www.elsevier.com/locate/humpath

Human Pathology (2013) 44, 837–843

Original contribution

YWHAE-FAM22 endometrial stromal sarcoma: diagnosis byreverse transcription–polymerase chain reaction informalin-fixed, paraffin-embedded tumor☆,☆☆

Anna Isphording BSc a,1, Rola H. Ali MDb,1, Julie Irving MDb, Angela Goytain BSc a,Nataliya Nelnyk MSc c, Lien N. Hoang MDb, C. Blake Gilks MDa,b,David G. Huntsman MDb,c, Torsten O. Nielsen MD, PhDa,b,Marisa R. Nucci MDc,d,⁎,2, Cheng-Han Lee MD, PhDa,b,⁎,2

aDepartment of Pathology and Laboratory Medicine, and Genetic Pathology Evaluation Center, Vancouver General Hospital,Vancouver, British Columbia, Canada V5Z 4E3bDepartment of Pathology and Laboratory Medicine, University of British Columbia, Vancouver, British Columbia,Canada V6T 2B5cCentre for Translational and Applied Genomics (CTAG), British Columbia Cancer Agency, Vancouver, British Columbia,Canada V5Z 4E6dDepartment of Pathology, Brigham and Women's Hospital, Boston, MA 02115, USA

Received 13 July 2012; revised 8 August 2012; accepted 10 August 2012

C

MG

0h

Keywords:RT-PCR;Endometrial stromalsarcoma;YWHAE;14-3-3;FAM22

Summary A subset of endometrial stromal sarcoma harbors t(10;17)(q23;p13), which results in thegenetic fusion between YWHAE and 1 of 2 highly homologous FAM22 family members—FAM22A orFAM22B. In contrast to classic low-grade endometrial stromal sarcoma with JAZF1-SUZ12 fusions,YWHAE-FAM22 endometrial stromal sarcoma displays high-grade histologic features and is associatedwith more aggressive disease course. Ancillary fluorescence in situ hybridization assay demonstratingthe presence of YWHAE rearrangement can be used to support the diagnosis, but the detection of fusiontranscript would be the most definitive test. We describe here an optimized reverse transcription–polymerase chain reaction assay for detection of YWHAE-FAM22 fusion transcript in formalin-fixedand paraffin-embedded tumor samples. We studied a series of 6 YWHAE-FAM22 endometrial stromalsarcomas, 7 JAZF-SUZ12 endometrial stromal sarcomas, 3 JAZF1-PHF1/EPC1-PHF1 endometrialstromal sarcomas, 6 undifferentiated endometrial sarcomas, 4 uterine leiomyosarcomas, and 4 uterineadenosarcomas. All 6 YWHAE-FAM22 endometrial stromal sarcomas were confirmed by fluorescence

☆ Anna Isphording was supported by funding from the Interdisciplinary Oncology Program at the University of British Columbia, Vancouver, Britisholumbia, Canada.

☆☆ Disclosure: The authors have no conflicts of interest to disclose.⁎ Corresponding authors. M. R. Nucci is to be contacted at Department of Pathology, Harvard Medical School and Brigham andWomen's Hospital, Boston,

A 02115, USA. C.-H. Lee, Department of Pathology and Laboratory Medicine, University of British Columbia and Anatomical Pathology, Vancouvereneral Hospital, Vancouver, BC, Canada V5Z 4E3.E-mail addresses: mnucci@partners.org (M. R. Nucci), chenghan@mail.ubc.ca (C. -H. Lee).1 Both authors contributed equally to the study.2 Co-senior authors.

046-8177/$ – see front matter © 2013 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.humpath.2012.08.007

838 A. Isphording et al.

in situ hybridization assay, whereas all non–YWHAE-FAM22 tumors were confirmed to lack YWHAErearrangement by fluorescence in situ hybridization assay. The reverse transcription–polymerase chainreaction assay optimized for formalin-fixed and paraffin-embedded samples detected YWHAE-FAM22fusion transcripts in all 6 YWHAE-FAM22 endometrial stromal sarcomas and none of the 24 non–YWHAE-FAM22 uterine sarcomas. These findings show that this reverse transcription–polymerasechain reaction assay is sensitive and specific for detection of YWHAE-FAM22 fusion transcript and canserve as a useful adjunct diagnostic assay to confirm the diagnosis of YWHAE-FAM22 endometrialstromal sarcoma in formalin-fixed and paraffin-embedded tumor samples.© 2013 Elsevier Inc. All rights reserved.

