Anaplastic Meningioma Versus Meningeal Hemangiopericytoma: Immunohistochemical and Genetic Markers

Transcription

1 Anaplastic Meningioma Versus Meningeal Hemangiopericytoma: Immunohistochemical and Genetic Markers VEENA RAJARAM, MD, DANIEL J. BRAT, MD, PHD, AND ARIE PERRY, MD Anaplastic meningiomas (MIIIs) and meningeal hemangiopericytomas (HPCs) display significant morphologic and immunohistochemical overlap, including occasional cases of otherwise classic HPC with focal epithelial membrane antigen (EMA) positivity. The availability of several new biomarkers prompted us to examine the potential diagnostic roles of ancillary immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH) studies. From the archival university neuropathology and consult files of 1 of the authors (A.P.), 19 meningeal HPCs and 19 MIIIs were retrieved for further study. IHC was performed by using EMA, CAM 5.2, CD99, Bcl-2, claudin-1 and Factor XIIIa (FXIIIa) antibodies. FISH was performed with NF2, 4.1B (DAL-1), chromosome 1p32, and 14q32 probes. HPCs showed strong CD99 (85% of cases), strong bcl-2 (86%), focal EMA (33%), focal claudin-1 (13%), and scattered individual cell FXIIIa (100%) positivity. MIIIs showed strong EMA (89%), strong claudin-1 (54%), weak or focal CD99 (15%), weak or focal bcl-2 (31%), and individual cell FXIIIa (84%) positivity. Focal CAM 5.2 expression was seen in 26% of HPCs and 15% of MIIIs. Deletions were extremely common in MIIIs: 1p (94%), 14q (67%), NF2 (100%), and 4.1B (67%). HPCs showed no 14q or 4.1B deletions, with 1 case each of 1p and From the Department of Pathology, Washington University School of Medicine, St. Louis, MO; and the Department of Pathology, Emory University, Atlanta, GA. Accepted for publication July 29, Address correspondence and reprint requests to Arie Perry, MD, Division of Neuropathology, Box 8118, Washington University School of Medicine, 660 S. Euclid Avenue, St. Louis, MO /$ see front matter 2004 Elsevier Inc. All rights reserved. doi: /j.humpath NF2 deletions (6%). The sensitivities and specificities of the 3 most useful IHC markers (EMA, CD99, bcl-2) were 85%-89% and 67%- 84%, respectively. The sensitivity and specificity of claudin-1 for MIII were 54% and 86%, respectively. The specificity and positive predictive value of combined CD99 and bcl-2 expression for the diagnosis of HPC was 95%. The sensitivities of individual genetic markers were 67%-100%, with specificities of 94%-100%. Our 3 conclusions were as follows: (1) EMA, CD99, bcl-2, and claudin-1 IHC and 1p, 14q, NF2, and 4.1B FISH are particularly useful for distinguishing anaplastic meningiomas from meningeal HPCs. (2) Focal EMA expression does not preclude a diagnosis of HPC. (3) The characteristic FXIIIa staining pattern reported for HPC also is encountered frequently in anaplastic meningiomas and therefore is nonspecific in this diagnostic setting. HUM PATHOL 35: Elsevier Inc. All rights reserved. Key words: meningioma, hemangiopericytoma, FISH, immunohistochemistry, tumor genetics. Abbreviations: MIII, anaplastic meningiomas; HPC, meningeal hemangiopericytomas; EMA, epithelial membrane antigen; FISH, fluorescence in situ hybridization; FXIIIa, Factor XIIIa. Anaplastic meningiomas (MIII) and meningeal hemangiopericytomas (HPC) are both dural-based tumors with characteristic histologic patterns that can be distinguished clearly in classic examples. However, occasional cases display sufficient morphologic overlap such that ancillary tests are necessary to distinguish between them. MIIIs are World Health Organization (WHO) grade III tumors that occasionally metastasize and that have a median survival of less than 2 years, whereas HPCs are considered primary dural sarcomas that metastasize outside the CNS in 25%-60% of cases, typically after 1 or more recurrences. They have a median survival of 5 to 12 years, predominantly influenced by whether they are low grade (WHO grade II) or high grade (WHO grade III). Currently, the most useful immunohistochemical and histochemical stains to distinguish the 2 are the following: (1) epithelial membrane