Automated Bright-Field Dual-Color In Situ Hybridization for MDM2

Transcription

1 Automated Bright-Field Dual-Color In Situ Hybridization for MDM2 Interobserver Reproducibility and Correlation With Fluorescence In Situ Hybridization in a Series of Soft Tissue Consults Gloria Zhang, MD; Christopher P. Lanigan, MS; John R. Goldblum, MD; Raymond R. Tubbs, DO ; Erinn Downs-Kelly, DO Context. Atypical lipomatous tumors/well-differentiated liposarcomas contain alterations in the 12q13-15 region resulting in amplification of MDM2 and nearby genes. Identifying MDM2 amplification is a useful ancillary test, as the histologic mimics of atypical lipomatous tumors/ well-differentiated liposarcomas have consistently shown a lack of MDM2 amplification. Objective. To assess the interobserver reproducibility of a bright-field assay for MDM2 amplification (dual-color, dual-hapten in situ hybridization [DDISH]) among reviewers with varying degrees of experience with the assay and to assess the concordance of MDM2 DDISH with MDM2 fluorescence in situ hybridization (). Design. In total, 102 cases were assessed in parallel for MDM2 by and DDISH. MDM2 amplification was defined as an MDM2 to chromosome 12 ratio of 2.0 or greater, whereas an MDM2 to chromosome 12 ratio of less than 2 was nonamplified. Fluorescence in situ hybridization was scored in the routine clinical laboratory and DDISH was evaluated by 3 different pathologists blinded to the final diagnosis and results. Results. Fluorescence in situ hybridization categorized 27 cases (26%) as MDM2 amplified and 75 cases (74%) as nonamplified; the consensus DDISH diagnosis was 98% concordant with. Agreement between MDM2 DDISH by each reviewer and MDM2 was highly concordant (99%, 98%, and 98%, respectively, for reviewers 1, 2 and 3). The j agreement of the 3 reviewers scoring DDISH was excellent (j ¼ 0.949, 0.95, and 0.95, respectively, for reviewers 1, 2, and 3). Conclusions. This study highlights excellent concordance between DDISH and in MDM2 copy number assessment. Moreover, excellent interobserver reproducibility of the DDISH assay was found among reviewers with varying levels of experience evaluating bright-field assays. (Arch Pathol Lab Med. 2016;140: ; doi: /arpa OA) Accepted for publication February 17, Deceased. From the Departments of Anatomic Pathology (Drs Zhang and Goldblum) and Molecular Pathology (Mr Lanigan and Dr Tubbs), Robert J. Tomsich Pathology and Laboratory Medicine Institute, Cleveland Clinic, Cleveland, Ohio; and the Department of Pathology, Huntsman Cancer Center, University of Utah, Salt Lake City (Dr Downs-Kelly). The authors have no relevant financial interest in the products or companies described in this article. Portions of this work were presented as a poster at the 2013 United States and Canadian Academy of Pathology Annual Meeting; March 6, 2013; Baltimore, Maryland. Reprints: Erinn Downs-Kelly, DO, Department of Pathology, Huntsman Cancer Hospital, 1950 Circle of Hope, Room N3100, Salt Lake City, UT ( erinn.downs-kelly@path.utah.edu). MDM2 is an oncogene that plays an important role in regulating the cell cycle and tumorigenesis by its inhibitory interaction with TP53. 1 MDM2 gene amplification has been identified in approximately 20% of soft tissue tumors, including undifferentiated pleomorphic sarcoma and osteosarcoma, at variable rates; this alteration has consistently been identified in atypical lipomatous tumor/ well-differentiated liposarcoma (ALT/WDL) and its dedifferentiated counterpart (dedifferentiated liposarcoma). 1 3 Molecular and cytogenetic studies have shown that ALT/ WDLs have ring and giant marker chromosomes that harbor amplified sequences in the 12q13-15 amplicon involving several genes including MDM2, HMGIC, CPM, CDK4, and SAS Importantly, MDM2 amplification has not been identified within the histologic mimics of ALT/ WDL, making this an attractive ancillary testing strategy in this histologically challenging entity. To that end, our group and others have previously shown the diagnostic utility of detecting MDM2 amplification by fluorescence in situ hybridization () within adipocytic neoplasms falling within the histomorphologic differential diagnosis of ALT/WDL. 10,13 16 Although for MDM2 is a reliable assay, there are some drawbacks, including the time, labor, and loss of fluorescent signal over time as well as the technical skill required to perform and interpret. As such, our group has developed and previously published data on a second-generation automated bright-field dualcolor in situ hybridization assay for MDM2. 17 The advantages of this assay include an automated platform with quick turnaround time (roughly 13 hours) as well as highresolution permanent signal. The current study was Arch Pathol Lab Med Vol 140, October 2016 Bright-Field In Situ Hybridization for MDM2 Zhang et al 1111

