FISH Diagnostic Assessment of MDM2 Amplification in Liposarcoma: Potential Pitfalls and Troubleshooting Recommendations

MDM2 amplification represents the leading oncogenic pathway and diagnostic hallmark of liposarcoma, whose assessment is based on Fluorescence In Situ Hybridization (FISH) analysis. Despite its diagnostic relevance, no univocal interpretation criteria regarding FISH assessments of MDM2 amplification have been established so far, leading to several different approaches and potential diagnostic misinterpretations. This study aims to address the most common issues and proposes troubleshooting guidelines for MDM2 amplification assessments by FISH. We retrospectively retrieved 51 liposarcomas, 25 Lipomas, 5 Spindle Cell Lipoma/Pleomorphic Lipomas, and 2 Atypical Spindle Cell Lipomatous Tumors and the corresponding MDM2 FISH analysis. We observed MDM2 amplification in liposarcomas cases only (43 out of 51 cases) and identified three MDM2-amplified patterns (scattered (50% of cases), clustered (14% of cases), and mixed (36% of cases)) and two nonamplified patterns (low number of signals (82% of cases) and polysomic (18% of cases)). Based on these data and published evidence in the literature, we propose a set of criteria to guide MDM2 amplification analysis in liposarcoma. Kindled by the compelling importance of MDM2 assessments to improve diagnostic and therapeutic liposarcoma management, these suggestions could represent the first step to develop a univocal interpretation model and consensus guidelines.


