013; 51: 805–9. 8. Foster AG. Rapid Identification of microbes in positive blood cultures by use of the vitek MS matrix-assisted laser desorption ionization-time of flight mass spectrometry system. J Clin Microbiol 2013; 51: 3717–9. 9. Chen JH, Ho PL, Kwan GS, et al. Direct bacterial identification in positive blood cultures by use of two commercial matrix-assisted laser desorption ionization-time of flight mass spectrometry systems. J Clin Microbiol 2013; 51: 1733–9. 10. Ferreira L, Sa´nchez-Juanes F, Mun˜oz-Bellido JL, Gonza´lez-Buitrago JM. Rapid method for direct identification of bacteria in urine and blood culture samples by matrix-assisted laser desorption ionization time-of-flight mass spectrometry: intact cell vs. extraction method. Clin Microbiol Infect 2011; 17: 1007–12. 11. Idelevich EA, Schu¨le I, Gru¨nastel B, Wu¨llenweber J, Peters G, Becker K. Rapid identification of microorganisms from positive blood cultures by MALDI-TOF mass spectrometry subsequent to very short-term incubation on solid medium. Clin Microbiol Infect 2014; Apr 3;. (Epub ahead of print). 12. Clerc O, Prod’hom G, Vogne C, Bizzini A, Calandra T, Greub G. Impact of matrix- assisted laser desorption ionization time-of-flight mass spectrometry on the clinical management of patients with Gram-negative bacteremia: a prospective observational study. Clin Infect Dis 2013; 56: 1101–7.

DOI: 10.1097/PAT.0000000000000206

Evaluation of HemoCue white blood cell differential counter at a remote health centre in Australia’s Northern Territory Sir, The total white blood cell (WBC) count and differential is a frequently ordered pathology test to detect infection, inflammation or as part of a full blood count or complete blood count in a routine health assessment.1 Differentiation of the number of white blood cells into subtypes (neutrophils, lymphocytes, monocytes, eosinophils and basophils) provides additional clinical information; for example, to distinguish between a potential bacterial, viral or parasitic infection.2

91

Traditionally the differential WBC count is measured by automated laboratory cell counters. The ability to measure a differential WBC count by point-of-care testing (POCT) would be useful in extra-laboratory (e.g., emergency departments and outpatient clinics) and primary care (e.g., general practices and Aboriginal Medical Services) settings where the convenience, accessibility, portability and immediacy of the POCT result would be advantageous.3 In 2008, HemoCue AB (Sweden) developed a POCT device for total WBC count. In 2012, this device was further refined to additionally provide a 5-part differential count. In this study we examined the analytical performance characteristics and usefulness of the HemoCue WBC Diff counter in a remote Aboriginal medical service, where access to pathology services is difficult due to the geographical isolation from the nearest laboratory and the high burden of infection among its Aboriginal community members. The study was implemented in a remote Indigenous community in Australia’s Northern Territory, located over 600 km from the nearest major town and laboratory service (Mt Isa, Queensland). The community has a population between 500 and 800 people, of which 94% are Aboriginal.4 The Remote Health Centre is serviced by two Remote Area Nurses (RAN) and a visiting medical practitioner. During the wet season the community is frequently cut off via road due to flooding. Ethics registration for this project was obtained in March 2013 from the Menzies School of Medicine Ethics Committee. Informed verbal consent was obtained from each patient routinely presenting to clinic with symptoms indicating a WBC count was needed. Patient participation in the study was voluntary. A 10 mL sample of capillary or venous blood is loaded into a HemoCue WBC Diff microcuvette. The red blood cells are lysed and the WBCs stained with methylene blue. Thirty seven images are then taken of the stained WBCs using microphotographs, images are focused and mathematical algorithms classify and count cell types. Total WBC and a 5-part differential count are displayed in less than 5 min. The measuring range for the total WBC count is 0.3–30
109/L. The analyser is small, lightweight and portable (size 190
160
160 mm, weight 1.3 kg). The analyser operates by batteries or by AC power adaptor and has external connectivity capability. Venous whole blood samples sent to the laboratory were measured on a Sysmex XE-2100 analyser (Sysmex America, USA). Fluorescent labelling is used to measure the nucleusplasma ratio of each individually stained cell, enabling differentiation and reporting of six WBC populations. The XE-Series utilises an adaptive cluster analysis system (ACAS) to separate cell populations into well-defined three-dimensional clusters. This study comprised 62 adult patients routinely presenting to clinic with symptoms indicating a WBC differential count was needed. Each patient result was recorded using a unique code number to maintain confidentiality in the study. The Flinders University International Centre for Point-of-Care Testing (iPOCT) developed a training resource package and delivered on-site training to the clinical team at the health service. A RAN collected the venous sample to be sent to the laboratory and tested a small aliquot of the venous sample on the WBC Diff prior to despatch to the laboratory. A fingerpick capillary sample was also taken at the same time from each patient and tested on the WBC Diff on-site. The venous sample was then sent via Mount Isa to the nearest accredited pathology laboratory in Brisbane some 2000 km from the remote community. The