2 undertaken to assess the interobserver reproducibility of this automated bright-field assay for MDM2 amplification (dualcolor, dual-hapten in situ hybridization [DDISH]) across reviewers with variable experience scoring bright-field assays and to assess the concordance of MDM2 DDISH with MDM2 as used in a series of soft tissue consultative material. MATERIALS AND METHODS Tissues evaluated included 102 mesenchymal lesions seen in a soft tissue consultative practice. MDM2 (Vysis, Abbott Molecular Inc, Des Plaines, Illinois) was performed as an ancillary diagnostic test, as the cases either had histomorphologic features worrisome for ALT/WDL, were deep seated or large (.10 cm) adipocytic neoplasms, or were spindle cell lesions within the retroperitoneum with no definitive ALT/WDL component. Of these, 20 (20%) were core needle biopsies and 82 (80%) represented whole tissue sections from resection material. Fluorescence in situ hybridization was performed according to the manufacturer s instructions and established laboratory protocol in the Cleveland Clinic reference laboratory, with the results incorporated into the final diagnosis. Briefly, unstained 4-lm sections from tissue blocks were placed on electrostatically charged slides (SuperFrost, Fisher, Hampton, New Hampshire) and evaluated individually using dual-color probes (Vysis, Abbott Molecular). A 209-kb probe labeled in SpectrumOrange hybridizes to the MDM2 gene at 12q15 and the centromeric enumeration probe (chromosome 12 [CHR12]) labeled in SpectrumGreen hybridizes to sequences found within the centromere of CHR12 (12p11.1-q11). Cases are scored by enumerating MDM2 signal and CHR12 signal in 40 nonoverlapping nuclei and then calculating the MDM2:CHR12 ratio. In accordance with prior probe validation, a calculated MDM2:CHR12 ratio of 2.0 or greater is considered amplified whereas an MDM2:CHR12 ratio of less than 2 is considered nonamplified. The status of a specimen relies on the calculated ratio with no threshold of positive cells established for defining MDM2 status. Each of these 102 cases was also subjected to MDM2 DDISH. This automated assay is performed on the Ventana BenchMark XT with XT Dual Color Open Probe Software (Ventana Medical Systems, Tucson, Arizona). The MDM2 probe is labeled with dinitrophenyl and hybridizes to an approximately 628-kb region on CHR12 that includes the MDM2 gene; the CHR12 probe is digoxigenin labeled and hybridizes per the manufacturer to CHR12 (Ventana Medical Systems). The probes are sold separately and mixed together into a cocktail via a User Fillable/Prep Kit (Ventana Medical Systems). The cocktail is dispensed as a single drop onto unstained 4-lm sections. Following probe hybridization and visualization, the DDISH cases were independently scored by 3 reviewers with varying levels of experience with bright-field assays. One reviewer had no experience scoring DDISH assays, one reviewer had moderate experience evaluating DDISH assays (prior publications or abstracts using the methodology, which required scoring at least 300 cases), and one reviewer had a high level of experience with the assay including assay development and years of evaluating bright-field in situ hybridization assays. Scoring instructions were discussed among the reviewers and included the following: (1) MDM2 signals were visualized as a discrete black signal whereas CHR12 signals were detected as a discrete red signal; (2) cases were considered acceptable for evaluation if endogenous MDM2 and CHR12 signals were identified in nontumor cells; (3) 40 nonoverlapping tumor nuclei were to be enumerated per case, recording the