Virchows Archiv

, Volume 456, Issue 1, pp 71–76 | Cite as

Multinucleated floret-like giant cells in sporadic and NF1-associated neurofibromas: a clinicopathologic study of 94 cases

  • Gaetano Magro
  • Paolo Amico
  • Giada Maria Vecchio
  • Rosario Caltabiano
  • Marine Castaing
  • Denisa Kacerovska
  • Dmitry V. Kazakov
  • Michal Michal
Original Article

Abstract

Multinucleated floret-like giant cells (MNFGCs), similar to those commonly observed in pleomorphic lipoma and giant cell fibroblastoma, have been occasionally reported in gynecomastia and neurofibromas from patients affected by neurofibromatosis type 1 (NF1). Accordingly, it has been suggested that their detection, especially in an otherwise typical neurofibroma, could be a morphological clue to diagnosis of NF1. The aim of the present study was the identification of MNFGCs in a large series (94 cases) of sporadic and NF1-associated neurofibromas, to assess if their presence may indeed be a morphological marker of NF1. Numerous MNFGCs, namely, those that were easily apparent at low magnification (×50 and ×100), were identified only in 5.3% of cases. In 18.1% of cases, a low number of these cells could be observed but only after a careful search, especially at higher magnification (×200 and ×400). Immunohistochemically, all MNFGCs were stained with vimentin and CD34, but not with S-100 protein. Interestingly, there was no statistically significant correlation between MNFGCs (presence or absence) and NF1 (p = 0.73), gender (p = 0.59), age (p = 0.43), and site of tumor (cutaneous vs deep-seated soft tissue; p = 0.27). Our clinicopathologic findings suggest that MNFGCs in an otherwise typical neurofibroma are not a reliable marker of NF1, likely representing a morphological reactive change of the indigenous dermal or endoneurial fibroblasts or dendritic cells in response to unknown microenvironmental stimuli.

Keywords

Neurofibroma Neurofibromatosis type I Multinucleated floret-like cells CD34 

Introduction

Neurofibroma is a relatively common benign peripheral nerve sheath tumor which can occur both sporadically or as part of neurofibromatosis type I (NF1) [1, 2]. Several histologic variants have been recognized over time, including classic, plexiform, cellular, epithelioid, myxoid, hyalinized, pacinian, pigmented, and granular cell variant [3, 4, 5]. In addition, some cases of neurofibroma with epithelial, rhabdomyomatous and lipomatous differentiation, or dendritic cells with pseudorosette formation have also been described [1, 6, 7, 8]. Although neurofibroma may present a wide morphological spectrum, only the plexiform neurofibroma, a tumor mass grossly consisting of tortuous expanded nerved branches, is virtually pathognomonic of NF1 [1, 2]. Apart plexiform neurofibroma, which is present in up 30% of cases of NF1 patients [9], most neurofibromas occurring in the context of this syndrome exhibit a morphology indistinguishable from those usually encountered in sporadic cases.

One of us (GM) first reported the presence of numerous multinucleated floret-like giant cells (MNFGCs), similar to those typically observed in pleomorphic lipoma and giant cell fibroblastoma, in a diffuse type, cutaneous neurofibroma from a patient affected by NF1 [10]. This unusual finding was subsequently confirmed in other three neurofibromas from patients with NF1 [11, 12, 13]. Notably, one of these patients presented MNFGCs in all 20 cutaneous neurofibromas which had been surgically excised at different periods [13]. The same study also documented occasional MNFGCs in sporadic neurofibromas, raising the question whether the presence of such cells in an otherwise typical neurofibroma may represent a reliable morphological marker of NF1 [13]. To the best of our knowledge, there is no study in the literature, which compares the presence of MNFGCs in sporadic neurofibromas versus NF1-associated neurofibromas. Accordingly, the aim of the present study was the identification of MNFGCs in a series of 94 cases of neurofibromas occurring either sporadically or in NF1 patients to establish if these cells could serve as a morphological clue to diagnosis of NF1. Moreover, a clinicopathologic correlation between the presence of MNFGCs with age, gender, site, histological type, and malignant peripheral nerve sheath tumors (MPNSTs) arising from a pre-existing neurofibroma is provided.