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

92

Pathology (2015), 47(1), January

CORRESPONDENCE

average transport time was between 12 and 24 h, while the turnaround time for the result to be reported to the treating doctor at the remote health service was between 4 and 7 days. Results of POCT performed on both venous and capillary samples and comparative laboratory test results were

entered into an Excel spreadsheet for subsequent statistical analysis. To assess accuracy, Passing–Bablok linear regression analysis (to calculate the slope, intercept and correlation coefficient) and Bland–Altman analysis (to calculate mean bias,

Total white cell count (× 109/L) capillary v laboratory

16 14 12 10 8 6 4 POCT = 1.01 × Lab – 0.38 r = 0.76 n = 53

2

16 14 12 10 8 6 4

0 0

18

0

12 10 8 6 4 POCT = 0.80 × Lab – 0.33 r = 0.81 n = 53

2

8

6

4

2

POCT = 0.81 × Lab – 0.46 r = 0.88 n = 55

0 0

2

4 6 8 10 12 Laboratory neutrophils (× 109/L)

14

0

Lymphocytes (× 109/L) capillary v laboratory

4

3

2

POCT = 0.79 × Lab + 0.87 r = 0.75 n = 53

0 3 1 2 4 Laboratory lymphocytes (× 109/L)

5

2

4 6 8 10 Laboratory neutrophils (× 109/L)

12

Lymphocytes (× 109/L) venous v laboratory

5

POCT lymphocytes - venous (× 109/L)

POCT lymphocytes - capillary (× 109/L)

18

10

0

0

2 4 6 8 10 12 14 16 Laboratory total white cell count (× 109/L) Neutrophils (×109/L) venous v laboratory

12

POCT neutrophils - venous (× 109/L)

POCT neutrophils - capillary (× 109/L)

2 4 6 8 10 12 14 16 Laboratory total white cell count (× 109/L) Neutrophils (× 109/L) capillary v laboratory

14

1

POCT = 0.88 × Lab – 0.56 r = 0.87 n = 55

2

0

5

Total white cell count (×109/L) venous v laboratory

18 POCT total white cell count - venous (× 109/L)

POCT total white cell count - capillary (× 109/L)

18

4

3

2

1

0

POCT = 0.92 × Lab + 0.30 r = 0.84 n = 55

0

3 1 2 4 Laboratory lymphocytes (× 109/L)

5

Fig. 1 Comparison plots for HemoCue capillary and venous samples versus laboratory method (Sysmex XE-2100).

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

CORRESPONDENCE

Monocytes (× 109/L) capillary v laboratory

1.2

1.2 1 0.8 0.6 0.4 POCT = 0.23 × Lab + 0.35 r = 0.22 n = 53

0.2

1 0.8 0.6 0.4 POCT = 0.06 × Lab + 0.44 r = 0.00 n = 55

0.2 0

0 0

0.2

0.4 0.6 0.8 1 1.2 Laboratory monocytes (× 109/L)

0

1.4

Eosinphils (× 109/L) capillary v laboratory

1

0.8

0.6

0.4 POCT = 0.76 × Lab + 0.03 r = 0.87 n = 53

0.2

0

0.2

0.4 0.6 0.8 1 1.2 Laboratory monocytes (× 109/L)

1.4

Eosinphils (× 109/L) venous v laboratory

1.2

POCT eosinphils - venous (× 109/L)

1.2

POCT eosinphils - capillary (× 109/L)

Monocytes (× 109/L) venous v laboratory

1.4

POCT monocytes - venous (× 109/L)

POCT monocytes - capillary (× 109/L)

1.4

93

1

0.8

0.6

0.4

POCT = 0.97 × Lab + 0.00 r = 0.92 n = 55

0.2

0 0

0.2

0.4 0.6 0.8 1 Laboratory eosinphils (× 109/L)