number of MDM2 signals and CHR12 signals per tumor nucleus; (4) aneusomy of CHR12 was considered if the average CHR12 copy number per cell was greater than 3; and (5) if a subpopulation of apparent tumor cells was amplified, the tumor cells were scored unselectively to determine the overall MDM2:CHR12 ratio. All DDISH cases were scored blinded to the histologic diagnosis and status. The MDM2:CHR12 ratio was calculated for each case, and, in keeping with the prior validation, an MDM2:CHR12 ratio of 2.0 or greater was considered amplified whereas an MDM2:CHR12 ratio of less than 2 was considered nonamplified. The DDISH results were compared with the results; correlation between DDISH and and agreement among observers were assessed with j statistics. RESULTS Both and DDISH were assessed and categorized as either MDM2 amplified or nonamplified in a total of 102 cases(seefigure1,athroughd).thehistopathologyand results were incorporated to arrive at a final diagnosis, with categorizing 27 cases (26%) as MDM2 amplified. For the MDM2-amplified cases, the resulting diagnoses were ALT/WDL (n ¼ 13) and dedifferentiated liposarcoma (n ¼ 14). For the 75 cases considered nonamplified, diagnoses were lipoma (n ¼ 46); intramuscular lipoma (n ¼ 11); pleomorphic lipoma (n ¼ 1); myofibroblastoma (n ¼ 1); sclerosing mesenteritis (n ¼ 2); lipoma with fat necrosis (n ¼ 4); pleomorphic sarcoma, not otherwise specified (n ¼ 2); low-grade spindle cell neoplasm not further classifiable (n ¼ 1); and benign mesenchymal myxoid neoplasm not further classifiable (n ¼ 1). Six cases were sent solely for MDM2 assessment without a request for tissue diagnosis; as such, the final diagnosis is not available on those cases. Although evaluable signal was present in all 102 cases by both and DDISH, 14 (14%) of the DDISH cases had to be repeated (1 time) because of weak red signal. This was found to be an artifactual loss of red signal if alcohol was used to dehydrate slides prior to cover slipping. Overall, there was concordance in 100 of 102 cases (98%) between the consensus DDISH result (agreement of 2 of the 3 reviewers or better) and. With respect to DDISH for each reviewer, agreement between the DDISH result and was highly concordant (99%, 98%, and 98%, respectively for reviewers 1 3). Overall, 3 cases were discordant between DDISH and. The first discrepancy was a whole-tissue slide from a resection of a highgrade spindle cell lesion from the retroperitoneum. On DDISH, reviewers 2 and 3 identified this case as being heterogeneous but overall having net MDM2 amplification, whereas identified this as nonamplified. On rereview of the slide, rare scattered amplified cells were identified (see Figure 2, A through D). With the second discrepancy (also resection material), reviewer 1 considered the DDISH as nonamplified with aneusomy of CHR12, whereas reviewers 2 and 3 identified the material as amplified with aneusomy of CHR12; the corresponding was amplified and the final diagnosis was that of an ALT/WDL. The third discrepancy (needle core biopsy) was with reviewers 2 and 3 identifying the material as aneusomic for CHR12 with low-level MDM2 amplification (MDM2:CHR 12 ratio of 2.2) whereas reviewer 1 considered the material nonamplified. The corresponding was nonamplified and the final diagnosis was a pleomorphic lipoma. With respect to interobserver reproducibility, when the j value exceeds 0.81, interobserver interpretative reproducibility has been inferred from published literature to be almost perfect agreement. 18 Using these guidelines, j agreement of the reviewers was almost perfect (j ¼ 0.949, 0.95, and 0.95, respectively, for reviewers 1 3). The DDISH and results and j values are listed in the Table Arch Pathol Lab Med Vol 140, October 2016 Bright-Field In Situ Hybridization for MDM2 Zhang et al