fixed, paraffin-embedded (FFPE) tumor sample for diagnos-

1. Introduction

Endometrial stromal sarcoma (ESS) is genetically,biologically, and histologically heterogeneous. Tumorsshowing classic low-grade histology frequently harbortranslocation associated genetic fusions between JAZF1and members of polycomb complex genes (SUZ12, PHF1,and EPC1) [1-7]. In contrast, tumors with high-grade (ormixed low- and high-grade) appearances frequently harborchromosomal translocation t(10;17)(q22;p13), which re-sults in YWHAE-FAM22A or YWHAE-FAM22B geneticfusions (collectively referred to as YWHAE-FAM22 ESS)[8,9]. Clinically, YWHAE-FAM22 ESS is more aggressivethan JAZF1 ESS (encompassing JAZF1/SUZ12/PHF1/EPC–rearranged cases), characterized by frequent andearly recurrence (local and/or systemic). Furthermore, incontrast to JAZF1 ESS, the high-grade component ofYWHAE-FAM22 ESS consistently lacks estrogen receptorand progesterone receptor expression, suggesting thathormonal therapy typically used in the management ofJAZF1 ESS will likely be ineffective against YWHAE-FAM22 ESS [10-12]. The distinction between YWHAE-FAM22 ESS and JAZF1 ESS is therefore clinicallyrelevant as the management would be different.

Although nearly all primary YWHAE-FAM22 ESSpossess high-grade areas characterized by larger nuclei,more irregular nuclear contours and greater mitotic activitycompared with JAZF1 ESS [8,9], it is sometimes difficult tohistologically distinguish between these 2 tumor types asapproximately half of YWHAE-FAM22 ESS contain ahistologically low-grade component, which can mimicJAZF1 ESS or other classic low-grade ESS morphologically.Immunohistochemically, in contrast to the high-gradecomponent of YWHAE-FAM22 that shows diffuse strongnuclear cyclin D1 positivity in the absence of significantestrogen receptor (ER), progesterone receptor (PR), andCD10 immunostaining, the low-grade component typicallydisplays diffuse strong ER, PR, and CD10 immunopositivity,with focal weak cyclin D1 nuclear staining, an immunopro-file that is similar to most JAZF1 ESS [13]. Because of suchmorphologic and immunophenotypic overlap, molecularcharacterization is necessary to determine the tumorgenotypes in some cases. We have previously describedfluorescence in situ hybridization (FISH) assay on formalin-

tic purposes, and we describe here a reverse transcription–polymerase chain reaction (RT-PCR) assay that is optimizedfor the detection of YWHAE-FAM22 in FFPE tumor samplesas a definitive ancillary diagnostic test.

2. Materials and methods

2.1. Study samples

FFPE tumor tissues were obtained from the pathologyarchives at Brigham and Women's Hospital and VancouverGeneral Hospital, and these include ESS cases in whichFISH studies had demonstrated either YWHAE-FAM22rearrangement (n = 6) or JAZF1/SUZ12/PHF1/EPC1rearrangement (n = 10). The histologic features of all tumorswere reviewed by 2 authors (C. H. Lee and M. R. Nucci) andpreviously described [8,9]. For comparison, FFPE tumortissues from 6 undifferentiated endometrial sarcomas, 4uterine leiomyosarcomas, and 4 uterine adenosarcomas, alllacking evidence of YWHAE and JAZF1/SUZ12/PHF1/EPC1 genetic rearrangements by FISH studies, wereincluded. The study was approved by the institutional reviewboards at the respective institutions.

2.2. FISH study

Split-apart FISH assays interrogating for genetic rearran-gements involving YWHAE, JAZF1, SUZ12, PHF1, andEPC1 were performed as previously described [1]. Probeswere prepared using the following BAC clones (CHORI):JAZF1 3V: RP11-597H8 and RP11-78F4, JAZF1 5V: RP11-466B23 and RP11-945M23, SUZ12 3V: RP11-55J8 andRP11-299H3, SUZ12 5V: RP11-398A1 and RP11-112D12,PHF1 3V: RP11-94D23, PHF1 5V: RP11-242N19 andRP11-908F2, EPC1 3V: RP11-983E11 and RP11-108P17,EPC1 5V: RP11-104F5 and RP11-73A23, YWHAE 5V:RP11-100F18 and RP11-60C18, and YWHAE 3V: RP11-61B16 and RP11-629C16. FISH was performed on 4-μm-thick full sections and tissue microarray sections using astandard pretreatment and hybridization protocol. The slideswere reviewed manually with at least 50 tumor nucleievaluated for each case, and a cutoff of more than 30% nuclei