Materials and methods

A retrospective review of all lesions coded as “neurofibroma” have been selected from the surgical pathology files at the Anatomic Pathology, Department G.F. Ingrassia, University of Catania, Catania, Italy and Sikl’s Department of Pathology, Charles University Hospital, Pilsen, Czech Republic and Bioptical Laboratory, Pilsen, Czech Republic. All available slides stained with hematoxylin and eosin were reviewed. For all cases, the clinical data were obtained from pathology reports. Neurofibromas manifested a wide anatomic distribution including arm, forearm, wrist, breast, head, eyelid, neck, knee, back, scalp, abdomen, chest, thigh, vulva, and paravertebral region.

All sections were meticulously scrutinized for the presence of MNFGCs, namely, medium-sized to large cells exhibiting multiple nuclei arranged either randomly or in a wreath-like configuration, similar to floret-like cells typically observed in pleomorphic lipoma and giant cell fibroblastoma [1]. Their quantification was performed semiquantitatively, and the cases examined were arbitrarily divided into two groups: “neurofibromas with high number of MNGFCs” and “neurofibromas with low number of MNGFCs.” In the former group, MNFGCs were readily apparent at low magnification (×50 and ×100), whereas in the latter group, they were identified only after a careful search, especially at higher magnification (×200 and ×400).

In addition, according to the site of their origin, neurofibromas of the present series were divided into cutaneous or deep-seated soft tissue neurofibromas. The former lesions were superficially located, with involvement of dermis and/or subcutis and were further histologically subdivided into diffuse or nodular type, according to the growth pattern exhibited by tumor [1]. The deep-seated soft tissue neurofibromas were represented by nodular masses with no evidence of skin or subcutis involvement.

Representative paraffin-embedded tissue blocks from all cases (22 of 94) containing MNFGCs were cut and processed for immunohistochemical analyses. One of these cases was previously published as case report [10]. Immunohistochemical studies were performed with the labeled streptavidin–biotin peroxidase detection system using the Ventana automated immunostainer (Ventana Medical Systems, Tucson, AZ). The antibodies used included vimentin (clone V9; dilution 1:50), CD34 (Clone Q Bend-10; dilution 1:50), S-100 protein (polyclonal; 1:5000), epithelial membrane antigen (clone GP1.4; dilution: 1:50), pancytokeratins (MNF116: dilution 1:100), and CD68 (clone KP1; dilution: 1:50; all purchased from DakoCytomation, Glostrup, Denmark).

Correlations between MNFGCs and the characteristics of the study population were performed through chi-square test and exact Fisher test when requested. p values < 0.05 were considered statistically significant.

Results

All relevant clinical features in our patients are summarized in Table 1. There were 41 men and 50 women. Gender was not available in three cases. The patients ranged in age from 10 to 88 years, with an average age of 45.1. Age was not available in seven cases. Most neurofibromas (62.8%) were cutaneous, exhibiting a nodular or a diffuse pattern in 14.9% and 47.9% of the cases, respectively (Table 1). Deep-seated soft tissue neurofibromas represented 37.2% of the cases, and among these, a plexiform pattern was observed in 17.1% of cases.
Table 1

Selected characteristics of the study population

 

Patients (total number = 94)

n

Percentage

Gender

Men

41

43.6

Women

50

53.2

Missing

3

3.2

Age (years)

<45

46

48.9

≥45

41

43.6

Missing

7

7.5

NF1

No

44

46.8

Yes

50

53.2

MNFGCs

High number

5

5.3

Low number

17

18.1

Absence

72

76.6

Site of tumor

Diffuse cutaneous

45

47.9

Nodular cutaneous

14

14.9

Deep-seated soft tissues

35

37.2

MPNST

No

86

91.5

Yes

8

8.5

n number of patients, NF1 neurofibromatosis type I, MNFGCs multinucleated floret-like giant cells, MPNST malignant peripheral nerve sheath tumors

An established history of NF1 was known in 50 patients (53.2%), while the remaining 44 patients (46.8%) had no clinical signs of this syndrome (Table 1). All patients without history of NF1 clinically presented with a single lesion. On the contrary, multiple cutaneous lesions (ranging from two to five) were available in three cases of patients with NF1. The six cases of plexiform type, deep-seated soft tissue neurofibromas were all associated with NF1.