3 Figure 1. A, MDM2 fluorescence in situ hybridization () showing an overall nonamplified state; most cells have 2 copies of the MDM2 locus (red signal), whereas the chromosome 12 (CHR12) copy number (green signal) varies from 2 to 4 copies per cell. B, MDM2 showing an amplified status with MDM2 copy number ranging from 14 to greater than 20 copies per nucleus (red signals) and 2 copies of CHR12 per nucleus (green signals). C, MDM2 nonamplified dual-color, dual-hapten in situ hybridization (DDISH) with equal numbers of MDM2 copies per nucleus (black signal) and CHR12 (red signal). D, DDISH showing an MDM2-amplified state with more than 20 MDM2 gene copies (black signal) in the majority of neoplastic nuclei and 2 copies of CHR12 per nucleus (red signal) (original magnifications 3100 [A and B] and 340 [C and D]. DISCUSSION The evaluation of MDM2 copy number status has become an important ancillary diagnostic tool in soft tissue pathology, especially given the limited material often present on needle core biopsy. Numerous studies have documented the utility of MDM2 in the challenging histologic differential of adipocytic tumors. 10,13,15,19 22 Other modalities to assess MDM2 include immunohistochemistry and reverse transcriptase polymerase chain reaction, each having limitations. In some studies, immunohistochemistry has not proven to be as sensitive or specific as for the detection of MDM2, 14,21 and reverse transcriptase polymerase chain reaction prohibits the evaluation of tissue morphology. Fluorescence in situ hybridization also is not without limitations, which include a labor-intensive assay, requirement of a fluorescence microscope, and a signal that diminishes over time. Moreover, in evaluating large pieces of tissue composed predominately of adipocytes, the nuclear density is low and care must be taken to evaluate the tissue in its entirety to avoid missing heterogeneity. The advantages of a bright-field assay to detect MDM2 gene status include the assessment of tissue morphology with ease of identification of nuclei. This latter point is critically important, as amplification may be heterogeneous and identifying signals by bright-field is less cumbersome than with fluorescence. Also, DDISH is less labor intensive, as it is fully automated and is completed in an overnight run. There were 2 discrepancies between the consensus DDISH and diagnosis. One of these was a consensus DDISH diagnosis of amplified whereas the corresponding was nonamplified. On rereview of the, scattered amplified cells were identified that were not appreciated in the original count. This discrepancy highlights the need for a thorough review of the entire slide when evaluating these assays and also points out the bright-field advantage in certain situations (ie, heterogeneous populations that are more easily identified on bright-field than in dark room conditions). This discrepancy also serves as a reminder that one cannot implicitly rely on the results of an ancillary test to establish a diagnosis; the histopathology and site of disease continue to be a cornerstone for diagnosis. The second discordance between consensus DDISH and involved a needle core biopsy wherein the consensus DDISH was low-level amplification with aneusomy of CHR12 (MDM2:CHR 12 ratio of 2.2) whereas the corresponding was nonamplified and the final diagnosis was a pleomorphic lipoma. This case suggests that in Arch Pathol Lab Med Vol 140, October 2016 Bright-Field In Situ Hybridization for MDM2 Zhang et al 1113

4 Figure 2. A, Hematoxylin-eosin stained photomicrograph of the discrepant whole-tissue section that was considered MDM2 amplified by consensus dual-color dual-hapten in situ hybridization (DDISH) and nonamplified by the original fluorescence in situ hybridization () assay. B, Corresponding MDM2 with rare scattered amplified cells. Note that 3 of the cells present have an MDM2 copy number of 8 copies per nucleus (red signal) and 2 copies of CHR12 signals (green signal); the remainder of the cells within this field have equal numbers of red and green signals. C, DDISH on the same case showing a field that contains nonamplified cells. D, DDISH on the same case showing a field that contains scattered amplified cells with 8 to 10 MDM2 copies per nucleus (black signal) and 2 or 3 CHR12 copies per nucleus (red signal) whereas many of the cells present are nonamplified (original magnifications 310 [A], 3100 [B], and 340 [C and D]. situations where there is potential low-level amplification, counting additional cells (more than 40) may be beneficial in arriving at a more accurate assessment of the MDM2 status of a lesion. The third overall discrepancy was on