Table Summary of the study cases and RT-PCR results

Sample Tumor type FISH, geneticrearrangement

RT-PCR forYWHAE-FAM22

1 ESS JAZF1/SUZ12 (−)2 ESS JAZF1/SUZ12 (−)3 ESS JAZF1/SUZ12 (−)4 ESS JAZF1/SUZ12 (−)5 ESS JAZF1/SUZ12 (−)6 ESS JAZF1/SUZ12 (−)7 ESS JAZF1/SUZ12 (−)8 ESS YWHAE/FAM22 (+) a

9 ESS YWHAE/FAM22 (+) a

10 ESS YWHAE/FAM22 (+) a

11 ESS YWHAE/FAM22 (+) a

12 ESS YWHAE/FAM22 (+)13 ESS YWHAE/FAM22 (+)14 UES (−) (−)15 UES (−) (−)16 UES (−) (−)17 UES (−) (−)18 UES (−) (−)19 UES (−) (−)20 LMS (−) (−)21 LMS (−) (−)22 LMS (−) (−)23 LMS (−) (−)24 Adenosarcoma (−) (−)25 Adenosarcoma (−) (−)26 Adenosarcoma (−) (−)27 Adenosarcoma (−) (−)28 ESS JAZF1/PHF1 (−)29 ESS EPC1/PHF1 (−)30 ESS JAZF1/PHF1 (−)

Abbreviations: ESS, endometrial stromal sarcoma; UES, undifferentiatedendometrial sarcoma; LMS: leiomyosarcoma (uterine).NOTE. Samples 8 to 12 correspond to case numbers 1 to 5 of the previousseries of YWHAE-FAM22 ESS [8].

a Previously published cases with RT-PCR confirmation on frozentumor samples.

839YWHAE-FAM22 endometrial stromal sarcoma

showing a split signal was considered positive for rearrange-ment of the flanked gene. Normal paired signals are defined asan orange and green signal less than 3 signal diameters apart ora single yellow (overlapping) signal, whereas unpaired signalsare those separated by greater than or equal to 3 signaldiameters. For tissue microarray sections, all cores withcomplete or partial loss of tumor tissue (b50 tumor nuclei) orwith weak signals were excluded from the FISH analysis.

2.3. RT-PCR and sequencing

RNA was extracted from FFPE tumor blocks (60 μmeach) with RecoverAll Total Nucleic Acid Isolation Kit(Ambion, Foster City, CA, USA), following the manufac-turer's protocol. Complementary DNA (cDNA) was gener-ated using SuperScript III Reverse Transcriptase (Invitrogen,Burlington, Ontario, Canada) with 1.5-μg maximum totalRNA, according to the manufacturer's protocol. Polymerasechain reaction (PCR) for YWHAE-FAM22 and GAPDH(internal positive control) was performed with 0.5-μL TaqDNA Polymerase (Invitrogen, 2 U/μL), 2 μL 10x PCRReaction Buffer (Invitrogen), 0.6 μL 50 mM MgCl2, 0.5 μL10 mM dNTP-Mix (Invitrogen), 0.5 μL forward and reverseprimer, 2 μL cDNA, and 13.4 μL H2O per reaction mix. Theforward primers were YWHAE E5+38F: 5V-CTGGA-TACGCTGAGTGAAGAAAG CTA-3V, YWHAE E5+105F:5V-TGACACTATGGACTTCAGACATGCAG-3V. The re-verse primer used was FAM22A/B E2-654R: 5V-CAGAGCCGTGAACACAGACAGG-3V. The primer setfor GAPDH control were 5V-TCGACCAGTCAGCCG-CATCTTCTTT-3V (GAPDH forward primer) and 5V-ACCAAATCCGTTGACTCCGACCTT-3V (GAPDH re-verse primer). The cycling conditions were set as follows: 1cycle at 94°C for 2 minutes, 35 cycles of 94°C for 0.5 minute,57°C for 0.5 minute, and 72°C for 1.5 minute, followed by72°C for 10 minutes. The PCR products were evaluatedalongside 1kb DNA ladder (Invitrogen) on 3% agarose gelsstained with ethidium bromide. For sequencing, the purifiedPCR product of sample 12was used in a reactionwith BigDyeTerminator v3.1 cycle sequencing kit (Applied Biosystems,Carlsbad, CA, USA) and read on a 3130 Genetic Analyzer(Applied Biosystems). In these reactions, the outer forwardprimer was used to ensure highest quality of the reads. Thechromatogram was analyzed with Applied Biosystemssequencing analysis software and Ridom TraceEdit.

