Abstract
Anaplastic sarcoma of the kidney is a rare tumor (≤25 reported cases) characterized by the presence of cysts, and solid areas composed of bundles of undifferentiated spindle cells, showing marked cellular anaplasia (usually accompanied by TP53 overexpression). These tumors often feature prominent areas of cartilage or chondroid material. Germline mutations in DICER1, encoding the microRNA (miRNA) processor DICER1, cause an eponymous syndrome. Recent reports suggest that anaplastic sarcoma of the kidney should be included in DICER1 syndrome as germline DICER1 mutations are associated with the occurrence of such tumors. Therefore, we sought to determine the following: (1) what proportion of anaplastic sarcoma of the kidney have DICER1 mutations; (2) whether the identified mutations affect both alleles of DICER1 (ie, are biallelic); (3) whether somatic missense mutations in the DICER1 RNase IIIb domain impact miRNA generation; and (4) whether TP53 alteration always occurs in these tumors. DICER1 mutations were evaluated by Sanger sequencing and next-generation sequencing in nine tumor/normal pairs. Impact of DICER1 mutations on miRNA generation was evaluated via an in vitro DICER1 cleavage assay. TP53 status was assessed by immunohistochemistry and next-generation sequencing. Eight of the nine cases had at least one RNase IIIb DICER1 mutation that impacted the generation of miRNAs. There were six tumors with truncating DICER1 mutations and in four of them, the mutation found in the tumor was also detected in adjacent normal tissue, and therefore was likely to be either mosaic or germline in origin. Analysis of mutation phase revealed that two of three tumors had biallelic DICER1 mutations. Six of nine anaplastic sarcomas of the kidney had aberrant TP53 immunohistochemisty with damaging TP53 mutations identified in three cases. Taken together, these data suggest that the great majority of anaplastic sarcomas of the kidney have DICER1 mutations and confirm that these tumors are part of the DICER1 syndrome.
Similar content being viewed by others
Main
Anaplastic sarcoma of the kidney was first described as a novel pediatric renal neoplasm in 2007.1 Previous to this formal description of anaplastic sarcoma of the kidney, some of these lesions could have been included in renal tumors broadly characterized as embryonal sarcomas of the kidney.2, 3 In a series of 25 such tumors,2 3 possessed anaplasia, so perhaps these tumors could be considered to be unrecognized anaplastic sarcomas of the kidney. Others may have been primary renal synovial sarcomas.4 The one almost certain anaplastic sarcoma of the kidney identified before the paper of Vujanic et al in 2007 was published by Faria and Zerbini5 as a dedifferentiated cystic nephroma. In this case, the renal tumor occurred in a 26-month-old girl. It was largely cystic, but had a solid anaplastic region with cartilaginous and rhabdomyoblastic differentiation.5 In view of subsequent findings, it is perhaps surprising that neither germline nor somatic DICER1 mutations were identified in this person or the tumor 18 years later.6
Anaplastic sarcoma of the kidney presents as a large renal mass, and its major gross and histologic features include the presence of cysts, marked anaplasia in the spindle cell component, and areas of benign or malignant cartilage or chondroid differentiation.1 There is female predominance. In 2007, there were no definitive genetic mutations or molecular markers linked to these tumors. These tumors remain amongst the rarest renal neoplasms, with no more than 25 cases reported to date.
Recent studies have reported that a few isolated cases of ASK harbor mutations in DICER1,6, 7, 8, 9 making a case that anaplastic sarcomas of the kidney should be included with the other lesions of the pleiotropic tumor predisposition syndrome known as DICER1 syndrome (OMIM 606241). DICER1 syndrome tumors include pleuropulmonary blastoma,10 cystic nephroma,11 and many other rare tumor entities, mainly occurring in the pediatric and adolescent age range.12 DICER1 is an endoribonuclease central to generating microRNAs (miRNAs), small RNA molecules that downregulate the expression of ~30% of protein-coding genes.12 DICER1 utilizes its RNase IIIa and IIIb endonuclease domains to cleave precursor (pre)-miRNA stemloops, thus releasing the mature single-strand miRNA. Of note, mature miRNAs can be coded within either the 5′ (5p) or 3′ (3p) arms of pre-miRNA stemloops.12 DICER1-related tumors usually possess two DICER1 mutations: one predicted to result in a truncated protein and the other a missense mutation at specific residues within exons encoding the RNase IIIb domain of the DICER1 protein.12 Previous studies of DICER1 syndrome-related tumors have shown that when two DICER1 mutations are present in the tumor, one mutation is present on one of the two DICER1 alleles and the other is present on the alternate allele13, 14, 15, 16, 17, 18, 19, 20 (ie, the mutations are said to be in trans, or are biallelic). In contrast, if both the mutations occur on one allele, then they are in cis, or monoallelic.