Histologically, all tumors of the present study shared the typical morphological features of neurofibroma, namely, a proliferation of spindle-shaped cells with scant cytoplasm and elongated, often wavy nuclei. Extracellular matrix varied greatly from myxoid to fibrous. Among the NF1 patients, a MPNST was identified in eight cases (8.5%; Table 1). In all these cases, a pre-existing neurofibroma could be histologically identified, with a plexiform pattern detected in three cases. All MPNSTs but one were histologically classified as high grade, according to World Health Organization classification [2].

MNFGCs were identified in 23.4% of the cases (Table 1). Neurofibromas which contained high or low number of MNFGCs represented 5.3% and 18.1% of the series, respectively (Table 1; Figs. 1 and 2). MNFGCs were morphologically similar both in sporadic and NF1-associated neurofibromas (Fig. 3a, b). They were found to be intermingling with neurofibroma tumor cells, from which they were easily distinguished by the presence of multiple nuclei arranged randomly or in a wreath-like configuration (Fig. 3a, b). Some of the nuclei were large in size, vesicular, and with a prominent nucleolus (Fig. 4). MNFGCs had a relatively abundant eosinophilic cytoplasm with dendritic-like processes (Figs. 3a, b and 4). Notably, one case of NF1-associated neurofibroma focally exhibited areas with numerous MNFGCs which lined pseudovascular spaces, namely, optically empty spaces without erythrocytes, evoking a giant cell fibroblastoma (Fig. 5). Immunohistochemically, MNFGCs were stained with vimentin and CD34, while no immunoreactivity was obtained with all other antibodies. CD34 staining highlighted the dendritic-like cytoplasmic processes of the tumor cells (Fig. 6).
Fig. 1

Neurofibroma from a NF1 patient: numerous MNFGCs are easily identified at low magnification

Fig. 2

Neurofibroma from a NF1 patient: only rarely MNFGCs are evident at higher magnification

Fig. 3

MNFGCs are morphologically similar in both sporadic neurofibromas (a) and NF1-associated neurofibromas (b): large-sized cells with eosinophilic cytoplasm and multiple nuclei with a floret-like arrangement

Fig. 4

Sporadic neurofibroma. MNFGCs exhibiting vesicular nuclei with a prominent nucleolus

Fig. 5

Neurofibroma from a NF1 patient: area with pseudovascular spaces lined by MNFGCs, closely reminiscent of giant cell fibroblastoma

Fig. 6

Immunostaining with CD34 highlighting cytoplasmic dendritic-like processes

Correlations between clinicopathologic features and MNFGCs are summarized in Table 2. Interestingly, there was no statistically significant association between NF1 and MNFGCs (high, low number, or absence; p = 0.93; Table 2). Similar results were obtained with the dichotomization, presence or absence of MNFGCs (p = 0.73; not shown). Only in one of the three cases of NF1-associated neurofibromas with multiple cutaneous lesions was a low number of MNFGCs found in all surgically excised samples (two cutaneous neurofibromas, diffuse type). The remaining two cases were devoid of MNFGCs.
Table 2

Correlations between MNFGCs and characteristics of the study population

 

Patients (total number = 94)

High number of MNFGCs (n = 5)

Low number of MNFGCs (n = 17)

Absence of MNFGCs (n = 72)

p

n (%)

n (%)

n (%)

 

Gender

Men

3 (60)

9 (52.9)

29 (40.3)

0.59

Women

2 (40)

8 (47.1)

40 (55.6)

 

Missing

0

0

3 (4.2)

 

Age (years)

<45

2 (40)

7 (41.2)

37 (51.4)

0.43

≥45

3 (60)

10 (58.8)

28 (38.9)

 

Missing

0

0

7 (9.7)

 

NF1

No

2 (40)

9 (64.3)

33 (45.8)

0.93

Yes

3 (60)

8 (35.7)

39 (54.2)

 

Site of tumor

Cutaneous

2 (40)

13 (76.5)

44 (61.1)

0.27

Deep-seated soft tissues

3 (60)

4 (23.5)

28 (38.9)

 

n number of patients, NF1 neurofibromatosis type I, MNFGCs multinucleated floret-like giant cells, MPNST malignant peripheral nerve sheath tumors

With regard to clinical features, no correlation was found between MNFGCs and gender (p = 0.59), age (<45 vs ≥45; p = 0.43), and site of tumor (cutaneous vs deep-seated soft tissues; p = 0.27; Table 2). Similarly, analyzing the presence or absence of MNFGCs, the association between MPNSTs and MNFGCs in pre-existing neurofibroma was not statistically significant, with only in one out of eight cases (p = 0.68; not shown), containing a low number of these cells.