3. Results

3.1. Study sample

FFPE tumor blocks were obtained from 6 YWHAE-FAM22ESSs. Five of the 6 were previously reported (Table) [8]. The1 additional case (case no. 13) was a myopermeative uterinetumor that contained a mix of low- and high-grade areas as

illustrated in Fig. 1. The low-grade component showedvarying degree of cellularity and the stroma ranged fromfibrous to myxoid (Fig. 1B-D). The more cellular areademonstrated histologic features that were well accepted forclassic low-grade ESS (Fig. 1D), with small monomorphicovoid nuclei, low mitotic activity (1 mitotic figures [MF]/10high-power field [HPF]), an arborizing network of stromalarterioles, concentric periarteriolar whorls of tumor cells andthe presence of fibrous bands. The high-grade component incomparison showed more irregular nuclear contour and highmitotic activity (up to 35 MF/10 HPF) (Fig. 1E-F). Focalstromal hyalinization was also present in the high-grade area(Fig. 1F). This tumor exhibited the usual immunohistochem-ical profiles with diffuse CD10, ER, and PR staining observedin the low-grade component and absent CD10, ER, and PRstaining in the high-grade component (Fig. 2A-C). The high-grade component showed diffuse strong nuclear cyclin D1

Fig. 1 The histologic appearance of a YWHAE-FAM22 ESS (case no. 13) with low-grade and high-grade components. A, This uterine tumorshows permeative growth through the myometrium (original magnification ×40). The low-grade component can be fibrous (B), myxoid (C), orcellular (D) and is weakly active mitotically (up to 1 MF/10 HPF) (original magnification ×200). E and F, The high-grade component consistsof round cells with more irregular nuclear contour and increased mitotic activity (up 35 MF/10 HPF).

840 A. Isphording et al.

staining, in contrast to the weak and focal staining seen in thelow-grade component (Fig. 2D). A virtual representativehematoxylin and eosin stain slide of this tumor is available forviewing online (www.gpecimage.ubc.ca/aperio/images/ess/case13.html/).

All 6 YWHAE-FAM22 demonstrated YWHAE andFAM22 genetic rearrangements by FISH assay, and 4 of 6with available frozen tumor material were confirmed to

possess YWHAE-FAM22 fusion transcripts by RT-PCRanalysis of frozen tissue samples [8,9]. A total 7 JAZF1-SUZ12 ESSs, 3 JAZF1-PHF1/EPC1-PHF1 ESSs, 6 undif-ferentiated endometrial sarcomas, 4 leiomyosarcomas, and 4adenosarcomas were also included the study, all of whichlack YWHAE rearrangement by FISH. The JAZF1-SUZ12,JAZF1-PHF1, and EPC1-PHF1 ESS were confirmed byFISH assays as described previously [1].

Fig. 2 The low-grade (left) and high-grade (right) components of a YWHAE-FAM22 ESS (case no. 13) display differing patterns ofimmunostaining (original magnification ×200). Hematoxylin and eosin stain (A), CD10 immunostain (56C6, 1:10 dilution; Vector lab,Burlingame, CA, USA) (B), ER immunostain (SP1, 1:60 dilution; Thermo Scientific, Fremont, CA, USA) (C), and cyclin D1immunostain (clone SP4, rabbit monoclonal, 1:50; Thermo Scientific) (D).