Pleuropulmonary blastomas with a germline DICER1 mutation can occur with and without deleterious mutation in TP53.13, 18 Pleuropulmonary blastoma and anaplastic sarcoma of the kidney could be analogous tumors occurring in different organs; it has been suggested that the stages of DICER1-dependent pleuropulmonary blastomas could be reminiscent of the possible progression of DICER1-dependent cystic nephroma to anaplastic sarcoma of the kidney6 but it is unknown whether all anaplastic sarcomas of the kidney arise from pre-existing cystic nephromas,1, 5, 6, 7, 8, 9, 21, 22, 23 partly because of the rarity of anaplastic sarcomas of the kidney but also because the surgical removal of cystic nephromas will prevent the observation of development of anaplastic sarcoma of the kidney from cystic nephroma. Nevertheless, cystic nephromas can pre-exist in regions of the kidney where anaplastic sarcoma of the kidney have later been observed.6, 7, 8 In this report, we sought to determine in nine anaplastic sarcomas of the kidney—(1) the frequency of DICER1 mutations; (2) whether these mutations are biallelic; (3) whether the identified DICER1 mutations occurring in the RNase IIIb domain affect pre-miRNA processing; and (4) does aberrant TP53 status always accompany DICER1 mutations in anaplastic sarcomas of the kidney.
Materials and methods
Sample Acquisition and Histopathological Description of the Anaplastic Sarcomas of the Kidney
Nine anaplastic sarcomas of the kidney and matched normal kidney formalin-fixed paraffin-embedded specimens samples were obtained. Cases 1, 2, and 6 were described in the original description of anaplastic sarcoma of the kidney1 and case 8 was presented by current author Watanabe,23 however neither study investigated DICER1 mutation status. Cases 7 and 9 have been reported by our group (refs 7 and 8, respectively) as DICER1-related anaplastic sarcomas of the kidney. The diagnosis in case 5 was more equivocal than for the other eight cases, and it was considered that it could represent an anaplastic Wilms tumor (the original diagnosis, Table 1).
The study was approved by the Institutional Review Board (IRB) of the Faculty of Medicine of McGill University, Montreal, QC, Canada (numbers A05-M60-14B, A08-M61-09B, and A12-M117-11A). Participants were recruited to the study in compliance with the second edition of the Canadian Tri-Council Policy Statement of Ethical Conduct of Research involving Humans and, where indicated because of young participant age, eligible relatives signed a consent form in accordance with the above-mentioned IRB protocols.
The anaplastic sarcomas of the kidney showed characteristic gross and histological features, including the presence of cysts (in the majority of cases) and solid areas composed of undifferentiated spindle cells with marked anaplastic changes, and prominent areas of benign or malignant cartilage or chondroid differentiation.1 Other, less common features include blastema-like areas, foci of rhabdomyoblastic differentiation, and small islands of osteoid.1
Screening DICER1 Mutations in Anaplastic Sarcoma of the Kidney
DNA was extracted from formalin-fixed paraffin-embedded samples and DICER1 RNase IIIa/b domain Sanger sequencing was performed as previously described.20 The ability of missense mutations to cause exon skipping was assessed as reported by our group.20 Additional fresh frozen tissue was available for three samples. In these cases, full DICER1 Sanger sequencing was performed and the phase of mutations was determined via cloning20 (Supplementary Data). Phase refers to whether two mutations are on the same copy of the gene (in cis) or whether there is one mutation on each chromosomal copy of DICER1 (in trans or biallelic). In addition, DNA extracted from formalin-fixed paraffin-embedded tissues (normal and tumor) was subjected to a custom-designed standard HaloPlex panel containing 11 genes (full gene or targeted regions of DICER1, SMARCA4, SMARCB2, CTNNB1, APC, BRAF, and PTCH1 as well as the exonic regions of DROSHA, FGF3, FGFR1, and TP53) according to a modified version of a previously published protocol.24 This gene panel was designed by our group to use for several projects; however, only DICER1 and TP53 were of interest for the current study. The subsequent deep sequencing was performed at the McGill University and Genome Quebec Innovation Center. Sequence analysis was carried out using a modification of established protocols (Supplementary Materials).
