Chinese Medical Journal 2007;120(10):876-881
Clinicopathologic features of pleomorphic hyalinizing angiectatic tumor of soft parts
KE Qi, Erbolat, ZHANG Hong-ying, BU Hong, LI Sheng, SHI Dan-ni, YANG Guang-hua, CHEN Hui-jiao, WEI Bing
KE Qi (Department of Pathology, West China Hospital, Sichuan University, Chengdu 610041, China)
Erbolat (Department of Pathology, Kuitun Hospital, Yili 833200, Xinjiang，China)
ZHANG Hong-ying (Department of Pathology, West China Hospital, Sichuan University, Chengdu 610041, China)
BU Hong (Department of Pathology, West China Hospital, Sichuan University, Chengdu 610041, China)
LI Sheng (Department of Pathology, Langzhong Hospital, Langzhong 637400, China)
SHI Dan-ni (Department of Pathology, Luzhou Hospital, Luzhou 646000, China)
YANG Guang-hua (Department of Pathology, West China Hospital, Sichuan University, Chengdu 610041, China)
CHEN Hui-jiao (Department of Pathology, West China Hospital, Sichuan University, Chengdu 610041, China)
WEI Bing (Department of Pathology, West China Hospital, Sichuan University, Chengdu 610041, China)Correspondence to:ZHANG Hong-ying,Department of Pathology, West China Hospital, Sichuan University, Chengdu 610041, China (Tel: 86-28-85423847. Fax:86-28-85422699. E-mail:firstname.lastname@example.org)
Background Pleomorphic hyalinizing angiectatic tumor (PHAT) of soft parts is a rare soft tissue tumor, which is generally considered low-grade. To distinguish the tumor from other soft tissue lesions, we analyzed the clinicopathologic and ultrastructural features, immunophenotypes, and flow cytometric DNA ploidy of PHAT in 9 cases.
Methods PHAT specimens were collected from 9 patients with PHAT from 1990 to 2004. Each specimen was cut into pieces and stained with hematoxylin-eosin, phosphotungstic acid-hematoxylin, Prussian blue, and Masson trichrome, respectively. Immunohistochemical stains for vimentin, S-100 protein, CD34, CD31, CD99, VEGF, desmin, CD117, α-SMA, and MIB-1 were performed with the Envision system. Flow cytometry was used in four specimens, two of which were observed by electron microscopy.
Results In the 9 cases, the PHAT occurred at the lower extremity in 2 patients, inguinal in 2, waist in 1, forearm in 1, buttock in 1, foot in 1, and the chest wall in 1. All the lesions presented in the superficial subcutaneous tissues. Follow-up data were available in 7 of the patients, among whom 2 (28.6%) had recurrence after primary therapy. Microscopically, typical PHAT was characterized by sheet-like proliferation of spindle or pleomorphic cells and clusters of thin-walled hyalinized cstatic vessels. In some areas of the tumor, hemosiderin-laden spindle cells, numerous small single vessels, and myxoid extracellular matrix could be identified, indicating an “atypical PHAT”. Mitotic figures were rare in all the cases. In 5 of the 9 patients (55.6%), the tumor was typical PHAT; and in the other 4 (44.4%), typical and atypical PHAT coexisted. Immunohistochemically, the neoplastic cells were positive for vimentin, CD34, CD99, and VEGF, but negative for S-100 protein, desmin, SMA, and CD31. In all the cases, the MIB-1 proliferative activity of the neoplastic cells was lower than 2%. Ultrastructural analysis did not reveal any evidence of specific differentiation. Aneuploidy was not detected by flow cytometry.
Conclusions Histologically, typical PHAT is characterized by spindle and pleomorphic cells associated with an angiectatic vasculature. The neoplastic cells often express vimentin and CD34, and may be positive for CD99 and VEGF. Ultrastructurally, the tumor usually has no specific differentiation. The low MIB-1 index and the absence of aneuploidy in PHAT indicate a non-malignancy. However, we consider the tumor as a borderline neoplasm because of its aggressive behaviour, and suggest wide local resection with tumor-free margin for the treatment of the disease.
