Differential diagnosis of lipoma and atypical lipomatous tumor/well‐differentiated liposarcoma by cytological analysis

Abstract Background Adipocytic tumors are the most common soft tissue tumors, with lipomas and atypical lipomatous tumor/well‐differentiated liposarcomas (ALT/WDL), which comprise most cases. Preoperative differential diagnosis of lipoma or ALT/WDL can provide important information for decisions regarding treatment. We evaluated the cytological findings of 20 cases of lipoma and ALT/WDL. Methods Fluorescence in situ hybridization (FISH) was performed on formalin‐fixed paraffin‐embedded specimens (FFPE) to examine mouse double minute 2 homolog (MDM2) amplification in all cases. Tissue samples were collected from the center of the surgical materials, stained with Pap, and evaluated for 12 cytological parameters by six cytotechnologists. Results The findings regarding large atypical cells, multinucleated cells, and nuclear pleomorphism were highly concordant among the cytotechnologists and were associated with MDM2 amplification. Large atypical cells, considered a highly specific feature of ALT/WDL, were not observed in lipoma cases. However, the sensitivity of the large atypical cell findings was not high (67%); therefore, comprehensive evaluation of multinucleated cells and pleomorphism is crucial for predicting ALT/WDL diagnosis. FISH of MDM2 on Pap‐stained specimens was performed in four cases. In two, the results were similar to those of MDM2 FISH performed on FFPE sections and were reproducible, whereas in the other two, the signal could not be evaluated because of the strong background coloration. Conclusions Cytology specimens may be useful for the preoperative diagnosis of adipocytic tumors, particularly if the FISH conditions for Pap‐stained specimens and the detection accuracy of MDM2 amplification can be improved.


| INTRODUCTION
Adipocytic tumors are the most common soft tissue tumors and mostly comprise lipomas and atypical lipomatous tumor/welldifferentiated liposarcomas (ALT/WDL). ALT and WDL primarily differ on the anatomical location and resectability of the tumor. Tumors that occur in the limbs or trunk and can be resected are called ALT, while those that occur in the retroperitoneum or mediastinum and must be resected at the margins are called WDLs. The fifth edition of the WHO classification classifies ALT and WDL into the same category. 1 ALT/WDL can become highly malignant by dedifferentiation or recurrence, thereby making it important to differentiate them from lipomas, which are benign tumors, for appropriate treatment and determination of the follow-up period after tumor resection. 2 The genomic abnormalities described in dedifferentiated liposarcoma support the fact that this tumor corresponds to a malignant adipocytic tumor showing progression from ALT/WDL to non-lipogenic sarcoma of variable aspect and grade. 3 The mouse double minute 2 homolog (MDM2) protein suppresses TP53, a tumor suppressor gene. Fluorescence in situ hybridization (FISH) analysis of formalin-fixed paraffin-embedded (FFPE) tissues has revealed MDM2 amplification in ALT/WDL and is currently used to differentiate these tumors. [4][5][6] With the development of new molecular tests showing high diagnostic specificity, fine-needle aspiration cytology (FNAC) has gained acceptance for the preoperative assessment of soft tissue tumors. 7 FNAC represents a versatile, low-cost, well-tolerated diagnostic strategy with advantages over histological biopsies. 8 However, a detailed comparison of the cytological findings of lipoma and ALT/WDL with MDM2 amplification has not yet been reported. In this study, we compare the cytological features of lipoma and ALT/WDL and use cytological findings for differential diagnosis. We also performed FISH to examine MDM2 amplification in cytological specimens and evaluate the usefulness of cytology for the differential diagnosis of adipocytic tumors.