In Vitro Cleavage Assay
HEK 293 cells were transduced to stably express FLAG-tagged versions of the somatically acquired DICER1 RNase IIIb mutations. The ability of the FLAG-immunoprecipitated mutant proteins to cleave internally radiolabeled in vitro-transcribed pre-miR122 was evaluated in a time course by autoradiography of RNA products resolved by denaturing urea polyacrylamide electrophoresis. The details of this assay have previously been described.19
Immunohistochemistry
Immunohistochemistry for cases 1–8 was performed at the Segal Cancer Centre Research Pathology Facility (Jewish General Hospital). Tissue samples were cut at 4 μm, placed on SuperFrost/Plus slides (Fisher), and dried overnight at 37 °C, before immunohistochemical processing. The slides were then loaded onto the Discovery XT Autostainer (Ventana Medical System). All solutions used for automated immunohistochemisty were from Ventana Medical System unless otherwise specified. Slides underwent de-paraffinization and heat-induced epitope retrieval (CC1 pre-diluted solution Ref: 950–124, standard protocol). Immunostaining for TP53 was performed in an automated manner using a heat protocol. Briefly, pre-diluted mouse monoclonal anti-TP53 antibody (Clone Bp53-11, Ventana Medical Systems) was applied for 32 min at 37 °C then followed by the appropriate detection kit (OmniMap anti-Mouse-HRP, Ref: 760–4310) for 8 min, followed by ChromoMap-DAB (Ref: 760–159). A negative control was performed by omission of the primary antibody. Slides were counterstained with hematoxylin. Sections were scanned using the Aperio A Turbo and analyzed by Drs G Vujanic and A Spatz. Case 9 was stained with Cell Marque antibody P53 (DO7) at 1:300 dilution.
Results
Histological characteristics of some of the anaplastic sarcomas of the kidney are described in the legend of Figure 1. Clinicopathological features and follow-up are presented in Table 1. Somatically acquired DICER1 RNase IIIb mutations were identified in eight of the nine anaplastic sarcomas of the kidney studied (Figure 2a). These were determined to be somatically acquired as the mutation is present in the tumor DNA but absent in the matched normal tissue. Seven manifested as missense mutations (cases 2: c.5437G>A [p.E1813K]; 3: c.5113G>A [p.E1705K]; 4 and 7: c.5425G>A [p.G1809R]; 6 and 8: c.5125G>A [p.D1709N]; and 8 (second hit): c.5138A>T [p.D1713V]; Figure 2). In cases 1 and 9 the identified variant c.5438A>G has the potential to cause exon skipping (p.E1788fsX41) in addition to resulting in an altered amino acid (p.E1813G).8 Inactivating DICER1 mutations were observed in cases 1, 6, 7, and 9 (c.5023_5025delTACinsAG [p.Y1675RfsX30]; c.2026C>T [p.R676X]; c.2062C>T [p.R688X]; and c.2450delC [p.P817LfsX15], respectively; Figure 2a). As these inactivating mutations were also observed in matched adjacent normal tissue, we deemed them to be of germline origin but cannot exclude the possibility of a mosaic origin (Figure 2a). Inactivating mutations were also detected in cases 2 (c.4684_4685inC [p.C1562SfsX34]) and 3 (c. 1630C>T [p.R544X]; Figure 2a). The inactivating mutations in these two cases are likely somatically acquired as they were not present in matched normal tissue (Figure 2a; Supplementary Table 2). Cloning experiments to determine the phase of mutations in cases 7, 8, and 9 determined that the pairs of DICER1 mutations in each cases were most likely biallelic for cases 7 and 8, and show a trend toward biallelism in case 9 (see comment in Supplementary Data). The DICER1 mutations in cases 1, 2, and 3 are also presumed to be biallelic based on the DICER1 mutation phase analyses of other DICER1 syndrome lesions13, 14, 15, 16, 17, 18, 19, 20 but we were unable to confirm this presumption due to lack of fresh frozen tissue on which to perform cloning experiments. In summary, 8/9 anaplastic sarcomas of the kidney possessed at least one DICER1 mutation and 7/8 contained two DICER1 mutations (likely in trans, see above). In 3 of these 7 cases, both mutations seen in the tumor were of somatic origin (cases 2, 3, and 8; Figure 2b). In 4 of the 7 cases where at least one DICER1 mutation was present, the truncating mutation was also detected in adjacent normal tissue, and therefore was likely to be either mosaic or germline in nature (cases 1, 6, 7, and 9; Figure 2b; Supplementary Table 2).
