The Role of USP6 Gene Rearrangement by FISH Analysis in Nodular Fasciitis and Histological Mimics


 Background: Nodular fasciitis (NF) is a common benign fibroblast and myofibroblast tumor. Although there have been constant concerns about its histological features and differential diagnosis, NF is still the most easily misdiagnosed mesenchymal lesions, especially over diagnosed as a sarcoma. USP6 gene rearrangement was reported recently to be a reproducible and specific gene change in NF, which strongly assists the diagnosis of NF. We aim to study USP6 Gene Rearrangement in NF and other spindle cell lesions mimicking NF, and to investigate the use of USP6 Gene Rearrangement in differentiating NF and its mimics.Methods: USP6 translocation by fluorescence in situ hybridization (FISH) was studied on 40 cases of NF and 54 cases of other spindle cell lesions (including 10 cases of spindle cell sarcoma) that mimic NF. Results: Thirty-four cases (85.5%) of NF showed USP6 gene translocation, including 10 cases which previously diagnosed or considered as malignant diseases by the referring pathologists. All of other 54 cases of spindle cell lesions were negative for USP6 translocation.Conclusions: USP6 gene rearrangement by FISH analysis is a reliable test for diagnosis of NF, and also a helpful adjunct to prevent over diagnosis of NF. Positive USP6 gene rearrangement by FISH could be a useful marker to exclude malignant lesions that mimics NF.


Introduction
NF is a commonly encountered, benign broblast and myo broblast neoplasm, which usually presents as a small but rapidly growing mass [1][2][3] . It predominant occurs in young to middle-aged adults between 20 and 40 years, without sexual predisposition. Upper extremities are the most common affected site, followed by the trunk and head and neck. Cases occur in extremities and hands and feet are infrequent [2][3] . Most cases are located in the super cial subcutaneous fascia, but some can involve deep muscle tissues regardless of its location. Histologically, NF displays spindle cells with myo broblasts features arranged into loose fascicles, showing storiform or S-like appearance. The spindle cells are usually bland looking, with variable myxoid vascularized to densely sclerotic stroma, and erythrocyte extravasations foci, with frequent foci of erythrocyte extravasations.
These histological features of NF make it easily misdiagnosed as other spindle cell malignant tumors, such as lowgrade myo broblastic sarcoma, myxo brosarcoma, synovial sarcoma, and so on, which is harmful for the patients with risk of overtreatment. On the other hand, it should also be differentiated from benign or intermediate lesions such as desmoid bromatosis, in ammatory myo broblastic tumor, dermato broma, etc. Despite being recognized for over six decades and heightened awareness of this entity constantly, NF is still among the most common benign mesenchymal lesions misdiagnosed as a sarcoma. A recent report showed that the misdiagnosis rate of NF was more than 50% and the over diagnosis rate was more than 30%. Making the diagnosis accurately and avoid over diagnosis of NF is still a great challenge for surgical pathologists, due to its confusing morphology and non-speci c immunophenotype.
USP6 gene, also known as TRE17, encodes an ubiquitin-speci c protease (ubiquitin-speci c peptidase 6). USP6 protein is involved in regulation of ubiquitination and transport. It may also act as a driving oncogene due to gene rearrangement. A clear association between nodular fasciitis and USP6 gene rearrangement had been established. Molecular ndings showed that MYH9-USP6 translocation is the most common rearrangement in NF. This nding resulted in its rede nition as a benign true neoplasm. In 2013, Amary et al conducted a series of studies identi ed USP6 rearrangements by FISH in 31 of 34 NF cases, with a sensitivity of about 91%. Patel and colleagues reported similar ndings, identi ed USP6 gene rearrangements in 20 of 26 NF cases. These results suggest that USP6 gene rearrangement test may serve as an important adjunct in the diagnosis and differential diagnosis of NF.
Up till now, only few literatures studied the USP6 gene rearrangement in NF and its mimics. Herein, we collected a cohort of NF and a group of its mimics, reviewed the clinical pathological features and morphology, and studied the USP6 rearrangement by FISH, aiming to analyze the use of USP6 rearrangement by FISH in diagnosis and differential diagnosis of NF.