antigen (EMA), 1-4 which is typically positive in meningiomas and usually negative in HPCs; (2) a characteristic individual cell Factor XIIIa staining pattern seen in HPCs and not in benign fibrous meningiomas 5,6 ; and (3) reticulin, which shows a dense network of intercellular deposition in HPCs, with most meningiomas being relatively reticulin poor. Other stains, including S100 protein and Leu 7, have been investigated for their utility in distinguishing these tumor types but were not found to have sufficient sensitivities and specificities for clinical use. 7 Similarly, CD34 was initially thought to be specific for HPCs, 1,8 but studies have shown that in contrast to the strong diffuse staining seen in solitary fibrous tumors, most HPCs are either negative or show patchy CD34 positivity, with 60% of meningiomas being similarly positive. 5,8 Diagnostic confusion sometimes results from focal EMA positivity in otherwise classic HPCs. We therefore decided to reexplore the issue of ancillary studies in these 2 tumor types, adding several recently developed biomarkers. CD99 (O13) 6,9 has been reported to be frequently positive in HPCs, and bcl-2 is strongly positive in the closely related solitary fibrous tumor. 6,10 Claudin-1, an integral structural component of tight junctions, is a useful marker for perineuriomas and has now been reported to be positive in meningiomas by immunohistochemistry Additionally, meningiomas have characteristic genetic deletions of 1p, 14q, NF2, and 4.1B (previously known as DAL-1: differentially expressed in adenocarcinoma of the lung ) that can be detected by fluorescence in situ hybridization (FISH). The latter marker, localizing to chromosome 1413

2 HUMAN PATHOLOGY Volume 35, No. 11 (November 2004) TABLE 1. Antibodies and Methods Antigen Antibody and Manufacturers Dilution Automated Staining Method Antigen Retrieval EMA Dako, clone E29, Carpinteria, CA 1:2000 Envision mouse Citrate HIER ph 6.0 Factor XIIIA Calbiochem, polyclonal, La Jolla, CA 1:400 Envision Rabbit Citrate HIER ph 6.0 CAM5.2 BD, monoclonal, San Diego, CA 1:80 Envision mouse EDTA HIER ph 8.0 CD99 Signet, clone O13, Dedham, MA 1:100 Envision mouse Citrate HIER ph 6.0 bcl-2 Dako, clone 124, Carpinteria, CA 1:200 Envision mouse EDTA HIER ph 8.0 Claudin-1 Zymed, rabbit polyclonal, South San Francisco, CA 1:70 Envision labeled polymer horseradish peroxidase Abbreviation: EDTA, ethylenediaminetetraacetic acid; HIER, heat induced epitope retrieval. Steaming for 15 min 18p11.3, was found to have considerable homology to NF2 and is similarly a member of the protein 4.1 family. In addition to frequent deletions, its role as a potential tumor suppressor gene is supported by the fact that reexpression of 4.1B in 4.1B-deficient meningioma cells reduces cell proliferation. 19 In contrast to meningiomas, HPCs currently do not have any well-characterized or signature genetic alterations. MATERIALS AND METHODS Formalin-fixed paraffin-embedded 5- m-thick tissue sections were cut or retrieved from previously diagnosed cases of HPC and MIII in our files at Washington University. These consisted of both in-house and consultation cases, including 19 cases of HPC and 19 cases of MIII. The histology of these tumors was rereviewed for diagnostic accuracy. Immunohistochemical stains were performed as summarized in Table 1. Immunohistochemical results were stratified according to extent and intensity of staining, respectively, as follows: negative, focal, or diffuse; and negative, weak, or strong. Although this represents a subjective approach, it reflects the one used daily in routine diagnostic pathology. Dual-colored FISH was performed as reported elsewhere, 14 and the probes used are summarized in Table 2. The tissue sections were deparaffinized and subjected to target retrieval by steam cooking in citrate buffer (20 minutes), cooling (20 minutes), and washing (5 minutes), followed by pepsin digestion (4 mg/ml at 37 C for 30 minutes). The slides were washed in 2 SSC for 5 minutes and air dried. The probes were diluted 1:25 in DenHyb hybridization buffer (Insitus Laboratories, Albuquerque, NM) and were paired (1p/14q, NF2/4.1B). Ten microliters of the hybridization mix was applied to each slide, and the probe and the target DNA were simultaneously denatured at 90 C for 13 minutes. Overnight hybridization was performed at 37 C in a humidified chamber. The slides were then washed at room temperature in 50% formamide 1 SSC for 5 minutes, followed by a wash TABLE 2. DNA Probes and Sources Probe Clone or Product Source 1p32 (green) RP11-260I23 Genome Sequencing Center, Washington University, St. Louis, MO 14q32 (red) RP11-299G2 Research Genetics, CIT978 SK-A, Huntsville, AL NF2 (red) E2 and E15 cosmids Mia MacCollin, MGH, Boston, MA 4.1B (green) PI-8297 Irene Newsham, Henry Ford Hospital, Detroit, MI in 2 SSC for 5 minutes. The slides were air dried and counterstained with 6-diamino-2-phenylindole (DAPI; 0.5 L/mL; Insitus Laboratories). Sections were viewed under an Olympus BX60 fluorescent microscope with appropriate filters (Olympus, Melville, NY), and those showing 90% nuclei with signals were evaluated, with 100 to 200 intact nonoverlapping nuclei scored for the number of fluorescent signals. Cutoffs for deletions were based on counts from nonneoplastic control specimens (temporal lobectomy specimens for seizure control) for each probe. Interpretation of deletion required 50% of nuclei containing a single signal for the probe. Hybridizations were digitally photographed with a highresolution black and white charged couple device (CCD) camera (Cohu Inc., San Diego, CA), with a Z-stack motor programmed to capture images at sequential DAPI (1 level), FITC (5 levels), and rhodamine (5 levels) filter settings. Reconstruction into a single superimposed image with blue, green, and red pseudocolors was accomplished by software from the CytoVision basic workstation (Applied Imaging, Santa Clara, CA). RESULTS Microscopically, the MIIIs fulfilled WHO criteria for anaplasia, 20 including a mitotic count of 20 per 10 high-power fields or pleomorphic spindled or epithelioid malignant cells whose differential diagnosis included carcinoma, sarcoma, or melanoma (Figs1A and 2A). The HPCs contained monomorphic oval to spindled cells arranged as dense hypercellular sheets, often interrupted by pale hypocellular islands, as well as abundant thin-walled, gaping vessels, some with the characteristic arborizing ( staghorn ) pattern (Figs 1B and 3A). The immunohistochemical and FISH results are summarized in Tables 3 and 4. Almost all of the MIIIs showed diffuse or focal membranous positivity for EMA (16/18), whereas less than one third of the cases showed focal or weak staining with CD99 (3/19) or bcl-2 (6/19; Figs 2B, D, and E). Most of the HPCs showed diffuse or focal, strong membranous positivity for CD99 (18/19) and cytoplasmic positivity for bcl-2 (16/19), and most were EMA negative, though a few cases showed focal, usually weak positivity (7/19) (Fig 3B, D, and E). The characteristic pattern of scattered single cells with delicate Factor XIIIa (FXIIIa) positive processes was seen in all the HPCs (19/19) and in the majority of the MIIIs (16/19; Figs 2C and 3C). Diffuse or focal membranous positivity was seen for claudin

3 FIGURE 1. Representative hematoxylin and eosin staining. (A) Anaplastic meningioma with epithelioid cells, intranuclear clearing, moderate amount of cytoplasm, and mitoses. (B) Hemangiopericytoma with identical monomorphic cells and a single dilated staghorn vessel (original magnification, 400). FIGURE 2. Representative anaplastic meningiomas. (A) Hematoxylin and eosin, epithelioid cells with clear cytoplasm and mitoses imparting a carcinoma-like appearance. (B) Epithelial membrane antigen. (C) Factor XIIIa. (D) CD99. (E) Bcl-2. (F) Claudin-1. (Original magnification, 200). 1415