841YWHAE-FAM22 endometrial stromal sarcoma

3.2. RT-PCR analysis for YWHAE-FAM22fusion transcript

The FFPE tumor samples ranged from 1993 to 2011 inage (1998 to 2010 for YWHAE-FAM22 ESS cases). Becauseof the fragmented nature of the messenger RNA in FFPEtissue samples, 2 primer sets were designed (PRIMER3software) such that the PCR amplicon length was under 200bases. The expected product size for the YWHAE E5+38Fand FAM22A/B E2-654R primer set (YWHAE-FAM22primer set 1) was 168 bases, and the product size for theYWHAE E5 +105F and FAM22A/B E2-654R primer set(YWHAE-FAM22 primer set 2) was 101 bases, comparedwith the 94-base product for GAPDH control. PCR productscorresponding to the size of the expected YWHAE-FAM22transcripts were detected in all 6 YWHAE-FAM22 ESSsexamined using both primer sets, and these products wereabsent in all non–YWHAE-FAM22 tumors, including all 7JAZF1-SUZ12 ESSs, 3 JAZF1-PHF1/EPC1-PHF1 ESSs, 6undifferentiated endometrial sarcomas, 4 uterine leiomyo-sarcomas, and 4 uterine adenosarcomas (Table; Fig. 3).

Subsequent sequencing of the PCR product from primer set 1confirmed the presence of YWHAE-FAM22 fusion transcripts(Fig. 4). The limited amount of FAM22 sequence includeddid not allow for determination of whether it was FAM22Aor FAM22B variant involved.

4. Discussion

We have validated in the current study an RT-PCR assayoptimized for FFPE tumor samples that is sensitive andspecific for detecting YWHAE-FAM22 fusion transcript inuterine sarcomas. Primer set 1 produces a 168-base productthat can be subjected to sequence confirmation by aconventional sequence analyzer. This provides a definitiveancillary diagnostic method for identifying YWHAE-FAM22 ESS in clinical cases where only FFPE tissue isavailable. In particular, the distinction of YWHAE-FAM22ESS from other ESS with JAZF1/SUZ12/PHF1/EPC1rearrangements is important because of its more aggressiveclinical course.

Fig. 3 Identification of YWHAE-FAM22 fusion transcripts in FFPE tumor samples. Representative agarose gel images showing thedetection of 168-base and 101-base fusion transcripts in 3 FISH-confirmed YWHAE-FAM22 (Y-FAM) ESS cases but not in other uterinesarcoma types. GAPDH transcript is detected in all tumor samples studied. Blank control refers to a reaction without input tumor cDNAsample. Abbreviations: LMS, uterine leiomyosarcoma; AS, uterine adenosarcoma.

842 A. Isphording et al.

Although nearly all primary YWHAE-FAM22 ESSreported to date exhibited areas that were histologicallyhigher grade appearing than JAZF1 ESS, approximately halfof YWHAE-FAM22 cases contains an admixed low-gradecomponent that morphologically and immunophenotypicallyresembles JAZF1 ESS (and other classic low-grade ESSwithout demonstrable genetic rearrangements). The low-grade component consists of mitotically inactive to weaklyactive ovoid to spindle-shaped cells, and the nuclei areuniform in appearance with smooth contours. The stroma canrange from fibrous to myxoid. As shown in a new caseillustrated in this study, the low-grade component can bemorphologically indistinguishable from that of usual classiclow-grade ESS, with features that include an arborizingnetwork of stromal arterioles, concentric periarteriolarwhorls of tumor cells, and the presence of fibrous bands[14]. The presence of such a low-grade and betterdifferentiated area with features that overlap with that ofclassic low-grade ESS is why we consider these tumors torepresent a high-grade genotype of ESS rather than anundifferentiated endometrial sarcoma (UES) [15].

We recently identified cyclin D1 as being a usefulimmunomarker for the high-grade component of YWHAE-

Fig. 4 Representative chromatogram (case no. 12) showing the fusionYWHAE and the start of exon 2 of FAM22A/B.

FAM22 ESS, showing diffuse strong nuclear staining, incontrast to the focal and weak staining that is seen in JAZF1ESS and in most of the other uterine sarcoma/sarcomatousmalignancies [13]. However, cyclin D1 staining in the low-grade component of YWHAE-FAM22 ESS is generally weakand focal and cannot be used reliably to separate the low-grade component of YWHAE-FAM22 ESS from JAZF1 ESS.Molecular tests using RT-PCR for YWHAE-FAM22 fusiontranscripts or FISH for YWHAE-FAM22 genetic rearrange-ments are, therefore, needed to accurately determine thetumor genotype. A potential advantage of an RT-PCR assayis that one can interrogate for a number of genetic fusions inparallel without incurring significantly more technical cost.Ideally, given the prognostic/therapeutic implications and thehistologic overlap, RT-PCR assay interrogating the presenceof various known ESS fusion transcripts (including YWHAE-FAM22, JAZF1-SUZ12, and JAZF1-PHF1) in a multiplexreaction [2,3,5,6,15] should be performed in cases wheremolecular confirmation is sought.