Pleomorphic hyalinizing angiectatic tumor (PHAT) of soft parts is a rare soft tissue neoplasm, which was first described in 1996 by Smith and colleagues.1 The tumor usually occurs within the subcutaneous tissues, especially in the lower limbs of adults of either sexes. It is often mistaken for other soft tissue lesions, such as malignant fibrous histiocytoma and schwannoma. From 1990 to 2004, we collected 9 cases of PHAT and analyzed the clinicopathologic, ultrastructural, and flow cytometric features, and immunophenotypes of the tumor to distinguish it from the other soft tissue neoplasms.
From 1990 to 2004, we collected data from 9 cases with PHAT from the routine surgical files of the West China Hospital (7 cases, Chengdu, China) and the consultation files of the Department of Pathology of the West China Hospital (2 cases, who underwent surgical treatments at the Langzhong Hospital and Luzhou Hospital in Sichuan Province, respectively).
The clinical and follow-up informations of the patients were obtained from the records of the three hospitals. The age, sex, symptoms, preoperative duration, tumor location, size, and depth, treatments, additional treatments (e.g. radiotherapy), outcomes, and follow-up period were recorded and analyzed.
PHAT tissues were obtained from the 9 cases, fixed in 4% buffered-formaldehyde, embedded in paraffin, and then cut into 4-μm thick pieces. The sections were stained with hematoxylin-eosin, phosphotungstic acid-hematoxylin, Prussian blue, and Masson trichrome, respectively.
The PHAT tissue sections were treated with aminopropyl- triethoxy-silane (Sigma, UK) and dried overnight at 37˚C. 3% alcoholic hydrogen peroxide was used to block endogenous peroxidase activity (10 minutes). Heat-induced epitope retrieval was performed by microwaving for 12 minutes in 10 mmol/L citrate buffer at pH 6.0. Then the sections were immunostained by vimentin (clone V9, dilution 1:200), S-100 protein (polyclonal, 1:100), CD34 (clone Qbend10, 1:100), CD31 (clone JC70A, 1:100), CD99 (MIC-2, clone 12E7, 1:100), CD117 (c-kit, polyclonal, 1:50), VEGF (1:150), desmin (D33, 1:100), α-SMA (clone 1A4, 1:100), and Ki-67 (clone MIB-1, 1:150), respectively (all the primary antibodies were purchased from the Dako Co., USA). Tumor proliferation was assessed by MIB-1 immunostaining. At least 2000 cells were counted to determine the proportion of tumor cells expressing MIB-1 antigen. The immunostaining were performed with an Envision kit on a Dako automatic stainer (Dako, USA) with 3, 3’-diaminobenzidine used as the chromogen, and hematoxylin as the counterstain. Antibody binding was visualized and photographed using a light microscope. Positive and negative controls were included in each run.
Nuclear and cytoplasmic staining with S-100 protein, nuclear staining with MIB-1, membranous staining with CD34, CD31, CD99, CD117, and VEGF, and cytoplasmic staining with the other markers were considered positive.
Electron microscopy was performed on the cases 4 and 5. The specimens were rinsed in carbodylate buffer, refixed in 2.5% glutaraldehyde, post-fixed in 1% osmium tetroxide, embedded in epoxy resin, and then stained with uranyl acetate and lead citrate, and examined under a H600-Ⅳ transmission electron microscope (Hitachi, Japan). For each case, 3 to 8 micrographs were taken (6 on average).
Flow cytometry was performed on the cases 4, 5, 6, and 8 using a Facscan flow cytometer (EPICS ELITE ESP, USA). The sections (50-μm thick) were deparaffinized by xylene and rehydrated in ethanol-distilled water solutions, and then treated with pepsin and RNase. Propidium iodide was used to stain the nuclei, which were then run on a flow cytometer to determine the DNA content at wavelengths between 488 and 610 nm. In each sample, nuclear DNA content was measured by counting 5000 cells in 4 sections at least.