The in vitro cleavage data demonstrate that all the DICER1 RNase IIIb mutations, when acting as missense mutations, are incapable of producing 5p miRNAs but instead produce 3p and an incompletely processed RNA molecule we term 5p+loop (Figure 3). If the DICER1 RNAse IIIb mutation of cases 1 and 9 produce a protein lacking exon 25, then no miRNAs are produced.8
The co-occurrence of a TP53 mutation with DICER1 was confirmed in three out of nine cases: cases 1 (c.630G>C [p.R210S]); 4 (c.212G>T [p.R71L]); and 6 (c.521G>A [p.R174Q]). Six of nine anaplastic sarcomas of the kidney (namely, cases 1, 4, 5, 6, 8, and 9) had aberrant TP53 immunohistochemistry (Figure 4a).
Discussion
To our knowledge, this is the largest collection of anaplastic sarcomas of the kidney from which DICER1 sequencing has been attempted. There is a high prevalence of somatically acquired DICER1 RNase IIIb mutations in our collection (8/9; Figure 2), which we suggest is an important genetic feature of the disease. Biallelic DICER1 mutations were observed in two anaplastic sarcomas of the kidney (cases 7 and 8) and a trend toward biallelism was seen in the third (case 9) where this could be assessed. In the case of cases 2, 3, and 8, the DICER1 mutations were both somatically acquired (Figure 2a). Therefore, the two-hit hypothesis of tumor formation is supported by the data for DICER1 mutations.
All the somatically acquired DICER1 RNase IIIb mutations seen in these anaplastic sarcomas of the kidney affect the cleavage of the 5p arm of the pre-miR122 stemloop (Figure 3). Using pre-miR122 as a surrogate for all DICER1-dependent pre-miRNAs, it appears that the tumors with the combination of one inactivating and one RNase IIIb DICER1 mutations are deficient in producing 5p miRNAs and/or have deleterious effects due to the presence of 5p plus loop RNA structures as we and others have suggested in previous studies.7, 8, 19, 25 Alternatively, the presence of only 3p miRNAs could predispose a cell to tumorigenesis. Given the observation that inherited mutations in genes involved in miRNA processing have been observed in patients with Wilms tumor20, 26 and cystic nephroma,27 it has been suggested that availability of a wide range of miRNAs is important for kidney development.26 It seems clear that pediatric cystic nephroma27 and much more rarely, Wilms tumor20, 26, 28, 29 can arise from the kidney of a proband with an inherited DICER1-inactivating mutation, but that the pathway to Wilms tumor does not appear to pass through a pre-existing pediatric cystic nephroma (Figure 5). On the other hand, our recent case report of a microscopic nascent anaplastic sarcoma of the kidney occurring in a pediatric cystic nephroma8 (case 9 here), taken together with previous publications,6, 7 further supports the notion that an anaplastic sarcoma of the kidney can arise within a cystic nephroma (Figure 5).
Overexpressed TP53 as observed by immunohistochemistry as seen in 6/9 of the anaplastic sarcomas of the kidney (Figure 4a) has been used as a surrogate method of TP53 mutation status. In addition, potentially damaging TP53 mutations were identified by HaloPlex in cases 1, 4, and 6 (Figure 4b; Supplementary Table 3). We suggest that mutations in the promoter, 5′-UTR, or 3′-UTR of TP53 not included in the Sanger sequencing and HaloPlex could account for lack of mutations detected in cases 5, 8, and 9. Nevertheless, our reported proportion of anaplastic sarcomas of the kidney with aberrant TP53 expression by immunohistochemistry is similar with that reported for DICER1-dependent pleuropulmonary blastomas.18
Case 5, which showed neither DICER1 nor TP53 mutations, but did overexpress TP53, shared some diagnostic features of anaplastic sarcomas of the kidney (Figures 1g–j), such as widespread anaplastic changes, but lacked some other commonly seen features such as cysts and convincing chondroid differentiation, making this case somewhat atypical and indistinguishable from anaplastic Wilms tumor. In addition, as this was the only tumor that we studied here that did not possess a DICER1 mutation, and as DICER1 mutations are much rarer in Wilms tumors than in anaplastic sarcomas of the kidney, we are inclined, in retrospect, to consider this to be an anaplastic Wilms tumor.