In conclusion, we have established an RT-PCR assay fordetection of YWHAE-FAM22 fusion transcripts in FFPEtumor samples, and this assay is sensitive and specific for theidentification of YWHAE-FAM22 ESS.

transcript across the point of fusion between the end of exon 5 of

843YWHAE-FAM22 endometrial stromal sarcoma

References

[1] Chiang S, Ali R, Melnyk N, et al. Frequency of known generearrangements in endometrial stromal tumors. Am J Surg Pathol2011;35:1364-72.

[2] Hrzenjak A, Moinfar F, Tavassoli FA, et al. JAZF1/JJAZ1 gene fusionin endometrial stromal sarcomas: molecular analysis by reversetranscriptase-polymerase chain reaction optimized for paraffin-em-bedded tissue. J Mol Diagn 2005;7:388-95.

[3] Huang HY, Ladanyi M, Soslow RA. Molecular detection of JAZF1-JJAZ1 gene fusion in endometrial stromal neoplasms with classic andvariant histology: evidence for genetic heterogeneity. Am J SurgPathol 2004;28:224-32.

[4] Koontz JI, Soreng AL, Nucci M, et al. Frequent fusion of the JAZF1and JJAZ1 genes in endometrial stromal tumors. Proc Natl Acad SciUSA 2001;98:6348-53.

[5] Micci F, Panagopoulos I, Bjerkehagen B, Heim S. Consistentrearrangement of chromosomal band 6p21 with generation of fusiongenes JAZF1/PHF1 and EPC1/PHF1 in endometrial stromal sarcoma.Cancer Res 2006;66:107-12.

[6] Nucci MR, Harburger D, Koontz J, Dal Cin P, Sklar J. Molecularanalysis of the JAZF1-JJAZ1 gene fusion by RT-PCR andfluorescence in situ hybridization in endometrial stromal neoplasms.Am J Surg Pathol 2007;31:65-70.

[7] Oliva E, de Leval L, Soslow RA, Herens C. High frequency of JAZF1-JJAZ1 gene fusion in endometrial stromal tumors with smooth muscle

differentiation by interphase FISH detection. Am J Surg Pathol 2007;31:1277-84.

[8] Lee CH, Marino-Enriquez A, Ou W, et al. The clinicopathologicfeatures of YWHAE-FAM22 endometrial stromal sarcomas: ahistologically high-grade and clinically aggressive tumor. Am J SurgPathol 2012;36:641-53.

[9] Lee CH, Ou WB, Marino-Enriquez A, et al. 14-3-3 fusion oncogenesin high-grade endometrial stromal sarcoma. Proc Natl Acad Sci USA2012;109:929-34.

[10] Garavaglia E, Pella F, Montoli S, et al. Treatment of recurrent ormetastatic low-grade endometrial stromal sarcoma: three case reports.Int J Gynecol Cancer 2010;20:1197-200.

[11] Malouf GG, Duclos J, Rey A, et al. Impact of adjuvant treatmentmodalities on the management of patients with stages I-II endometrialstromal sarcoma. Ann Oncol 2010;21:2102-6.

[12] Sjoquist KM, Martyn J, Edmondson RJ, et al. The role of hormonaltherapy in gynecological cancers-current status and future directions.Int J Gynecol Cancer 2011;21:1328-33.

[13] Lee CH, Ali R, Rouzbahman M, et al. Cyclin D1 as a diagnosticimmunomarker for endometrial stromal sarcoma with YWHAE-FAM22 rearrangement. Am J Surg Pathol 2012;36:1562-70.

[14] Chang KL, Crabtree GS, Lim-Tan SK, Kempson RL, HendricksonMR. Primary uterine endometrial stromal neoplasms. A clinicopath-ologic study of 117 cases. Am J Surg Pathol 1990;14:415-38.

[15] Kurihara S, Oda Y, Ohishi Y, et al. Endometrial stromal sarcomas andrelated high-grade sarcomas: immunohistochemical and moleculargenetic study of 31 cases. Am J Surg Pathol 2008;32:1228-38.