It was noteworthy that among cutaneous neurofibromas, only diffuse type contained MNFGCs, whereas such cells were lacking in all cases of nodular type neurofibromas (not shown). In the former, MNFGCs were identified in 33.3% of the cases (not shown), with an associated NF1 history in 60% of cases (not shown). Only 20% of deep-seated soft tissue neurofibromas contained MNFGCs (Table 2), and among these, 28.6% of cases were associated with NF1 (not shown).

Discussion

MNFGCs cells have been well documented in the stroma of several anatomic locations, including lower female genital tract, testis, urinary bladder, anus, breast, and skin [14, 15, 16, 17]. Similar cells have also been described as part of some soft tissue tumors, including pleomorphic lipoma, giant cell fibroblastoma, giant cell collagenoma, and giant cell angiofibroma/solitary fibrous tumor [1]. Interestingly, MNFGCs have been occasionally reported in gynecomastia from patients bearing NF1, suggesting the possibility that their detection could help in recognizing patients with this syndrome [18, 19]. The possibility of a link between MNFGCs and NF1 was also supported by the detection of these cells and also in some cases of neurofibromas from NF1 patients [10, 11, 12, 13]. This latter finding led some authors to postulate that MNFGCs in an otherwise typical neurofibroma could be a morphological marker of NF1 [13]. To confirm this hypothesis, we first studied the incidence and quantification of MNFGCs in a large series of NFs occurring either sporadically or in the context of NF1.

Our results showed that the occurrence of MNFGCs in neurofibromas is relatively low, being found in 23.4% (22 of 94) of cases. Among these cases, most (17 of 22) were represented by neurofibromas containing only a few numbers of MNFGCs which would likely have been easily overlooked in daily practice. In fact, these cells could be identified only after a meticulous search, especially at higher magnification (×200 and ×400). Only in 5.3% of the series were MNFGCs found to be numerous and readily identifiable at low magnification (×50 and ×100). A previous study showed that 20 neurofibromas, all surgically excised from a 54-year-old patient affected by NF1, contained numerous MNFGCs [13]. This intriguing finding was compared with a series of 53 sporadic neurofibromas (NF1-associated neurofibromas were not included) [13]. The authors found MNFGCs in 5.6% of sporadic tumors, and interestingly, these cells were lower in number (i.e., identified only after a careful scrutiny), when compared with those found in the single patient with NF1 [13]. Therefore, it was suggested that a high number of these cells might be a clue to the presence of NF1 [13].

Based on the results of our study, we were not able to confirm this hypothesis. By comparing sporadic versus NF1-associated neurofibromas, there is no statistically significant association between MNFGCs and NF1. Moreover, the high number of MNFGCs is a reliable feature in distinguishing between NF1 and non-NF1 patients, because the cells were found in three versus two cases with or without NF1, respectively. In contrast to a previous report [13], our study failed to confirm the presence of numerous MNFGCs in multiple neurofibromas excised from the same NF1 patient. In addition, by analyzing other clinicopathologic characteristics, the presence of MNFGCs did not significantly correlate with patients’ age, gender, site (cutaneous vs deep-seated soft tissues), and MPNSTs arising from pre-existing neurofibroma.

The histogenesis of MNFGCs in neurofibroma is still to be established. The present study shows that these cells are not integral part of neurofibroma, as they are significantly identified only in 5.3% (five of 94) of cases, without differences between sporadic and NF1-associated cases. Morphologically similar, if not identical, MNFGCs have been described in the stroma of benign fibroepithelial neoplasms of the breast, mesodermal fibroepithelial polyps of the lower female genital tract, fibroepithelial polyps of anus, allergic-type nasal polyps, and bladder with inflammatory or neoplastic (urothelial carcinoma) diseases [15, 17, 20, 21]. As previously suggested for the above mentioned pathological conditions [17], it is likely to postulate that MNFGCs occasionally encountered in neurofibromas may merely represent a morphological reactive change of the indigenous dermal or endoneurial fibroblasts or dendritic cells in response to unknown stimuli. Immunohistochemical analysis, revealing exclusively vimentin and CD34 expression, would support the fibroblastic or dendritic nature of MNFGCs. Interestingly, some authors noted a close relationship between mast cells and MNFGCs, suggesting a potential role of the former in the formation of the latter cells [17]. Whether the well-known presence of mast cells in neurofibromas [1] could be closely related to that of MNFGCs is an intriguing hypothesis that deserves further clarification which is beyond the scope of the present paper.