| MATERIALS AND METHODS
We reviewed the clinical and histological data of 20 patients with lipoma and ALT/WDL who had undergone resection at Kanagawa Cancer Center between 2018 and 2020. One ALT/WDL case showed partially dedifferentiated areas. We evaluated age, sex, and maximum tumor diameter as clinical parameters. Tissue samples (2-3 mm in size) were randomly collected from the center of the surgical specimens, mimicking FNAC, and subjected to Pap-staining. If the case showed a dedifferentiated area, the sample was collected from the ALT/WDL area. Six cytotechnologists (CTs) evaluated the cell morphology, number of lipoblasts, size of adipocytes, nuclear pleomorphism, intranuclear vacuoles, multinucleated cells, nuclear enlargement, unequal size of nuclei, irregular nuclear borders, hyperchromasia, prominent nucleoli, large atypical cells, and background necrosis as a cytology routine observation. (Figure 1A-G). Large atypical cells were defined as cells with hyperchromasia and irregular nuclear enlargement ( Figure 1H-l).
The morphology of each cell was evaluated on a scale of 1 to 4 (1: almost none, 2: a little, 3: common, 4: prominent), and background necrosis was assessed on a scale of 0 to 1 (0: absent, 1: present). Furthermore, each CT estimated the histological type of the lipoma or ALT/WDL based on cytological findings. More than 200 cells were examined using a WinROOF2018 image analyzer (MITANI Corporation, Tokyo, Japan), and the short nuclear diameter was measured for samples subjected to Pap staining.
The FFPE blocks of surgical samples were prepared for histological examination. The resected specimens were fixed in 10% neutral-buffered formalin. At least one block per centimeter of the largest diameter of the tumor was prepared for histological evaluation. If the maximum diameter was 10 cm, more than 10 FFPE blocks were prepared for histological evaluation. The two pathologists evaluated the nuclear atypia of adipocytes and atypical stromal cells and made a histological diagnosis. Immunostaining for MDM2 and CDK4 and FISH examination for MDM2 were performed in all cases using FFPE specimens. In addition, in four cases, FISH for MDM2 was performed using Pap-stained specimens.
The histological type determined via at least four of the six CTs was used as the cytological diagnosis result for analysis. We used the Mann-Whitney U test to analyze the association between clinical findings, histological diagnosis, cytological diagnosis, and length of the short diameter of the cell nuclei and the presence of MDM2 amplification.
We analyzed the association between each cytological finding and each diagnosis based on Pap-stained specimens using Spearman's correlation and then examined the mean values. We also analyzed the association between the total score of the cytological findings of the six CTs and the presence of MDM2 amplification using Spearman's correlation. Statistical analysis was performed using SPSS version 26 software (SPSS Inc., Chicago, IL). Statistical significance was set at p < .05.

| Immunohistochemistry
Deparaffinized tumor sections were stained for CDK4 (Clone DCS-31, Thermo Fisher Scientific, Waltham, MA) and MDM2 (Clone IF2, Thermo Fisher Scientific) using the heat-induced epitope retrieval method. Appropriate positive and negative controls were used for all analyses. Immunostaining was evaluated based on the intensity and proportion of the tumor cells in each specimen. The intensity of staining was defined by applying Allred scoring 9 as follows: 3+, strong; 2+, moderate; 1+, weak; À, no staining. The proportion of staining was measured for each specimen and classified by applying Allred scoring as follows: 5, > 66%; 4, 66%-33%; 3, 33%-10%; 2, 10%-1%; 1, < 1%; 0, 0%. Cases were defined as MDM2-positive if the Allred score of the marker (defined as the combined value of the intensity score and proportion score) was more than 1. Moreover, cases were considered CDK4-positive if the Allred score of the marker was more than 5. FISH for MDM2 was performed using Pap-stained specimens to evaluate the cytological findings in four cases. Using FISH in FFPE samples, two cases showed MDM2 amplification, whereas two did not. After xylene was removed using 100% ethanol, the samples were destained with ethanol hydrochloride for 2-3 h. Then, the samples were washed with 100% ethanol and incubated overnight at room temperature (20-30 C). The specimens were immersed in 0.2% hydrochloride for 20 min, followed by immersion in distilled water for 1 min, a wash buffer for 5 min, and finally a protease solution (Abbott Molecular; pre-warmed to 37 ± 1 C) for 10 min. The samples were again immersed in wash buffer for 5 min, and the procedure was repeated. Next, the samples were immersed in 10% neutral-buffered formalin for 10 min, followed by immersion in wash buffer for 5 min, and the procedure was repeated. Finally, the Vysis ® LSI ® MDM2 SpectrumOrange Probe was added to the denatured DNA, and hybridization was carried out at 73 C for 3 min, followed by overnight incubation at 37 C. The cells were washed to eliminate nonspecific signals by immersing them in hybridization wash buffer (2X SSC/0.3% NP-40; Abbott Molecular) preheated to 72 ± 1 C for 2 min. The specimens were immersed in wash buffer and DAPI was added, followed by observation with a fluorescence microscope Ti-E equipped with a triple bandpass filter set, DAPI/Green/Orange v2 (Nikon Corporation, Tokyo, Japan).