This study suggests that identification of DICER1 RNase IIIb mutations is a useful genetic marker for anaplastic sarcomas of the kidney. Also, DICER1-dependent evolution from cystic nephroma to anaplastic sarcoma of the kidney may parallel the evolution of DICER1-dependent type I cystic pleuropulmonary blastomas to more solid types II and III pleuropulmonary blastomas. As aberrant immunohistochemistry status of TP53 was only seen in a proportion of anaplastic sarcomas of the kidney, it is not a useful molecular marker of the disease. Given the high incidence of DICER1 mutations in our set of anaplastic sarcomas of the kidney we suggest that screening for both germline and somatic DICER1 mutations is warranted in suspected cases of anaplastic sarcoma of the kidney.
References
Vujanic GM, Kelsey A, Perlman EJ et al. Anaplastic sarcoma of the kidney: a clinicopathologic study of 20 cases of a new entity with polyphenotypic features. Am J Surg Pathol 2007;31:1459–1468.
Arnold MM, Beckwith JB, Faria P et al. Embryonal sarcoma of adult and pediatric kidneys. Mod Pathol 1995;8:409.
Delahunt B, Beckwith JB, Eble JN et al. Cystic embryonal sarcoma of kidney: a case report. Cancer 1998;82:427–433.
Argani P, Faria PF, Epstein JI et al. Primary renal synovial sarcoma: molecular and morphologic delineation of an entity previously included among embryonal sarcomas of the kidney. Am J Surg Pathol 2000;24:1087–1096.
Faria PA, Zerbini MC . Dedifferentiated cystic nephroma with malignant mesenchymoma as the dedifferentiated component. Pediatr Pathol Lab Med 1996;16:1003–1011.
Doros LA, Rossi CT, Yang J et al. DICER1 mutations in childhood cystic nephroma and its relationship to DICER1-renal sarcoma. Mod Pathol 2014;27:1267–1280.
Wu MK, Goudie C, Druker H et al. Evolution of renal cysts to anaplastic sarcoma of kidney in a child with DICER1 syndrome. Pediatr Blood Cancer 2016;63:1272–1275.
Wu MK, Cotter MB, Pears J et al. Tumor progression in DICER1-mutated cystic nephroma-witnessing the genesis of anaplastic sarcoma of the kidney. Hum Pathol 2016;53:114–120.
Yoshida M, Hamanoue S, Seki M et al. Metachronous anaplastic sarcoma of the kidney and thyroid follicular carcinoma as manifestations of DICER1 abnormalities. Hum Pathol 2017;61:205–209.
Hill DA, Ivanovich J, Priest JR et al. DICER1 mutations in familial pleuropulmonary blastoma. Science 2009;325:965.
Bahubeshi A, Bal N, Rio Frio T et al. Germline DICER1 mutations and familial cystic nephroma. J Med Genet 2010;47:863–866.
Foulkes WD, Priest JR, Duchaine TF . DICER1: mutations, microRNAs and mechanisms. Nat Rev Cancer 2014;14:662–672.
Pugh TJ, Yu W, Yang J et al. Exome sequencing of pleuropulmonary blastoma reveals frequent biallelic loss of TP53 and two hits in DICER1 resulting in retention of 5p-derived miRNA hairpin loop sequences. Oncogene 2014;33:5295–5302.
Sabbaghian N, Hamel N, Srivastava A et al. Germline DICER1 mutation and associated loss of heterozygosity in a pineoblastoma. J Med Genet 2012;49:417–419.
de Kock L, Sabbaghian N, Plourde F et al. Pituitary blastoma: a pathognomonic feature of germ-line DICER1 mutations. Acta Neuropathol 2014;128:111–122.
de Kock L, Sabbaghian N, Druker H et al. Germ-line and somatic DICER1 mutations in pineoblastoma. Acta Neuropathol 2014;128:583–595.