Introduction
Within the neoplastic soft tissue panorama, liposarcoma (LPS) represents the most common type of adult sarcoma, accounting for almost 20% of cases worldwide [1,2]. LPS is a clinical challenge, as it presents a high recurrence rate, unsatisfactory response to available treatments, and a challenging diagnostic workup, especially if based on morphology and immunohistochemistry alone [3].
The new WHO Classification of Soft Tissue Tumors edition identifies several specific subtypes of LPS, namely Atypical Lipomatous Tumors (ALT)/Well-Differentiated LPS (WDLPS), Myxoid LPS (MLPS), Dedifferentiated LPS (DDLPS), and Pleomorphic LPS (PLPS) [4][5][6]. The most common variants are WDLPS and MLPS, while DDLPS represents the aggressive evolution of WDLPS [7]. From a pathologist perspective, WDLPS and DDLPS represent the most challenging variants, as WDLPS can present overlapping features with benign tumors such as Lipomas, while DDLPS may show extensive dedifferentiated MDM2 acts as an E3 ubiquitin-ligase, binding p53 and promoting its ubiquitylation and consequent proteasome-dependent degradation [12] (Figure 1). Low levels of p53 activity are unable to (I) regulate the cell cycle and (II) induce apoptosis in DNA-damaged cells, leading to uncontrolled proliferation. On the other hand, p53 regulates the expression of MDM2, balancing its activity and influence on the cell cycle and proliferation. MDM2 also regulates the Retinoblastoma protein (RB) [13]. In human tumor cell lines, MDM2 enhances RB degradation through a proteasome-dependent mechanism in a similar process as seen for p53 [12].
MDM2 amplification emerged as an oncogenic pathway in several malignances, but it is mostly represented in soft tissue sarcomas [14][15][16][17] where, differently from other tumors, MDM2 amplification and p53 mutation are mutually exclusive [18][19][20][21]. In soft tissue sarcomas, the MDM2 amplification occurs mainly through the so-called double minutes chromosomes (Dmins) mechanism [22]. Dmins are small, generally acentric, and autonomously replicating chromatin bodies that act as an amplification mechanism for several oncogenes, including MDM2 [23][24][25][26]. The sarcomas showing the highest percentages of MDM2 amplification are the low-grade/periosteal osteosarcoma, ALT/WDLPS, DDLPS, MDM2 acts as an E3 ubiquitin-ligase, binding p53 and promoting its ubiquitylation and consequent proteasome-dependent degradation [12] (Figure 1). Low levels of p53 activity are unable to (I) regulate the cell cycle and (II) induce apoptosis in DNA-damaged cells, leading to uncontrolled proliferation. On the other hand, p53 regulates the expression of MDM2, balancing its activity and influence on the cell cycle and proliferation. MDM2 also regulates the Retinoblastoma protein (RB) [13]. In human tumor cell lines, MDM2 enhances RB degradation through a proteasome-dependent mechanism in a similar process as seen for p53 [12].
MDM2 amplification emerged as an oncogenic pathway in several malignances, but it is mostly represented in soft tissue sarcomas [14][15][16][17] where, differently from other tumors, MDM2 amplification and p53 mutation are mutually exclusive [18][19][20][21]. In soft tissue sarcomas, the MDM2 amplification occurs mainly through the so-called double minutes chromosomes (Dmins) mechanism [22]. Dmins are small, generally acentric, and autonomously replicating chromatin bodies that act as an amplification mechanism for several oncogenes, including MDM2 [23][24][25][26]. The sarcomas showing the highest percentages of MDM2 amplification are the low-grade/periosteal osteosarcoma, ALT/WDLPS, DDLPS, and intimal sarcoma [27]. Indeed, the evaluation of the MDM2 status is fundamental in the LPS diagnostics workup, as its