| RESULTS
The participants included 13 males and 7 females, with a mean age of 56 years. The most commonly affected site was the thigh (n = 10), and the mean maximum tumor diameter was 133 mm. Seven cases were histologically diagnosed as lipomas and 13 as ALT/WDLs. The  (Table 3). In the group with MDM2 amplification, the total scores of large atypical cells, multinucleated cells, and pleomorphism were higher than those in the group without MDM2 amplification. The mean total score of the above three findings was less than five in the  No large atypical cells were found in the lipoma cases (Table 4).
FISH examination was performed on Pap-stained specimens from two cases in which MDM2 amplification was confirmed by FISH using FFPE specimens (Table 5, Figure 2  In a few cases of ALT/WDL, atypical cells were identified relatively easily in cytological specimens, even when it was difficult to identify atypical cells in the histological specimens ( Figure 4).
Although tissue specimens can be examined over a wide area, cytological specimens can be used to evaluate nuclear atypia in more detail and may help improve diagnostic accuracy in combination with tissue specimens.
The length of the nucleus short diameter in the Pap-stained cytological specimens was significantly greater in cases showing MDM2 amplification. Additionally, a significant difference in SD values was observed, suggesting that unequal nuclear size was prominent in the MDM2 amplification group, which may reflect pleomorphism in cytological findings. In cases without MDM2 amplification, few cells with a nucleus short diameter > 7 μm were observed, but none of the cells had a nucleus short diameter greater than 9 μm. We did not perform immunostaining on cytological specimens since some were used for FISH examination. In contrast, we observed a significant difference between the expression of MDM2 and CDK4 as evaluated using immunostaining and the presence of MDM2 amplification via FISH in the tissue specimens. However, we found an association between the immunostaining and FISH results only after the detailed evaluations of immunostaining and comparison of multiple cases. Immunostaining of MDM2 in ALT/WDL has been reported to be weak, localized, and not highly positive. 10,11 Therefore, immunostaining of MDM2 may not be a useful tool for the differential diagnosis of lipoma and ALT/WDL in cytology. tumors may be predicted from these samples ( Figure 5).
It is possible to detect genes using small specimens, and minimally invasive FNAC may be used for preoperative diagnosis of various tumors, including soft tissue tumors. 5 In the era of genetic analysis of small tissue specimens, preoperative diagnosis using FNAC, a minimally invasive test, may play a major role in determining treatment plans. Familiarity with the cell morphology of cytological specimens derived from tumors is important for making a differential diagnosis.
Preoperative diagnosis of lipoma or ALT/WDL can provide important information for deciding the treatment plan. Examining the diagnostic accuracy of cytological morphology in actual aspiration biopsy specimens is essential. Given that the size of the patient data set was small (n = 20), we will consider a follow-up study with a larger cohort. Our results may facilitate differential diagnoses for patients with lipoma and ALT/WDL and assist clinicians in making treatment decisions.

| CONCLUSION
Lipoma or ALT/WDL can be predicted with a high probability by evaluating the cytological findings of multinucleated cells, pleomorphism, and large atypical cells in Pap-stained tissue specimens. However, it is sometimes difficult to confirm the diagnosis based only on cell morphology, and further confirmation of MDM2 amplification by FISH may therefore be required. With an improved accuracy of MDM2 FISH in cytological specimens, cytology may potentially be a useful tool for the preoperative differential diagnosis of adipocytic tumors.

ACKNOWLEDGMENTS
We are grateful to Dr. Masayuki Takagi for advice on the histological diagnosis and to the technical staff at Kanagawa Cancer Center (Mei Sawaguchi, Ryuji Nasu, Shigeko Iwanade, Tomomi Hasumoto, and Kazuhisa Kitamura) for evaluating the cytological findings and their expert technical assistance.

CONFLICT OF INTEREST
No conflict of interest declared.

ETHICS STATEMENT
We applied the opt-out method to obtain patient consent for this study. This study was approved by the Research Ethics Review Committee of the Kanagawa Cancer Center (approval no. 2020 epidemiology-22).

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.