Among the 9 patients with PHAT, 5 were female and 4 were male, with an average age of 45.7 years (range, 19−76). Most of the patients presented with a slowly growing painless mass; and 6 of them had had the mass for more than 1 year before operation. The size of the tumors ranged from 1.4 to 7.2 cm in diameter (average, 4.3 cm). All the tumors were located within the superficial subcutaneous tissues at the lower extremities (2 patients), inguinal (2), waist (1), forearm (1), buttock (1), foot (1), and chest wall (1). A mean of 60.6 months follow-up was available in 7 of the 9 patients (range, 23− 123). During the follow-up, 2 patients had local recurrence, no one developed local or distal metastasis (Table).
Table. Clinical information of the 9 patients with PHAT
Grossly, the lesion was lobulate in cases 1, 3, 4, 6, and 9; and was round or ovoid in the others. None of the tumors was encapsulated; 5 of them (cases 4, 6, 7, 8, and 9) had diffusely infiltrative margin, and 4 had well-circumscribed border. Most of the lesions were white-tan to maroon in the cut surface with foci of hemorrhage. The tumors were soft or rubbery; one of them had cystic change (case 1).
Histologically, all the tumors grew infiltratively and were characterized by the presence of alternating vascular and cellular areas. The most prominent feature of these lesions was the various sized thin-walled ectatic blood vessels (Fig. 1A), which tended to cluster, presented with varied shapes, and distributed at any part of the tumors. The ectatic vessels were lined by flat endothelial cells with a thick subjacent rim of amorphous eosinophilic material (Fig. 1B). Phosphotungstic acid hematoxylin (PTAH) and Masson trichrome stainings showed that the eosinophilic material was composed of fibrin that surrounded by collagen. In some areas, the amorphous eosinophilic material extended from the vessels into the hyalinized stroma. Some of the vessels were filled with thrombus/recanalized thrombus and surrounded by hemorrhage. The presence of neoplastic cells was another feature of these lesions. In a low-power field, the tumor cells were arranged in sheets or fascicles between the angiectatic vessels without obvious differentiation. The neoplastic cells were spindle or ovoid, pleomorphic, and varied in size with a small amount of eosinophilic cytoplasm and hyperchromatic large striking pleomorphic nuclei, mimicking malignant fibrous histiocytoma (Fig. 2). Mitotic figures were very rare (<1/50 HPF) despite of the prominent pleomorphism. Some of the neoplastic cells had larger nuclei with prominent nuclear pseudoinclusions, which are the features reminiscent of schwannoma (Fig. 3), or intracytoplasmic hemosiderin.
In some areas of the tumor, in addition to the typical features of PHAT, we also noted relatively monomorphic components, i.e. “atypical PHAT”, which were composed of generally bland spindle neoplastic cells with wavy nuclei scattered in a variably myxoid background (Fig. 4). Neoplastic cells were arranged in fascicles and numerous small single vessels could be observed. Notably, by Prussian blue staining, we found that the neoplastic spindle cells adjacent to the ectatic vessels contained variable amounts of intracytoplasmic hemosiderin (Fig. 5). Moreover, inflammatory cells including mast plasma cells, and lymphocytes, were identified in all of the lesions. The features described above are similar to those of the “hemosiderotic fibrohistiocytic lipomatous lesion (HFLL)”.2
In 5 of our cases (cases 1, 2, 4, 5, and 7; 55.6%), typical PHAT accounted for over 80% of the lesion. In case 7, numerous small single vessels, and myxoid extracellular matrix were noted, with some features of atypical PHAT. And in case 4, original resection margins were positive. In the other 4 cases (cases 3, 6, 8, and 9), typical and atypical PHAT coexisted in variable proportions. In 3 of them, typical PHAT were located at the center of the lesions, while the atypical components mostly scattered in the periphery of the lesions (Fig. 6A). However, in another case (case 8), typical and atypical components were admixed intimately. In some cases, we also noted that atypical PHAT infiltrating into the surrounding fat tissues (Fig. 6B). In case 8, the recurrent lesion was almost composed of typical PHAT only, while the primary lesion consisted of both typical and atypical PHAT. In cases 3, 6, and 8, original resection margins were positive. None of the tumors was composed of atypical PHAT only.