Heravi-Moussavi A, Anglesio MS, Cheng SW et al. Recurrent somatic DICER1 mutations in nonepithelial ovarian cancers. N Engl J Med 2012;366:234–242.
Seki M, Yoshida K, Shiraishi Y et al. Biallelic DICER1 mutations in sporadic pleuropulmonary blastoma. Cancer Res 2014;74:2742–2749.
Wu MK, de Kock L, Conwell LS et al. Functional characterization of multiple DICER1 mutations in an adolescent. Endocr Relat Cancer 2016;23:L1–L5.
Wu MK, Sabbaghian N, Xu B et al. Biallelic DICER1 mutations occur in Wilms tumours. J Pathol 2013;230:154–164.
Gomi K, Hamanoue S, Tanaka M et al. Anaplastic sarcoma of the kidney with chromosomal abnormality: first report on cytogenetic findings. Hum Pathol 2010;41:1495–1499.
Antonescu C, Bisceglia M, Reuter V et al. Sarcomatous transformation of cystic nephroma in adults. Mod Pathol 1997;10:391.
Watanabe N, Omagari D, Yamada T et al. Anaplastic sarcoma of the kidney: case report and literature review. Pediatr Int 2013;55:e129–e132.
de Kock L, Wang YC, Revil T et al. High-sensitivity sequencing reveals multi-organ somatic mosaicism causing DICER1 syndrome. J Med Genet 2016;53:43–52.
Wang Y, Chen J, Yang W et al. The oncogenic roles of DICER1 RNase IIIb domain mutations in ovarian Sertoli-Leydig cell tumors. Neoplasia 2015;17:650–660.
Rakheja D, Chen KS, Liu Y et al. Somatic mutations in DROSHA and DICER1 impair microRNA biogenesis through distinct mechanisms in Wilms tumours. Nat Commun 2014;2:4802.
Cajaiba MM, Khanna G, Smith EA et al. Pediatric cystic nephromas: distinctive features and frequent DICER1 mutations. Hum Pathol 2016;48:81–87.
Slade I, Bacchelli C, Davies H et al. DICER1 syndrome: clarifying the diagnosis, clinical features and management implications of a pleiotropic tumour predisposition syndrome. J Med Genet 2011;48:273–278.
Palculict TB, Ruteshouser EC, Fan Y et al. Identification of germline DICER1 mutations and loss of heterozygosity in familial Wilms tumour. J Med Genet 2016;53:385–388.
Acknowledgements
We thank the Drs N Benlimame, M Bayat, and D Grehan for performing the immunohistochemistry, and Dr A Spatz for his interpretation of the staining. We thank Drs R Grant and C Goudie for their clinical contribution to this study, and John R Priest for reading the manuscript. WDF is supported by Alex’s Lemonade Stand and a Canadian Institutes for Health Research (CIHR) Grant (FDN-148390); MRF by a CIHR grant (MOP-130425), and MKW by a Fonds de Recherche du Québec-Santé (FRQS) award.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Competing interests
The authors declare no conflict of interest.
Additional information
Supplementary Information accompanies the paper on Modern Pathology website
Rights and permissions
About this article
Cite this article
Wu, M., Vujanic, G., Fahiminiya, S. et al. Anaplastic sarcomas of the kidney are characterized by DICER1 mutations. Mod Pathol 31, 169–178 (2018). https://doi.org/10.1038/modpathol.2017.100
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/modpathol.2017.100
This article is cited by
-
Genomic characterization of DICER1-associated neoplasms uncovers molecular classes
Nature Communications (2023)
-
DICER1 tumor predisposition syndrome: an evolving story initiated with the pleuropulmonary blastoma
Modern Pathology (2022)
-
DICER1-associated sarcomas: towards a unified nomenclature
Modern Pathology (2021)
-
Clinicopathologic and molecular analysis of embryonal rhabdomyosarcoma of the genitourinary tract: evidence for a distinct DICER1-associated subgroup
Modern Pathology (2021)
-
Expanding the spectrum of dicer1-associated sarcomas
Modern Pathology (2020)