4 HUMAN PATHOLOGY Volume 35, No. 11 (November 2004) FIGURE 3. Representative hemangiopericytomas. (A) Hematoxylin and eosin. Characteristic staghorn vasculature and oval to spindled cells with high N/C ratios. (B) Epithelial membrane antigen. (C) Factor XIIIa. (D) CD99. (E) Bcl-2. (F) Claudin-1 (Original magnification, 200). in slightly more than half the MIIIs (7/13), whereas only 2 of 15 HPCs showed focal or weak positivity (Figs 2F and 3F). Cytoplasmic staining with CAM5.2 was focal and generally weak to equivocal in rare MIIIs (2/13) and HPCs (5/17). By FISH (Fig 4), all the MIIIs had NF2 deletions (15/15), and the majority had 1p32 (17/18), 14q32 (12/18), and 4.1B (10/15) deletions. In the HPCs, only 1 of 15 cases each had deletion for 1p32 or NF2. There were no deletions of 14q32 or 4.1B. The sensitivity for EMA as a marker for MIII was TABLE 3. Immunohistochemistry Results Variable HPC (n) MIII (n) EMA Weak focal 6 2 Weak diffuse 0 0 Strong focal 1 10 Strong diffuse 0 4 CD Weak focal 1 2 Weak diffuse 0 0 Strong focal 5 0 Strong diffuse 12 1 bcl Weak focal 6 4 Weak diffuse 0 1 Strong focal 2 1 Strong diffuse 8 0 Claudin Weak focal 1 2 Weak diffuse 0 0 Strong focal 1 1 Strong diffuse 0 4 FXIIIa Positive CAM Positive %, and the specificity was 67%. FXIIIa was not found to be useful, given positivity in 84% in MIIIs and 100% of the HPCs, with an identical staining pattern in both the tumors. CD99 (O13) was a very sensitive marker of HPC (95%), with a specificity of 84%. Similarly, bcl-2 was a sensitive marker of HPC (84%), with a specificity of 68%. When used in combination, positivity of both of the latter 2 markers yielded a sensitivity of 84%, specificity of 95%, and a positive predictive value of 93% for the diagnosis of HPC. Claudin-1 was not a highly sensitive marker of MIII (54%) but had a specificity of 86%. FISH studies for 1p32, NF2, DAL-1, and 14q32 were highly associated with the diagnosis of MIII, yielding sensitivities of 67%-100% and specificities of 94%- 100% (Table 4). The positive predictive value of having a single deletion was 94%. DISCUSSION Meningeal HPC was initially described as the angioblastic variant of meningiomas, but as the differences in its clinicopathologic, immunohistochemical, 3,21 and ultrastructural features 22 from meningiomas became apparent, it has come generally to be considered a dural-based sarcoma, analogous to its soft tissue counterpart. 23 EMA, FXIIIa, and reticulin have been touted as the most useful ancillary stains for the distinction of HPC from MIII, with S100 protein and Leu-7 generally considered unreliable because of low sensitivities and specificities. In our study, EMA remained a highly sensitive meningothelial marker but was also reactive in more than a third of the HPCs, albeit weak and focal in most. EMA expression has similarly been reported in rare HPCs elsewhere, 1,5,21 emphasizing that focal EMA ex- 1416

5 MENINGIOMA VS. HEMANGIOPERICYTOMA (Rajaram et al) pression does not entirely exclude the diagnosis of HPC. Surprisingly, the characteristic FXIIIa staining pattern, which is useful in differentiating HPCs from fibrous meningiomas, 5 was not very useful in the differential with MIII, because most of them had an identical pattern of FXIIIa staining. In fact, given the individual cell pattern of reactivity and the cytology of positive cells, it has been suggested that they represent entrapped or recruited dendritic cells, rather than the actual tumor cells. CAM5.2 was focally positive in both MIIIs and HPCs, but only in rare cases. Bcl-2, which has been very useful in the diagnosis of SFT, was also a sensitive marker for HPC, with only occasional MIIIs showing weak or focal staining. CD99 is another marker that has been reported to be positive in SFT, and in our study, it similarly proved to be a sensitive marker, further highlighting the extensive overlap between SFT and HPC. In fact, the majority of HPCs strongly coexpressed CD99 and bcl-2, with this pattern being highly specific because it was generally not encountered in the MIIIs. FIGURE 4. Representative fluorescence in situ hybridizations. (A) Metaphase III (MIII) with 1p and 14q deletions (1 green and 1 red signal in most nuclei). (B) Meningeal hemangiopericytomas (HPC) showing normal number of 1p/14q signals (up to 2 green and 2 red in most nuclei). (C) MIII with NF2/4.1B deletions (1 green and 1 red signal