amplification is present in 95% of WDLPS and DDLPS cases, while benign lipomatous lesions show no amplification at all [9, 10,12]. IHC for MDM2 expression is available and frequently (>95% of cases) positive in WDLS. However, previous reports found an unsatisfactory correlation between IHC for the MDM2 protein and MDM2 gene amplification status, particularly in poorly differentiated cases or in cases with MDM2 overexpression not related to gene amplification [8,11,[27][28][29]. Considering the clinical consequences of misdiagnosing these lesions, the molecular analysis of MDM2 is essential, and it is usually performed with Fluorescence In Situ Hybridization (FISH) [30]. Indeed, the MDM2 amplification evaluation by FISH is a crucial and well-established assay for liposarcoma diagnostic work-up and, nowadays, is considered as a diagnostic gold standard [31][32][33]. However, despite its diagnostic relevance, no definitive consensus has ever been defined for determining the MDM2 gene status. Considering the frequency of LPS in the adult population and the diagnostic and prognostic implications, the establishment of interpretation guidelines is an unmet need of crucial importance.
In this setting, our study aims to improve the diagnostic interpretation of the FISH assessment of MDM2 amplification in LPS by evaluating the potential drawbacks and pitfalls and suggesting a potential set of diagnostic criteria to achieve a standardized evaluation of this diagnostic test.

MDM2 Amplification Accurately Stratify Our Series
We collected and analyzed 27

MDM2 Amplification Patterns
MDM2 amplified cases presented a high number of MDM2 copies (mean 21; range 11-30.2), a low number of CEP12 (mean 2.5; range 1.4-3.7), and, consequently, an MDM2/CEP12 ratio > 2. MDM2 positive cases presented three distinctive FISH amplification patterns: (i) the first (50% of cases) was characterized by several scattered signals distributed over the whole nucleus ( Figure 2A); (ii) the second pattern (14% of cases) presented MDM2 signals clustered in specific areas of the nucleus ( Figure 2B); (iii) the third presented overlapping features of the two previous patterns, as signals were contemporarily clustered and scattered (36% of cases). Regardless of these patterns, MDM2 amplified cases also presented extra chromosomal signals that were referred to as being Dmin. These signals were so small that they could be read by the Metafer 4 in 18 of the 43 amplified cases (42%) only. Of note, the Dmin amplification could present focal areas with overlapping signals resulting in an overall blurry appearance [34][35][36][37]. On the other hand, cases resulting negative for MDM2 amplification presented two different patterns: (i) The most common one (82% of cases) showed less than three MDM2 gene signals (MDM2 < 3; mean 2.2; range 1.7-2.9). These cases were considered negative regardless of the MDM2/CEP12 ratio ( Figure 2C). (ii) The second pattern (18% of cases) showed instead three or more MDM2 gene signals (MDM2 ≥ 3; mean 3.7; range 3.1-4.3) together with a gain of the chromosome 12 centromere (mean 3.1; range 2.0-3.9) ( Figure 2D).
We also identified two subgroups of cells with "giant nuclei" (i.e., at least two times larger than nearby nuclei) that presented challenging features. The first subgroup consisted of chromosome 12 polysomy (more than ten MDM2 and CEP12 signals) ( Figure 2E-top inset) that was negative for MDM2 amplification by ratio. The evaluation of the EGFR/CEP7 dual probe expression in this subgroup revealed EGFR/CEP7 polysomy and confirmed the cells' polyploidy ( Figure 2F). The second subgroup is represented by amplified cells with "giant nuclei" harboring increased MDM2 gene signals and was positive by ratio ( Figure 2E-bottom inset).