The neoplastic cells were diffusely and strongly positive for vimentin, CD34, CD99, and VEGF (Figs. 7−9), and negative for S-100 protein, desmin, CD31, and α-SMA. The mast cells in the background were positive for CD117. In all the cases, the proliferative activity with MIB-1 was lower than 2%.
Fig. 1. A: Thin-walled ectatic blood vessels with red blood cells in the vascular lumen (case 1, HE, original magnification×100). B: Thin-walled angiectatic vessels lined by endothelium with a thick rim of eosinophilic material (case 2, HE, original magnification×200).
Ultrastructural analysis was performed on cases 4 and 5 (retrieved from paraffin block). We found that there were intermediate-sized filaments in the cytoplasm of the neoplastic cells, which was the feature of the fibroblasts; the intranuclear inclusions were formed by trapping cytoplasm; incomplete basal lamina was occasionally observed; and collagen bundles and some foci of long-spacing collagen were identified. No evidence of specific differentiation was noted.
Flow cytometric analysis of nuclear DNA content performed on four cases (cases 4, 5, 6, and 8) showed that aneuploidy DNA histogram were absent in 3 cases (cases5, 6, and 8). In the other case (case 4), the flow cytometry was failed because of too much cellular pieces.
Clinical characteristics of PHAT
PHAT, an independent entity of soft tissues, is a kind of non-metastasizing tumor. 3 To our knowledge, besides our 9 cases, there are only 69 cases reported in English literature.1,4-17 Of all the 78 patients, we reviewed 76 cases (42 women (55.3%) and 34 men (44.7%)) and found that the patients aged from 10 to 89 years (median 51; average 55.3); 73 (96.1%) had the PHAT located within the superficial subcutaneous tissues, and 3 (3.9%) in the muscles. The common site involved was the lower extremity (65.8%). Other sites (34.2%) included the arm, chest wall, axillary fossa, popliteal fossa, buttock, inguinal, perineum, and buccal mucosa. Most of the patients presented with slowly growing painless masses. The size of these tumors ranged from 0.3 to 19.7 cm in diameter. 1,4,6-13,15,16
Grossly, all of the PHATs are nonencapsulated, lobulate infiltrative masses, varied from white-tan to maroon in color. 1 Some of the lesions had cystic components. In our study, the tumors of the nine cases were relatively soft in consistency and some of them had hemorrhage that is the reason why the tumor was misdiagnosed as hematomas initially in 4 of our patients.
Histologically, the lesion usually has infiltrative borders. Typical PHAT is characterized by clusters of hyalinized ectatic vessels and hyperchromatic spindle or pleomorphic cells with nuclear pseudoinclusions.1 The ectatic vessels are lined by thin endothelial cells with an underlying rim of amorphous eosinophilic material. The tumor cells are arranged randomly between the angiectatic vessels. The cells vary from small spindle cells with inconspicuous nucleoli to large round or pleomorphic cells with large nucleoli and occasional intranuclear inclusions. Cytoplasmic hemosiderin can be observed in many of the neoplastic cells. Notably, besides the typical PHAT, in our series, we also found atypical components, consisting of bland, hemosiderin-laden spindle cells in a myxoid background, which is similar to the “HFLL” histologically,2 and numerous small single vessels in some areas of the tumor. In the largest series of patients with PHAT, which included 41 cases,4 12 had only HFLL-like areas without distinct features of PHAT. Thus, the authors recommended that HFLL-like lesions should be termed early PHAT, because it represents the precursor of typical PHAT. However, some other pathologists suggested that HFLL-like areas is not the “early PHAT”, but the “late PHAT”.18
In our opinion, HFLL-like lesion in PHAT is a kind of atypical PHAT and is neoplastic rather than reactive, and probably represents a precursor of typical PHAT. 19,20 First, in both the series of us and Folpe and Weiss,4 HFLL-like areas were not only located at the periphery of the lesions, but also intimately admixed with typical PHAT. Second, typical and atypical components share some similarities including nuclear pseudoinclusions, hemosiderin-laden neoplastic cells, and tiny ecstatic vessels. Third, in our study, a recurrent lesion was almost composed of typical PHAT only, while the primary lesion consisted of both typical and atypical PHAT. Fourth, in our two recurrent cases, the initial margin of the tumor was entirely composed of atypical components, suggesting an association between atypical and typical PHATs. Therefore, pathologists should recognize atypical PHAT, particularly in the periphery of the lesions. Clinicians should avoid removing only the central nodule but leaving behind the subtle areas of atypical PHAT.