Although not statistically significant, it was noteworthy that MNFGCs in our series were more frequently found in cutaneous versus deep-soft tissue neurofibromas. Among cutaneous neurofibromas, only diffuse type contained MNFGCs, whereas no case of nodular-type neurofibroma exhibited such cells. This finding would suggest that the type of growth pattern, namely, diffuse rather than nodular, may induce native dermal CD34+ fibroblasts/dendritic cells to acquire a multinucleated and floret-like appearance.

In conclusion, the presence of MNFGCs in an otherwise typical neurofibroma is not a significant indicator of NF1. Awareness that neurofibromas may contain MNFGCs is important for pathologist to avoid confusion with other tumors, especially giant cell fibroblastoma. Although the occasional presence of numerous MNSFCs in a neurofibroma may represent an alarming feature, these cells should not be misinterpreted as atypical tumor cells suggesting the possibility of atypical neurofibroma or MPNSTs arising from a pre-existing neurofibroma [1, 22]. This potential confusion should be raised because atypical tumor cells in a neurofibroma or MPNST may occasionally have multinucleation [1, 23]. Unlike MNSFCs, however, atypical tumor cells have hyperchromatic and irregular nuclei and may also exhibit prominent intranuclear pseudoinclusions [1, 23]. The presence of numerous atypical tumor cells uniformly distributed in a neurofibroma, in association with an increased cellularity and mitotic figures, is a reliable indicator of malignant transformation [1, 23]. Lastly, immunohistochemistry, showing a vimentin+/CD34+/S100/CD68 profile in MNSFCs, is helpful in ruling out atypical tumor cells which are characteristically S-100 protein positive.

Notes

Conflict of interest

The authors declare that they have no conflict of interest.