Discussion
To date, soft tissue lesions are initially approached and diagnosed mainly through the accurate combination of clinical data and morphological features of biopsy samples. This practice, although useful and accurate in several settings, presents crucial drawbacks, especially when dealing with equivocal and misleading histopathological findings, as in the LPS scenario. Indeed, LPS and, particularly, the ALT/WDLPS, DDLPS, and UPS variants present relevant overlapping morphologic features with other types of malignant sarcomas or even with benign entities, such as Lipomas, thus leading to a challenging diagnostic assessment and a broad differential diagnosis process. In this context, the

Discussion
To date, soft tissue lesions are initially approached and diagnosed mainly through the accurate combination of clinical data and morphological features of biopsy samples. This practice, although useful and accurate in several settings, presents crucial drawbacks, especially when dealing with equivocal and misleading histopathological findings, as in the LPS scenario. Indeed, LPS and, particularly, the ALT/WDLPS, DDLPS, and UPS variants present relevant overlapping morphologic features with other types of malignant sarcomas or even with benign entities, such as Lipomas, thus leading to a challenging diagnostic assessment and a broad differential diagnosis process. In this context, the evaluation of MDM2 status is of crucial support, as it represents a diagnostic hallmark of LPS. Regardless of its diagnostic relevance, no guidelines or consensus criteria regarding MDM2 gene amplification interpretation have ever been proposed, whereas previous studies assessing this molecular hallmark adopted significant methodological differences, including a broad spectrum of criteria for probe counting and diagnostic cut-off values [65,66]. This methodological heterogeneity could lead to confounding definitions and hamper the diagnostic reliability/reproducibility, eventually misleading patient clinical management.
Based on published evidence and our data, we hereby provide a set of considerations and recommendations regarding FISH interpretation criteria for MDM2 amplification assessments that should be considered in the LPS diagnostic workup (Figure 4). #1 The analyzed nuclei must be representative of the entire lesion. The FISH analysis is performed in a small, selected area, that, however, has to be representative of the whole lesion. We recommend that an expert pathologist in soft tissue tumors determines the area for FISH analysis, supplying marked hematoxylin-eosin (H&E)-stained slides. This criterion, common to every molecular analysis, should be strictly and routinely applied to avoid misleading results due to nonrepresentative sampling.
#2 The MDM2 amplification is secondary to a Dmin-based mechanism. Dmins are extra chromosomal elements that are considered cytogenetic hallmarks of high-gene amplification. Based on our data, Dmins were the main mechanism behind MDM2 amplification in all the cases evaluated.
#3 MDM2/CEP12 ratio is essential to define MDM2 status. We recommend determining the MDM2/CEP12 ratio and consider only cases with a ratio > 2 as amplified.
Generally, MDM2 amplified cases present many MDM2 gene signals together with a mean of two CEP12 signals. In our experience, observing a >2 MDM2/CEP12 ratio is necessary to differentiate amplified from polysomic cases. We did not experience cases with a ratio = 2, but we consider that this equivocal pattern could be solved by extending the cells count. Furthermore, we observed no cases with MDM2 signals < 4 and a ratio > 2. This eventuality could happen due to the loss of the CEP12 (CEP12 < 1.8), but we would have considered these cases as not amplified without any further analysis. Similarly, cases with several MDM2 signals (e.g., >6) but a <2 ratio should be considered nonamplified as well. In this scenario, the use of other enumeration probes could prove helpful in confirming cells' polysomic nature.
#4 Ambiguous cases require attention and critical review, especially if polysomy is suspected. In ambiguous cases and, in particular, if polysomy is suspected, we recommend analyzing at least 50 nuclei to avoid MDM2 status misinterpretation. In these cases, a careful assessment of the MDM2/CEP12 ratio is crucial to properly determine the amplification status.  Our study presents some limitations, mainly represented by the monocentric collection and analysis of our series and the absence of external validation, which is required to independently confirm our diagnostic recommendations.
This effort is especially important if we consider the evaluation of MDM2 amplification by FISH as a cornerstone of the LPS diagnostic workup, but it is still affected by the lack of consensus guidelines, potentially resulting in different and even misleading approaches. Our study represents the first attempt to solve this controversial scenario and develop formal guidance for the interpretation of this assay to optimize the diagnostic and clinical management of patients with suspected LPS. Screening and reviewing the literature regarding MDM2 amplification, we noticed that several studies used assessment criteria similar to the ones reported in breast cancer HER2 evaluation guidelines. Overall, we discourage this practice, as these guidelines were tailored for a different setting (breast cancer) and a diverse gene with its peculiar amplification pattern. As we already discussed, MDM2 amplification is developed through the Dmin mechanism and is usually characterized by the presence of several gene signals (MDM2 > 10) clustered or scattered over the whole nucleus, with a low centromere number ( Figure 5A). Differently, HER2 amplification mainly occurs with two distinct amplification mechanisms: 30% of HER2-positive breast tumors present a Dmin amplification mechanism with a pattern similar to the ones described for MDM2 ( Figure 5A), but the majority of HER2-positive breast cases (~60%) involve intrachromosomal regions called homogeneously staining regions (HSR) ( Figure 5B) [67,68]. HSR have also been identified in breast cancer cell lines, including BT474, SKBR3, and JIMT-1 [69]. HSR are unusual patterns associated with the gain (CEP > 2), loss (CEP < 2), or coamplification of the centromeric region [70]. The differences could depend on the HSR extension and the presence of other cytogenetic aberrations (translocations, inversions, and deletions) involving the gene loci and the CEP. Furthermore, breast cancer HER2 assessment guidelines are based on different aims which mainly include the prognostic and predictive assessment (the selection of eligible patients for Trastuzumab treatment). Differently, LPS MDM2 amplification harbors a diagnostic value only, to date. Based on this evidence and distinct characteristics, the application of HER2 criteria to define MDM2 status could increase, rather than reduce, inappropriate interpretations and diagnosis. . HSR are chromosomal segments of various lengths but uniform staining intensity. In the chromosomal region where HSR occur, a segment of the chromosome is amplified or duplicated several times. This pattern is the most common in HER2 amplified cases.

Materials and Methods
This study focuses on the FISH assessment of MDM2 status in LPS by combining data acquired at our Institution and the evidence published in the literature so far.