Immunohistochemically, neoplastic cells are usually strongly positive for vimentin and CD34, and negative for S-100 protein, CD31, desmin, smooth muscle actin, muscle specific actin, cytokeratin, and EMA.1,3,4 Further- more, in our series, the neoplastic cells were also strongly and diffusely positive for CD99 and VEGF, which is consistent with the results of Groisman and Matsumoto.5,6
PHAT might be misdiagnosed as other soft tissue lesions because of its rarity. First of all, malignant fibrous histiocytoma should be considered in the differential diagnosis, because both the tumors share prominent cellular and nuclear pleomorphism. In our study, case 5 was misdiagnosed as having a malignant fibrous histiocytoma initially. We should pay attention to intranuclear cytoplasmic inclusions, low MIB-1 index (<2%), and lacking significant mitotic activity that indicate PHATs (in malignant fibrous histiocytomas mitotic figures are frequent and MIB-1 index are usually much higher).21-28 In addition, immunostaining for CD34 is helpful for differentiation. Second, PHAT should be differentiated from a schwannoma, because both the lesions have ectatic, hyalinized blood vessels and infiltration of variable chronic inflammatory cells.29 In our study, case 8 was misdiagnosed as a schwannoma initially. Unlike a typical schwannoma, PHAT lacks a distinct capsule, grows infiltratively, without well-formed Verocay bodies and clear-cut Antoni A and B areas. Immunohistochemically, it is strongly positivie for CD34, VEGF, and CD99, and negativie for S-100 protein. Third, PHAT also mimics melanotic schwannoma because of intracytoplasmic pigment and intranuclear inclusions. However, the latter expresses S-100 protein, HMB-45, and Melan-A, whereas the former lacks these antigens.30
As for an atypical PHAT, the differential diagnosis involves the following lesions: angiosarcoma (especially Kaposi sarcoma), lipomatous tumor, leiomyosarcoma, nodular fasciitis, and vascular malformation. We should keep in our mind that both lipomatous tumor and nodular fasciitis are usually devoid of ectatic hyalinized vessels and hemosiderin-laden neoplastic cells. Furthermore, ropey collagen is not a feature of PHAT, whereas it is striking in spindle cell lipoma.31 The pronounced histopathologic characteristics of PHAT and the negativity for desmin and α-SMA might be helpful to exclude a leiomyosarcoma. Histologically, Kaposi sarcomas possess some similarities with PHAT, however they usually have strong and intensive HHV-8 positivity.32
Treatments and prognosis
A very low MIB-1 index, lacking significant mitotic activity, and the absence of aneuploidy on flow cytometry suggest that PHAT is not a fully malignant tumor. However, in a report, 33%−50% of the patients with PHAT experienced local recurrences,4 while local recurrences occurred in 2 (28.6%) of 7 patients in our study. Therefore, the lesion should be regarded as a borderline neoplasm.