in most nuclei). (D) HPC showing normal number of NF2/4.1B signals (up to 2 green and 2 red in most nuclei). Some signals are beyond the planes of focus in the photographs and there is also some truncation artifact, resulting in fewer signals than expected, because of the sectioning of nuclei (original magnification, 1000). TABLE 4. Variable Fluorescence In Situ Hybridization Results 1p32 14q32 NF2 4.1B HPC 1/15 (7) 0/15 1/15 (7) 0/15 MIII 17/18 (94) 12/18 (67) 15/15 (100) 10/15 (67) Abbreviations: HPC, meningeal hemangiopericytomas; MIIIs, anaplastic meningiomas. Claudin-1, a component of intercellular tight junctions, has been shown to be positive in meningiomas. In our study, only slightly more than half of the MIIIs showed claudin-1 expression. Nonetheless, only a few of the HPCs were positive, making claudin-1 a fairly specific, albeit not very sensitive, marker for MIII. In most cases, MIII can be differentiated from HPC by IHC or by the ultrastructural findings of interdigitating cell processes and desmosomal intercellular junctions in MIII and basal lamina-like amorphous material in HPC. On rare occasions, the IHC staining pattern may be equivocal and electron microscopy noncontributory, necessitating additional molecular studies. Loss of 1p, NF2, 4.1B, and 14q have been frequently identified in meningiomas, whereas no consistent chromosomal losses or gains have been seen in HPCs. Homozygous deletions of the CDKN2A on 9p21 have been reported in both HPCs and MIIIs Alterations on 6p, 7p, and 19q and t(12;19)(q13;q13.3) have been reported in occasional HPCs. 27,28 In our study, deletions for 1p32, 14q32, NF2, and 4.1B by FISH were extremely common in MIIIs and were exceptionally rare in HPCs, supporting the notion that these tumors are genetically and biologically distinct. Thus, it corroborates the current view that HPC is a dural-based sarcoma, rather than a variant of meningioma. The sensitivities and specificities of these markers were also sufficiently high that they provide firm evidence of meningothelial origin when 2 or more deletions are detected and virtually exclude the possibility of MIII when all 4 are absent. Therefore, they are particularly useful as an ancillary diagnostic aid in the most poorly differentiated dural-based neoplasms. In conclusion, EMA, CD99, and bcl-2 by IHC and 1p32, 14q32, NF2, and 4.1B deletions by FISH are useful ancillary markers in distinguishing MIIIs from HPCs. Focal EMA positivity may occasionally be seen in HPCs and therefore does not exclude the diagnosis. FXIIIa does not appear to be a useful marker in this differential, because it is seen in the majority of both HPCs and MIIIs. Additional studies are needed to identify specific molecular markers for meningeal HPC and to clarify its relationship with solitary fibrous tumor. Acknowledgment. The authors thank all the referring pathologists who have contributed their cases in consultation; Ruma Banerjee for assistance with FISH; Rodney Brown, Alma Johnson, Kevin Keith, Kevin Selles, and Lisa Snipes for technical support and immunohistochemistry performance; and Diane Lawson for claudin-1 immunohistochemistry. 1417

6 HUMAN PATHOLOGY Volume 35, No. 11 (November 2004) REFERENCES 1. Renshaw AA, Paulus W, Joseph JT: CD34 and epithelial membrane antigen distinguish dural hemangiopericytoma and meningioma. Appl Neurochem 3: , Nemes Z: Differentiation markers in hemangiopericytoma. Cancer 69: , Moss TH: Immunohistochemical characteristics of haemangiopericytic meningiomas: Comparison with typical meningiomas, haemangioblastomas and haemangiopericytomas from extracranial sites. Neuropathol Appl Neurobiol 13: , Iwaki T, Fukui M, Takeshita I, et al: Hemangiopericytoma of the meninges: A clinicopathologic and immunohistochemical study. Clin Neuropathol 7:93-99, Perry A, Scheithauer BW, Nascimento AG: The immunophenotypic spectrum of meningeal hemangiopericytoma: A comparison with fibrous meningioma and solitary fibrous tumor of meninges. Am J Surg Pathol 21: , Alawi F, Stratton D, Freedman PD: Solitary fibrous tumor of the oral soft tissues: A