Case Series Construction and FISH Analysis
Formalin-fixed paraffin-embedded (FFPE) tissue blocks of 83 adipocytic tumors (diagnosed from 2014 to 2019) were retrieved from the archive of the Pathology Unit of the Città della Salute e della Scienza Hospital, (Turin, Italy). From each FFPE block, two 4 μm thick tissue serial sections were cut for a tissue adequacy evaluation and FISH analysis. The slides for the tissue adequacy assessment were stained with hematoxylin and eosin and reviewed by two pathologists with soft-tissue expertise ( Figure 6).
The slides for FISH analysis were baked overnight at 58 °C and then deparaffinized. Later, samples were treated with the Invitrogen Spot-light tissue pretreatment kit (Invitrogen Corporation, Camarillo, CA, USA) at 98 °C for 15 min, and enzymatic digestion with a protease solution (pepsin) at 37 °C for 45 to 60 min was then performed. Finally, Our study presents some limitations, mainly represented by the monocentric collection and analysis of our series and the absence of external validation, which is required to independently confirm our diagnostic recommendations.
This effort is especially important if we consider the evaluation of MDM2 amplification by FISH as a cornerstone of the LPS diagnostic workup, but it is still affected by the lack of consensus guidelines, potentially resulting in different and even misleading approaches. Our study represents the first attempt to solve this controversial scenario and develop formal guidance for the interpretation of this assay to optimize the diagnostic and clinical management of patients with suspected LPS.

Materials and Methods
This study focuses on the FISH assessment of MDM2 status in LPS by combining data acquired at our Institution and the evidence published in the literature so far.

Case Series Construction and FISH Analysis
Formalin-fixed paraffin-embedded (FFPE) tissue blocks of 83 adipocytic tumors (diagnosed from 2014 to 2019) were retrieved from the archive of the Pathology Unit of the Città della Salute e della Scienza Hospital (Turin, Italy). From each FFPE block, two 4 µm thick tissue serial sections were cut for a tissue adequacy evaluation and FISH analysis. The slides for the tissue adequacy assessment were stained with hematoxylin and eosin and reviewed by two pathologists with soft-tissue expertise ( Figure 6).
The slides for FISH analysis were baked overnight at 58 • C and then deparaffinized. Later, samples were treated with the Invitrogen Spot-light tissue pretreatment kit (Invitrogen Corporation, Camarillo, CA, USA) at 98 • C for 15 min, and enzymatic digestion with a protease solution (pepsin) at 37 • C for 45 to 60 min was then performed. Finally, the sections were dehydrated in ethanol of different concentrations for the subsequent hybridization. The hybridization was performed indifferently using two commercially available dual-color probes: MDM2 (green spectrum)/CEP12 (orange spectrum) (Abbott Molecular, Chicago, IL, USA) and ZytoLight SPEC MDM2 (green spectrum)/CEN 12 (orange spectrum) (Zytovision. GmbH, Bremerhaven, Germany) (for consistency, the control probe has always been identified as CEP12). The slides were codenatured in an HYBrite System at 72 • C for 5 min (Abbott) or 75 • C for 10 min (Zytovision) and hybridized overnight at 37 • C. The slides were then washed in a 0.7xSSC/0.3% NP-40 solution at 73.5 • C for 3 min (Abbott) or in a 2xSSC/0.3% NP-40 solution at 73.5 • C for 3 min (Zytovision); then, they were dehydrated in ethanol of different concentrations, air-dried, and counterstained with 6-diamidino-2-phenylindole (DAPI). The presence of polyploidy was identified using a dual-color probe for EGFR (7p11) (orange spectrum; Abbott Molecular) and the centromere of chromosome 7 (D7Z1) (green spectrum; Abbott Molecular).
On each slide, five to ten tumor areas of interest were identified, selected, and automatically acquired with the motorized Metafer 4 Scanning System (Carl Zeiss MetaSystems GmbH. Jena, Germany) equipped with AxioImager epifluorescence microscope (one focus plane for DAPI and nine focus planes for green and red spots). An analysis of the MDM2/CEP12 and EGFR/CEP7 probe patterns was performed both with the Metafer 4 software and by counting the MDM2 and CEP12 spots on images taken through Metafer 4 and transferred into the Integrated Set of Information Systems (ISIS) software.
Successively, the MDM2 gene was evaluated on 20 to 200 nuclei in the selected representative areas. Only nuclei with both the MDM2 and CEP12 signals were assessed.
An MDM2/CEP12 ratio higher than two (MDM2/CEP12 > 2) was considered positive for MDM2 amplification, as assumed from the available literature [9, [18][19][20][21][22][23][24]. Cases with a ratio equal or smaller than two were considered not MDM2 amplified regardless of the absolute number of copies (e.g., polysomic cases with a relatively high number of MDM2 copies, but with a high number of CEP12 copies too). for 3 min (Abbott) or in a 2xSSC/0.3% NP-40 solution at 73.5 °C for 3 min (Zytovision); then, they were dehydrated in ethanol of different concentrations, air-dried, and counterstained with 6-diamidino-2-phenylindole (DAPI). The presence of polyploidy was identified using a dual-color probe for EGFR (7p11) (orange spectrum; Abbott Molecular) and the centromere of chromosome 7 (D7Z1) (green spectrum; Abbott Molecular).  Figure 2A). (B) An MDM2-amplified WDLPS (the MDM2 amplification pattern of this cases is represented in Figure 2B); (C,D) Representative images of two lipomas that were analyzed for MDM2 amplification (the corresponding negative MDM2 FISH patterns are represented in Figure 2C,D); (E) A non-MDM2-amplified undifferentiated pleomorphic sarcoma with giant cells (the corresponding MDM2 and EGFR FISH patterns are represented in Figure 2E (including the top inset) and Figure 2F, respectively); (F) An MDM2-amplified DDLPS with scattered giant cells (the corresponding MDM2 FISH pattern is provided in Figure 2E (bottom inset)).  Figure 2A). (B) An MDM2-amplified WDLPS (the MDM2 amplification pattern of this cases is represented in Figure 2B); (C,D) Representative images of two lipomas that were analyzed for MDM2 amplification (the corresponding negative MDM2 FISH patterns are represented in Figure 2C,D); (E) A non-MDM2amplified undifferentiated pleomorphic sarcoma with giant cells (the corresponding MDM2 and EGFR FISH patterns are represented in Figure 2E (including the top inset) and Figure 2F, respectively); (F) An MDM2-amplified DDLPS with scattered giant cells (the corresponding MDM2 FISH pattern is provided in Figure 2E (bottom inset)).