1.2

0

0.2

0.4 0.6 0.8 1 Laboratory eosinphils (× 109/L)

1.2

Fig. 1 (Continued)

95% confidence intervals and mean % bias) were performed on the comparison patient data using the Analyse-It statistical software program (UK; http://analyse-it.com). Three patient samples (with differing total white cell counts of 3.5
109/L, 11.5
109/L and 23.5
109/L, designated low, normal and high) were analysed 20 times each on the same day (and within 12 h of collection) to calculate withinday imprecision. For imprecision, the coefficient of variation (CV%) was calculated for each sample and compared to the analytical goals for each cell type set by the Royal College of Pathologists Quality Assurance Program Pty Ltd (RCPA QAP). Of the 62 patients tested at the clinic, seven were excluded as venous samples did not reach the laboratory within their specified transport times and sample integrity could not be guaranteed. Two samples were excluded in the capillary comparison as an error relating to the image being out-of-focus was obtained on the POCT device and a repeat sample could not be obtained. The remaining 53 matched sets of data were analysed for the capillary POCT versus laboratory comparison, while 55 matched sets of data were available for venous POCT versus laboratory comparison. The 62 patients comprised 32 males and 30 females, while the age range of the patient group was 19–83 years.

The results of the Passing–Bablok analysis are shown in Fig. 1. For both the capillary and venous whole blood comparisons, the correlation coefficient (r) was >0.75 for all cell types except monocytes. Results of Bland–Altman analysis for capillary and venous POC samples versus the laboratory method are shown in Table 1. The mean percentage difference between capillary (y) and venous (x) POCT results for each cell type was: total WBC 2.8%, neutrophils –2.6%, lymphocytes 12.5%, monocytes 4.2% and eosinophils –10%. The results of within-day imprecision studies (n ¼ 20) conducted on three patient samples with differing total WBC counts are summarised in Table 2. For all total and differential cell counts with a mean cell number greater than 4
109/L, the imprecision was less than 5%. For cell counts with a mean cell number between 0.5–4
109/L, the imprecision was less than 15%. For individual cell counts with cell numbers <0.5
109/L, the imprecision was greater than 20% in all cases. High levels of imprecision were reported for monocytes and eosinophils due to the very low number of cells present; for example, a CV of 34% was reported but all repeat tests results were either zero or 0.1
109/L. The basophil count was not

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

94

Pathology (2015), 47(1), January

CORRESPONDENCE

Table 1

Summary of Bland–Altman analysis, comparing capillary and venous POCT results with laboratory results

Cell type (
109/L)

Sample type

WBC Count

Capillary Venous Capillary Venous Capillary Venous Capillary Venous Capillary Venous

Neutrophils Lymphocytes Monocytes Eosinophils

Table 2

Number

Mean lab value

Range of values

Mean bias

95% confidence intervals

Mean % Bias

53 55 53 55 53 55 53 55 53 55

7.90 7.80 4.60 4.65 2.47 2.28 0.54 0.52 0.29 0.30

3.9 to 14.2 3.6 to 12.9 1.3 to 7.7 1.9 to 10.4 1.3 to 4.1 1.2 to 3.8 0.1 to 1.3 0.2 to 1.2 0 to 0.9 0.1 to 1.1

0.38 0.45 0.66 0.46 0.39 0.12 0.1 0.11 0.043 0

0.83 to 0.07 0.73 to 0.16 0.97 to 0.35 0.68 to 0.24 0.26 to 0.53 0.02 to 0.23 0.17 to 0.03 0.19 to 0.03 0.072 to 0.015 0.03 to 0.02

4.8% 5.8% 14.3% 9.9% 15.8% 5.3% 18.5% 21.2% 14.8% 0.0%

Observed within-day imprecision for total WBC and their subtypes for three patient samples

Cell type

WBC

Neutrophils

Lymphocytes

Monocytes

Eosinophils

Sample (n ¼ 20)

L

N

H

L

N

H

L

N

H

L

N

H

L

N

H

Mean (x109/L) SD CV%

3.4 0.28 8.3

10.5 0.28 2.7

23.5 0.61 2.6

2.8 0.13 4.8

4.8 0.22 4.7

20.3 0.74 3.7

0.2 0.06 24.7

4.9 0.24 4.9

2.0 0.24 11.7

0.2 0.07 37.8

0.74 0.11 14.8

0.9 0.22 24.7

0.2 0.12 52.3

0.09 0.03 34.2

0.2 0.06 33.7

CV, coefficient of variation; H, high total WBC count; L, low total WBC count; N, normal total WBC count; SD, standard deviation.