resection material, wherein reviewer 1 considered the DDISH as nonamplified with aneusomy of CHR12 whereas reviewers 2 and 3 identified the material as amplified with aneusomy of CHR12 (average MDM2:CHR12 ratio of 2.5); the corresponding was amplified and the final diagnosis was that of an ALT/WDL. Although this was not a consensus diagnosis discrepancy, it points out that there is some subjectivity in these assays. In actual clinical scenarios, cases are shared where the experience of the group can be beneficial. The latter 2 discrepancies were noted within welldifferentiated lipomatous lesions: an ALT/WDL of the extremity and a pleomorphic lipoma of the shoulder region. Zhang et al, 15 in their study of 405 extremity-based lipomatous lesions, which included lipomas, intramuscular lipomas, and ALT/WDLs, evaluated cases diagnosed historically by histomorphology and then assessed their molecular status (MDM2 and CPM amplification status). They found that 11 of 52 cases (21%) that were histologically called ALT/WDLs were reclassified as lipomas or intramuscular lipomas following molecular evaluation. On review, these cases had subtle cytologic atypia. On the converse, they found that 7 of 324 cases (2%) that were histologically classified as lipomas were reclassified as ALT/WDLs j Analysis of Correlation Between Fluorescence In Situ Hybridization () and Dual-Color, Dual-Hapten In Situ Hybridization (DDISH) by Reviewer Reviewer 1 Reviewer 2 Reviewer 3 Non- Non- Non- DDISH DDISH Non s j Abbreviations:, amplified; non-, nonamplified Arch Pathol Lab Med Vol 140, October 2016 Bright-Field In Situ Hybridization for MDM2 Zhang et al

5 following molecular testing. They note that these tumors were not cytologically atypical but tended to be deep and sizable (the majority measured 16.5 cm or larger). The authors suggest that molecular testing for lipomatous tumors be done in the setting of lipomatous tumors with equivocal cytologic dysplasia, recurrent lipomas, and deepseated lipomatous tumors larger than 15 cm without cytologic atypia. In addition, molecular testing of retroperitoneal tumors is warranted, irrespective of the degree of atypia, as suggested by Macarenco et al. 22 The bright-field assessment of MDM2 is akin to the bright-field assessment of HER2 (INFORM HER2 dual ISH, Ventana Medical Systems), which has been adopted in multiple laboratories, including community practices, academic centers, and reference laboratories. Assessment of HER2 gene amplification by this bright-field method has been evaluated in several large series, some of which have been enriched for challenging equivocal cases, with overall good correlation between DDISH and This sets precedence for the use of a bright-field assay for determining MDM2 status. The current study shows that MDM2 gene status as determined by DDISH is highly concordant with MDM2 and has excellent reproducibility among reviewers with varying levels of experience with the assay. References 1. Dal Cin P, Kools P, Sciot R, et al. Cytogenetic and fluorescence in situ hybridization investigation of ring chromosomes characterizing a specific pathologic subgroup of adipose tissue tumors. Cancer Genet Cytogenet. 1993; 68(2): Karakousis CP, Dal Cin P, Turc-Carel C, Limon J, Sandberg AA. Chromosomal changes in soft-tissue sarcomas: a new diagnostic parameter. Arch Surg. 1987;122(11): Pedeutour F, Suijkerbuijk RF, Forus A, et al. Complex composition and coamplification of SAS and MDM2 in ring and giant rod marker chromosomes in well-differentiated liposarcoma. Genes Chromosomes Cancer. 1994;10(2): Hostein I, Pelmus M, Aurias A, et al. Evaluation of MDM2 and CDK4 amplification by real-time PCR on paraffin wax-embedded material: a potential tool for the diagnosis of atypical lipomatous tumours/well-differentiated liposarcomas. J Pathol. 2004;202(1): Dei Tos AP, Doglioni C, Piccinin S, et al. Coordinated expression and amplification of the MDM2, CDK4, and HMGI-C genes in atypical lipomatous tumours. J Pathol. 2000;190(5): Erickson-Johnson MR, Seys AR, Roth CW, et al. Carboxypeptidase M: a biomarker for the discrimination of well-differentiated liposarcoma from lipoma. Mod Pathol. 