References

  1. 1.
    Weiss SW, Goldblum JR (2008) Enzinger and Weiss’s soft tissue tumors. Mosby, St LouisGoogle Scholar
  2. 2.
    Scheithauer BW, Louis DN, Hunter S et al (2000) Malignant peripheral nerve sheath tumour (MPNST). In: Louis DN, Ohgaki H, Webster OD, Cavenee WK (eds) Pathology and genetics. Tumours of the central nervous system. IARC, Lyon, pp 155–162Google Scholar
  3. 3.
    Megahed M (1994) Histopathologic variants of neurofibroma: a study of 114 lesions. Am J Dermatopathol 16:486PubMedCrossRefGoogle Scholar
  4. 4.
    Gonzalez-Vela MC, Val-Bernal JF, Gonzalez-Lopez MA et al (2006) Pure sclerotic neurofibroma: a neurofibroma mimicking sclerotic fibroma. J Cutan Pathol 33:47CrossRefPubMedGoogle Scholar
  5. 5.
    MacLennan SE, Melin-Aldana H, Yakuboff KP (1999) Pacinian neurofibroma of the hand: a case report and literature review. J Hand Surg [Am] 24:413–416CrossRefGoogle Scholar
  6. 6.
    Val-Bernal JF, Gonzales-Vela MC (2005) Cutaneous lipomatus neurofibroma: characterization and frequency. J Cutan Pathol 32:274–279CrossRefPubMedGoogle Scholar
  7. 7.
    Plaza JA, Wakely PE, Suster S (2006) Lipoblastic nerve sheath tumors. Report of a distinctive variant of neural soft tissue neoplasm with adipocytic differentiation. Am J Surg Pathol 30:337–344PubMedGoogle Scholar
  8. 8.
    Kazakov DV, Vanecek T, Sima R et al (2005) Dendritic cell neurofibroma with pseudorosettes lacks mutations in exons 1–15 of the neurofibromatosis type 2 gene. Am J Dermatopathol 27:286CrossRefPubMedGoogle Scholar
  9. 9.
    Ferner RE (2007) Neurofibromatosis 1 and neurofibromatosis 2: a twenty first century perspective. Lancet Neurol 6:340CrossRefPubMedGoogle Scholar
  10. 10.
    Magro G, Scavo S, Ruggieri M (2002) Floretlike multinucleated giant cell in a neurofibroma from a patient with NF1: an unusual finding for such a tumor. Virchows Arch 441:525–526CrossRefPubMedGoogle Scholar
  11. 11.
    Shaktawat SS, Golka D (2007) Floret-like multinucleated giant cells in neurofibroma. Diagn Pathol 8:2–47Google Scholar
  12. 12.
    Swick BL (2008) Floret-like multinucleated giant cells in a neurofibromatosis type 1-associated neurofibroma. Am J Dermatopathol 30:632–634CrossRefPubMedGoogle Scholar
  13. 13.
    Taungjaruwinai WM, Golbert LJ (2009) Multinucleate giant cells in neurofibromas: a clue to the diagnosis of neurofibromatosis. J Cutan PatholGoogle Scholar
  14. 14.
    Groisman GM, Amar M, Polak-Charcon S (2000) Multinucleated stromal cells of the anal mucosa: a common finding. Histopathology 36:224–228CrossRefPubMedGoogle Scholar
  15. 15.
    Groisman GM, Polak-Charcon S (1998) Fibroepithelial polyps of the anus: a histologic, immunohistochemical, and ultrastructural study, including comparison with the normal anal subepithelial layer. Am J Surg Pathol 22:70–76CrossRefPubMedGoogle Scholar
  16. 16.
    Ohtsuki Y, Furihata M, Iwata J et al (2000) Multinucleated giant cells in submucosal layer of human urinary bladder: an immunohistochemical and electron microscopic study. Pathol Res Pract 196:293–298PubMedGoogle Scholar
  17. 17.
    Pitt MA, Roberts IS, Agbamu DA et al (1993) The nature of atypical multinucleated stromal cells: a study of 37 cases from different sites. Histopathology 23:137–145CrossRefPubMedGoogle Scholar
  18. 18.
    Damiani S, Eusebi V (2001) Gynecomastia in type-1 neurofibromatosis with features of pseudoangiomatous stromal hyperplasia with giant cells. Report of two cases. Virchows Arch 438:513–516CrossRefPubMedGoogle Scholar
  19. 19.
    Zamecnik M, Michal M, Gogora M et al (2002) Gynecomastia with pseudoangiomatous stromal hyperplasia and multinucleated giant cells. Association with neurofibromatosis type 1. Virchows Arch 441:85–87CrossRefPubMedGoogle Scholar
  20. 20.
    Powell CM, Cranor ML, Rosen PP (1994) Multinucleated stromal giant cell in mammary fibroepithelial neoplasm. A study of 11 patients. Arch Pathol Lab Med 118:912–916PubMedGoogle Scholar
  21. 21.
    Ryska A, Reynolds C, Keeney GL (2001) Benign tumors of the breast with multinucleated stromal giant cell. Immunohistochemical analysis of six cases and review of the literature. Virchows Arch 439:768–775PubMedGoogle Scholar
  22. 22.
    Lin BT, Weiss LM, Medeiros LJ (1997) Neurofibroma and cellular neurofibroma with atypia: a report of 14 tumors. Am J Surg Pathol 21:1443–1449CrossRefPubMedGoogle Scholar
  23. 23.
    Jokinen CJ, Argenyi ZB (2009) Atypical neurofibroma of the skin and subcutaneous tissue: clinicopathologic analysis of 11 cases. J Cutan Pathol. doi:10.1111/j.1600-0560.2009.01293.x Google Scholar

Copyright information

© Springer-Verlag 2009

Authors and Affiliations

  • Gaetano Magro
    • 1
  • Paolo Amico
    • 1
  • Giada Maria Vecchio
    • 1
  • Rosario Caltabiano
    • 1
  • Marine Castaing
    • 2
  • Denisa Kacerovska
    • 3
  • Dmitry V. Kazakov
    • 3
  • Michal Michal
    • 3
  1. 1.Dipartimento G.F. IngrassiaPoliclinico Universitario G. Rodolico, Anatomia Patologica, Università di CataniaCataniaItaly
  2. 2.G.F Ingrassia DepartmentIntegrated Tumor Registry of Messina-Catania-Siracusa, Hygiene and Public Health InstituteCataniaItaly
  3. 3.Sikl’s Department of PathologyCharles University Medical Faculty HospitalPilsenCzech Republic

Personalised recommendations