clinicopathologic and immunohistochemical study of 16 cases. Am J Surg Pathol 25: , D Amore ES, Manivel JC, Sung JH: Soft-tissue and meningeal hemangiopericytomas: An immunohistochemical and ultrastructural study. HUM PATHOL 21: , Cummings TJ, Burchette JL, McLendon RE: CD34 and dural fibroblasts: The relationship to solitary fibrous tumor and meningioma. Acta Neuropathol (Berl) 102: , Guillou L, Gebhard S, Coindre JM: Lipomatous hemangiopericytoma: A fat-containing variant of solitary fibrous tumor? Clinicopathologic, immunohistochemical, and ultrastructural analysis of a series in favor of a unifying concept. HUM PATHOL 31: , Tihan T, Viglione M, Rosenblum MK, et al: Solitary fibrous tumors in the central nervous system. A clinicopathologic review of 18 cases and comparison to meningeal hemangiopericytomas. Arch Pathol Lab Med 127: , Bhattacharjee M, Adesina AM, Goodman C, et al: Claudin-1 expression in meningiomas and schwannomas: Possible role in differential diagnosis. J Neuropathol Exp Neurol 62:581, Folpe AL, Billings SD, McKenney JK, et al: Expression of claudin-1, a recently described tight junction-associated protein, distinguishes soft tissue perineurioma from potential mimics. Am J Surg Pathol 26: , Wolburg H, Wolburg-Buchholz K, Liebner S, et al: Claudin-1, claudin-2 and claudin-11 are present in tight junctions of choroid plexus epithelium of the mouse. Neurosci Lett 307:77-80, Cai DX, Banerjee R, Scheithauer BW, et al: Chromosome 1p and 14q FISH analysis in clinicopathologic subsets of meningioma: Diagnostic and prognostic implications. J Neuropathol Exp Neurol 60: , Gutmann DH, Donahoe J, Perry A, et al: Loss of DAL-1, a protein 4.1-related tumor suppressor, is an important early event in the pathogenesis of meningiomas. Hum Mol Genet 9: , Perry A, Cai DX, Scheithauer BW, et al: Merlin, DAL-1, and progesterone receptor expression in clinicopathologic subsets of meningioma: a correlative immunohistochemical study of 175 cases. J Neuropathol Exp Neurol 59: , Watson MA, Gutmann DH, Peterson K, et al: Molecular characterization of human meningiomas by gene expression profiling using high-density oligonucleotide microarrays. Am J Pathol 161: , Weber RG, Bostrom J, Wolter M, et al: Analysis of genomic alterations in benign, atypical, and anaplastic meningiomas: Toward a genetic model of meningioma progression. Proc Natl Acad Sci U S A 94: , Gutmann DH, Hirbe AC, Huang ZY, et al: The protein 4.1 tumor suppressor, DAL-1, impairs cell motility, but regulates proliferation in a cell-type-specific fashion. Neurobiol Dis 8: , Louis DN, Scheithauer BW, Budka H, et al: Meningiomas. In: Kleihues P CW, editor. World Health Organization classification of tumours. Pathology and genetics of tumours of the nervous system. Lyon, France: IARC Press, 2000: Winek RR, Scheithauer BW, Wick MR: Meningioma, meningeal hemangiopericytoma (angioblastic meningioma), peripheral hemangiopericytoma, and acoustic schwannoma. A comparative immunohistochemical study. Am J Surg Pathol 13: , Pena CE: Meningioma and intracranial hemangiopericytoma. A comparative electron microscopic study. Acta Neuropathol (Berl) 39:69-74, Jaaskelainen J, Louis DN, Paulus W, et al: Hemangiopericytoma. In: Kleihues P CW, editor. World Health Organization classification of tumors. Pathology and Genetics of tumors of the Nervous System. Lyon: IARC Press, 2000: Ono Y, Ueki K, Joseph JT, et al: Homozygous deletions of the CDKN2/p16 gene in dural hemangiopericytomas. Acta Neuropathol (Berl) 91: , Perry A, Banerjee R, Lohse CM, et al: A role for chromosome 9p21 deletions in the malignant progression of meningiomas and the prognosis of anaplastic meningiomas. Brain Pathol 12: , Tse JY, Ng HK, Lo KW, et al: Analysis of cell cycle regulators: p16ink4a, prb, and CDK4 in low- and high-grade meningiomas. HUM PATHOL 29: , Henn W, Wullich B, Thonnes M, et al: Recurrent t(12; 19)(q13;q13.3) in intracranial and extracranial hemangiopericytoma. Cancer Genet Cytogenet 71: , Herath SE, Stalboerger PG, Dahl RJ, et al: Cytogenetic studies of four hemangiopericytomas. Cancer Genet Cytogenet 72: ,