Literature Analysis
An extensive literature review of the published evidence regarding MDM2 status in LPS was performed by querying PubMed, Scopus, Embase, and Web of Science databases.
Up to 90 papers were identified using the following keywords: ("MDM2") AND ("amplification") AND ("FISH") AND ("LPS" OR "Liposarcoma"). Abstracts of conference presentations, case reports, and non-English written papers were excluded. The title and abstract of the selected papers were then screened and assessed for appropriateness, whereas references were double checked to identify potentially neglected relevant articles and ensure literature research adequacy. From each study, details about study design, material and methods, and FISH analysis outcomes were then evaluated and recorded.

Conclusions
In conclusion, clinical data and morphologic features are of crucial relevance to approach LPS diagnosis, but molecular techniques, such as FISH cytogenetic analysis, are increasingly required to achieve a conclusive diagnosis, particularly for specific variants such as ALT/WDLPS and DDPLS. Considering its crucial diagnostic role in the LPS diagnostic workup, MDM2 amplification assessment requires a clearly defined workflow and interpretation criteria.
Based on our experience as a tertiary referral center for LPS diagnostic assessment and considering the current literature evidence, we here proposed a set of criteria for MDM2 FISH assessment as a step towards the development of consensus-based formal guidelines. Informed Consent Statement: Patient consent was waived due to the retrospective nature of the research protocol and the fact that this study had no impact at all on the patients' care.

Data Availability Statement:
The data supporting the findings of this study are not publicly available due to privacy or ethical restrictions but can be obtained upon reasonable request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.