Because of the small number of cases reported, it is difficult to reach a consensus regarding treatments of the disease. In our opinion, surgical excision with tumor-free margin is a choice for this lesion. However, a wide, local excision of the area is necessary to prevent additional growth of a recurrent tumor. Low-dose radiotherapy could be given to the patients, particularly in whom incomplete resection of the lesion has been performed to avoid recurrence. Metastasis of PHAT has not been reported so far, 1,4-17 so high-dose radiotherapy or chemotherapy after the operation is not necessary.
Some immunohistochemical and ultrastructural studies indicates that the neoplastic cells might be a kind of undifferentiated, primitive mesenchymal cells, possibly related to stromal fibroblasts.1,4 PHAT, which is a fibrohistiocytic tumor, probably derive from proliferating microvascular CD34+ dendritic cells and FXⅢa+ dendrophage cell subsets.10
In summary, PHAT is a unique borderline neoplasm. We believe that atypical PHAT is neoplastic rather than reactive, and it is probably a precursor of typical PHAT. Appropriate long-term follow-up is required for all the patients with PHAT as it may recur after several years. Further studies, including cytogenetic or molecular biology, are necessary to better differentiate PHAT from other lesions.
1. Smith ME, Fisher C, Weiss SW. Pleomorphic hyalinizing angiectatic tumor of soft parts: a low-grade neoplasm resembling neurilenoma. Am J Surg Pathol 1996; 20:21-29.
2. Marshall-Taylor C, Fanburg-Smith JC. Hemosiderotic fibrohistiocytic lipomatous lesion: ten cases of a previously undescribed fatty lesion of the foot/ankle. Mod Pathol 2000; 13: 1192-1199.
3. Weiss SW. Pleomorphic hyalinizing angiectatic tumour of soft parts. In: Christopher DMF, Krishnan UK, Fredrik M, eds. World health organization classification of tumors: pathology and genetics-tumours of soft tissue and bone. Lyon: International Agency for Research on Cancer (IARC), 2002: 191.
4. Folpe AL, Weiss SW. Pleomorphic hyalinizing angiectatic tumor: analysis of 41 cases supporting evolution from a distinctive precursor lesion. Am J Surg Pathol 2004; 28: 1417-1425.
5. Groisman GM, Bejar J, Amar M, Ben-lzhak O. Pleomorphic hyalinizing angiectatic tumor of soft parts: immunohistochemical study including the expression of vascular endothelial growth factor. Arch Pathol Lab Med 2000; 124:423-426.
6. Matsumoto K, Yamamoto T. Pleomorphic hyalinizing angiectatic tumor of soft parts: a case report and literature review. Pathol Int 2002; 52: 664-668
7. Lin O, Crapanzano JP. Fine-needle aspiration cytology of pleomorphic hyalinzed angiectatic tumor: a case report. Diagn Cytopathol 2005; 32: 238-242.
8. Fujiwara M, Yuba Y, Wada A, Ozawa T, Tanaka T. Pleomorphic hyalinizing angiectatic tumor of soft parts: report of a case and review of the literature. J Dermatol 2004; 31: 419-423.
9. Husek K, Vesely K. Pleomorphic hyalinizing angiectatic tumor. Cesk Patol 2001; 37: 177-181.
10. Silverman JS, Dana MM. Pleomorphic hyalinizing angiectatic tumor of soft parts: immunohistochemical case study shows cellular composition by CD34+ fibroblasts and factor XIIIa+ dendrophages. J Cutan Pathol 1997; 24: 377-383.
11. Ide F, Shimoyama T, Horie N. Pleomorphic hyalinizing angiectactic tumor of the buccal mucosa. J Oral Pathol Med 2004; 33: 451-453.
12. Fukunaga M, Ushigome S. Pleomorphic hyalinizing angiectatic tumor of soft parts. Pathol Int 1997; 47: 784-788.
13. Lee JC, Jiang XY, Karpinski RH, Moore ED. Pleomorphic hyalinizing angiectatic tumor of soft parts. Surgery 2005; 137: 119-121.