included in the data analysis as none of the patient samples had a basophil count above 0.1
109/L (>1%). The analytical performance characteristics for the total WBC and differential cell counts were generally acceptable, except for monocytes. The results for basophils were not reported in this study due to low cell numbers in the samples tested. Basophils and monocytes counts are commonly imprecise, even for automated laboratory analysers,5 due to the relatively low numbers of cells present and similarities between monocytes and other cells that may interfere with the count (such as abnormal or atypical lymphocytes). Poor performance of monocyte counts on the HemoCue WBC Diff has also been reported previously in peer-reviewed publications in the Netherlands and United States.6–8 This device evaluation was performed in one of the most remote and challenging environments for POCT in Australia. Observed differences between POCT and laboratory results may result in part from long transport/turnaround times for venous samples to reach the laboratory (2000 km away). Supporting this premise, 11% of venous samples collected did not reach the laboratory in time to be accurately measured (>36 h). A limitation of this study was that there were very few acute patient presentations at the remote health service with low (<4
109/L) or high (>11
109/L) total WBC counts and therefore we cannot comment authoritatively on the accuracy of the HemoCue WBC Diff at these levels. The RAN performing the HemoCue tests commented that the analyser was ‘easy to use’ and the clinician involved commented that the ‘levels of imprecision observed would not change clinical judgement’. The logistical difficulties in transporting pathology samples from remote communities to the nearest pathology laboratory can often compromise the integrity of the sample, and lead to a patient’s results being suppressed. Being able to perform tests at the point of patient care (providing it is of equivalent analytical standard) can improve the operational effectiveness of pathology testing delivery. The ability to measure a total

WBC and a differential count on-site in a remote community provides a convenient and accessible service for the patient and obviates the need for follow-up visits by the patient. It also has clinical benefits for patient care in terms of timeliness of results enabling more rapid initiation of treatment of infections. The HemoCue WBC Diff was shown in this study to be a userfriendly POCT device by its clinical stakeholders and has demonstrated generally sound analytical performance for total WBC and differential count (except for monocytes), making it a useful surrogate analyser for use in remote communities.

Acknowledgements: The authors acknowledge the contribution of Darren Scott, SA Pathology, who assisted with the samples for the imprecision study. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Brooke A. Spaeth1 Mark D. S. Shephard1 Beverly McCormack2 Gary Sinclair2 1

Flinders University International Centre for Point-of-Care Testing, Flinders University, Adelaide, SA, and 2Central Australia Remote Health, Northern Territory Department of Health, NT, Australia Contact Brooke A. Spaeth. E-mail: [email protected] 1. Handin RI, Lux SE, Stossel TP, editors. Hematopoiesis. 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 2003; Chapter 7, Blood: principles and practice of hematology. 2. Higgins C. editor. Understanding Laboratory Investigations for Nurses and Health Professionals. 2nd ed. Oxford: Blackwell Publishing, 2007; Chapter 16, White cell count and differential.

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

CORRESPONDENCE

3. Shephard M. Point-of-care testing in Australia: the status, practical advantages, and benefits of community resiliency. Point of Care 2013; 12: 41–5. 4. Australian Bureau of Statistics. 2011 Census Counts – Aboriginal and Torres Strait Islander Peoples in Indigenous Regions. Cat. no. 2901.0. Canberra: Australian Bureau of Statistics, 2011. 5. Buttarella M, Plebani M. Automated blood cell counts. Am J Clin Pathol 2008; 130: 104–16. 6. Russcher H, van Deursen N, de Jonge R. Evaluation of the HemoCue WBC DIFF system for point-of-care counting of total and differential white cells in pediatric samples. Ned Tijdschr Klin Chem Labgeneesk 2013; 38: 140–1.

95

7. Johnsson E, Hockum S, Reed J. Novel POC analysis for determination of total and 5-part differential WBC count among a US population, in comparison to Beckman Coulter LH750. Point of Care 2014; 13: 12–4. 8. Lindberg S, Jo¨nsson I, Milsson M, et al. A novel technology for 5-part differentiation of leukocytes point-of-care. Point of Care 2014; 13: 27–30.

DOI: 10.1097/PAT.0000000000000202

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