2009;22(12): Fletcher CD, Akerman M, Dal Cin P, et al. Correlation between clinicopathological features and karyotype in lipomatous tumors: a report of 178 cases from the Chromosomes and Morphology (CH) Collaborative Study Group. Am J Pathol. 1996;148(2): Nilbert M, Rydholm A, Mitelman F, Meltzer PS, Mandahl N. Characterization of the 12q13-15 amplicon in soft tissue tumors. Cancer Genet Cytogenet. 1995;83(1): Shimada S, Ishizawa T, Ishizawa K, et al. The value of MDM2 and CDK4 amplification levels using real-time polymerase chain reaction for the differential diagnosis of liposarcomas and their histologic mimickers. Hum Pathol. 2006; 37(9): Sirvent N, Coindre JM, Maire G, et al. Detection of MDM2-CDK4 amplification by fluorescence in situ hybridization in 200 paraffin-embedded tumor samples: utility in diagnosing adipocytic lesions and comparison with immunohistochemistry and real-time PCR. Am J Surg Pathol. 2007;31(10): Gisselsson D, Hoglund M, Mertens F, et al. The structure and dynamics of ring chromosomes in human neoplastic and non-neoplastic cells. Hum Genet. 1999;104(4): Pedeutour F, Forus A, Coindre JM, et al. Structure of the supernumerary ring and giant rod chromosomes in adipose tissue tumors. Genes Chromosomes Cancer. 1999;24(1): Weaver J, Downs-Kelly E, Goldblum JR, et al. Fluorescence in situ hybridization for MDM2 gene amplification as a diagnostic tool in lipomatous neoplasms. Mod Pathol. 2008;21(8): Weaver J, Rao P, Goldblum JR, et al. Can MDM2 analytical tests performed on core needle biopsy be relied upon to diagnose well-differentiated liposarcoma? Mod Pathol. 2010;23(10): Zhang H, Erickson-Johnson M, Wang X, et al. Molecular testing for lipomatous tumors: critical analysis and test recommendations based on the analysis of 405 extremity-based tumors. Am J Surg Pathol. 2010;34(9): Kashima T, Halai D, Ye H, et al. Sensitivity of MDM2 amplification and unexpected multiple faint alphoid 12 (alpha 12 satellite sequences) signals in atypical lipomatous tumor. Mod Pathol. 2012;25(10): Zhang W, McElhinny A, Nielsen A, et al. Automated brightfield dual-color in situ hybridization for detection of mouse double minute 2 gene amplification in sarcomas. Appl Immunohistochem Mol Morphol. 2011;19(1): Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1): Weaver J, Goldblum JR, Turner S, et al. Detection of MDM2 gene amplification or protein expression distinguishes sclerosing mesenteritis and retroperitoneal fibrosis from inflammatory well-differentiated liposarcoma. Mod Pathol. 2009;22(1): Miura Y, Keira Y, Ogino J, et al. Detection of specific genetic abnormalities by fluorescence in situ hybridization in soft tissue tumors. Pathol Int. 2012;62(1): Kimura H, Dobashi Y, Nojima T, et al. Utility of fluorescence in situ hybridization to detect MDM2 amplification in liposarcomas and their morphological mimics. Int J Clin Exp Pathol. 2013;6(7): Macarenco RS, Erickson-Johnson M, Wang X, et al. Retroperitoneal lipomatous tumors without cytologic atypia: are they lipomas?: a clinicopathologic and molecular study of 19 cases. Am J Surg Pathol. 2009;33(10): Dietel M, Ellis IO, Hofler H, et al. Comparison of automated silver enhanced in situ hybridisation (SISH) and fluorescence ISH () for the validation of HER2 gene status in breast carcinoma according to the guidelines of the American Society of Clinical Oncology and the College of American Pathologists. Virchows Arch. 2007;451(1): Gao FF, Dabbs DJ, Cooper KL, Bhargava R. Bright-field HER2 dual in situ hybridization (DISH) assay vs fluorescence in situ hybridization (): focused study of immunohistochemical 2þ cases. Am J Clin Pathol. 2014;141(1): Koh YW, Lee HJ, Lee JW, Kang J, Gong G. Dual-color silver-enhanced in situ hybridization for assessing HER2 gene amplification in breast cancer. Mod Pathol. 2011;24(6): Lim SJ, Cantillep A, Carpenter PM. Validation and workflow optimization of human epidermal growth factor receptor 2 testing using INFORM HER2 dualcolor in situ hybridization. Hum Pathol. 2013;44(11): Mansfield AS, Sukov WR, Eckel-Passow JE, et al. Comparison of fluorescence in situ hybridization () and dual-ish (DISH) in the determination of HER2 status in breast cancer. Am J Clin Pathol. 2013;139(2): Arch Pathol Lab Med Vol 140, October 2016 Bright-Field In Situ Hybridization for MDM2 Zhang et al 1115