14. Brim SP, Allerding TJ, Buck K. Pleomorphic hyalinized angiectatic tumor of soft parts. J Am Podiatr Med Assoc 1999; 89: 307-311.
15. El-Tal AE, Mehregan D. Pleomorphic hyalinizing angiectatic tumor of soft parts: case report and literature review. J Cutan Pathol 2006; 33: 361-364.
16. Thomas Temple H, Fanburg-Smith J, Murphey MD. Unusual soft tissue mass in a 43-year-old man. Clin Orthop Relat Res 1998; 345: 249-252, 255-256.
17. Capovilla M, Birembaut P, Cucherousset J, Ploton D, de Saint-Maur PP, Flejou JF, et al. Pleomorphic hyalinizing angiectatic tumor of soft parts: ultrastructural analysis of a case with original features. Ultrastruct Pathol 2006; 30: 59-64.
18. Michal M, Kazakov DV. Relationship between pleomorphic hyalinizing angiectatic tumor and nemosiderotic fibrohistiocytic lipomatous lesion. Am J Surg Pathol 2005; 29: 1256-1257.
19. Suarez-Vilela D, Izquierdo-Garcia FM. Lipoblast-like cells in early pleomorphic hyalinizing angiectatic tumor. Am J Surg Pathol 2005; 29: 1257-1259.
20. Folpe AL, Weiss SW. Lipoblast-like cells in early pleomorphic hyalinizing angiectatic tumor. Author reply. Am J Surg Pathol 2005; 29: 1259.
21. Segal NH, Blachere NE, Shiu HY, Leejee S, Antonescu CR, Lewis JJ, et al. Antigens recognized by autologous antibodies of patients with soft tissue sarcoma. Cancer Immun 2005; 5: 4.
22. An JK, Oh KK. Malignant fibrous histiocytoma of chest wall. Yonsei Med J 2005; 46: 177-180.
23. Son E, Park J, Jeon H, Cho S. Malignant fibrous hstiocytoma (MFH) in axilla. Yonsei Med J 2004; 45: 736-738.
24. Kikuchi A, Akiyama M, Han-Yaku H, Shimizu H, Naka W, Nishikawa T. Solitary cutaneous malignant schwannoma. Immunohistochemical and ultrastructural studies. Am J Dermatopathol 1993; 15:15-19.
25. Mankin HJ, Hornicek FJ. Diagnosis, classification, and management of soft tissue sarcomas. Cancer Control 2005; 12: 5-21.
26. Hibshoosh H, Lattes R. Immunohistochemical and molecular genetic approaches to soft tissue tumor diagnosis: a primer. Semin Oncol 1997; 24: 515-525.
27. Singh HK, Kilpatrick SE, Silverman JF. Fine needle aspiration biopsy of soft tissue sarcomas: utility and diagnostic challenges. Adv Anat Pathol 2004; 11: 24-37.
28. Mentzel T, Fletcher CD. Recent advances in soft tissue tumor diagnosis. Am J Clin Pathol 1998; 110: 660-670.
29. Enzinger FM, Weiss SW. Soft tissue tumors, 4th ed. St. Louis: CV Mosby, 2001: 1472-1475.
30. Zhang HY, Yang GH, Chen HJ, Wei B, Ke Q, Guo H, et al. Clinicopathological, immunohistochemical, and ultrastructural study of 13 cases of melanotic schwannoma. Chin Med J 2005; 118: 1451-1461.
31. Enzinger FM, Weiss SW. Soft tissue tumors, 4th ed. St. Louis: CV Mosby, 2001. 590-597.
32. Lamovec J, Knuutila S. Kaposi sarcoma. In: Christopher DMF, Krishnan UK, Fredrik M, eds. World health organization classification of tumors: pathology and genetics-tumours of soft tissue and bone. Lyon: International Agency for Research on Cancer (IARC), 2002. 170-172.