Cases selection
Spindle lesions which diagnosed as NF, or have been mentioned as differentiated with NF, or have been tested USP6 gene translocation by FISH in the department of pathology, Sun Yat-sen University Cancer center from August 2017 to January 2020 were collected. To ensure the accuracy of the diagnosis, two senior pathologists speci ed in soft tissue (Yingchun Zhang and Huilan Rao) reviewed all the collected cases with initial routine hematoxylin and eosin (HE) and immunohistochemistry (IHC) slides. Some cases were also submitted to molecular test to con rm or exclude the corresponding lesions, for example, SS18-SSX fusion gene FISH analysis for synovial sarcoma, COL1A1-PDGFB fusion gene for protuberant dermato brosarcoma, etc. Following criteria of the World Health Organization (WHO, 2013), we divided those cases into three groups: NF, benign or intermediate lesions similar to NF, and malignant tumors similar to NF. Patients' age and sex, tumor location, related clinical information and follow-up data were collected. Patients diagnosed as nodular fasciitis and malignant tumors were followed up in detail. Sun Yat-sen University cancer center (SYSUCC) institutional ethics committee approved this study.
FISH analysis FISH analysis for USP6 rearrangement was performed formalin-xed, para n-embedded (FFPE) sections by using speci c Dual Color Break Apart Probes following the manufacturer's instructions (LBP Medicine Science and Technology Co., Ltd, Guangzhou, China). The FISH signals were assessed under an Olympus BX 51 microscope (Olympus, Japan) equipped with a triple-pass lter (DAPI/Green/Orange, Vysis) Images were acquired using the Cell Sens Standard Imaging System (version 1.7, Olympus Microsystems Inc., Japan). USP6 signals were independently scored by two experienced pathologists (and). Cells with an orange and green signal were scored as normal. The distance of break-apart signals between one of the signals, and one orange and green signal were more than twice, indicating a rearrangement for the USP6 gene. A total of 100 tumor cells were counted and the percentage of the cells with split signal was recorded. Positive cases with a rearrangement for the USP6 gene were de ned as those with > 10% of cells showing the break-apart signal.

Results
General clinical and demographic features of the total cohort After eliminating the cases of duplication and failure in detection, 94 cases of spindle cell lesions were retrieved, including 40 cases of NF, 44 cases of benign or intermediate lesions and 10 cases of malignant tumors histologically similar to NF. The cohort consisted of 52 males (55.3%) and 42 females (44.7%), ranging in age from 1 to 70 years (median, 30 years). The patient's clinico-pathological information is shown in Table 1 and Table 2.  Of the NF group, 9 were in-house cases and 31 were consult cases, with the latter being the majority. The most common site affected is head and neck (n = 15, 37.5%), followed by the trunk (n = 10, 25.0%), upper extremities (n = 7, 17.5%), and lower extremities (n = 8, 20.0%). Tumor size ranged from 1.0 to 6.2 cm (median, 2.0 cm). Tumor size of 30 cases was recorded, with 21 (73.3%) cases smaller than 3 cm. Median age was 32.5 years (ranging, 1-60 years). Twenty-ve (25/40, 62.5%) patients were in their second to fth decades, and 10 (10/40, 25.0%) patients were children and adolescents (age ≤ 15 years). The preoperative duration of 28 cases of patients were obtained, ranged from 1 day to > 2 years, 53.6% (15/28) of which were less than 3 months, while 28.6% (22/28) were less than 6 months. Thirty-one cases of NF were consult cases, and the initial submitted diagnoses and the reason for external consultation for 27 cases were recorded. Some of them were given de nite diagnosis by the referring pathologists, while others with inde nite diagnosis were given assessment of potential biological behavior (e.g., favoring a sarcoma or sarcomatoid neoplasm). Of these 27 cases with diagnostic opinions, only 4 cases were correctly diagnosed as NF, while 9(9/27, 33.3%) cases were initially diagnosed as malignant tumors, including lowgrade malignant bromyxoid sarcoma, myxoid brosarcoma, myxoid liposarcoma, synovial sarcoma, etc. Others were named as desmoid bromatosis, in ammatory myo broblast tumor and other benign or intermediate lesions.
Microscopically, our cases had a wide variety of morphologic features: many cases had abundant myxoid matrix, and a few cases had a higher cellularity, while some lesions was characterized by increased collagen production, sometimes with deposition of keloid-like collagen bers. All the cases were composed of plump, immatureappearing broblasts and myo broblasts that arranged in short intersecting fascicles. In the area with obvious mucoid degeneration, those cells were arranged in a haphazard fashion, often called "tissue culture appearance" or "feathery pattern", usually accompanied extravasated erythrocytes and scattered lymphocytes. Mitotic gures are fairly common, but cytological pleomorphism and nuclear atypia were absent. Most lesions were well circumscribed but unencapsulated, whereas a minority of cases had poorly de ned borders with in ltration into the surrounding skeletal muscle, adipose tissue, and entrapped nerve fascicles. The stroma was at least focally myxoid in all cases and there were frequent cystic spaces in these myxoid areas. In some cases, the cells were abundant and dense, with more mitotic images and similar morphology to malignant tumors (Fig. 1). Mitotic gures are fairly common, but atypical mitoses are virtually never seen. Possibly related to the duration of the lesion, there are three major subtypes of NF: mucinous, cellular and brous.
Due to its variable morphologic features, nodular fasciitis were easily misdiagnosed as other benign of even malignant lesions with mucinous appearance, of with proliferation of broblasts and myo broblasts, such as lowgrade myo broblast sarcoma, desmoid-type bromatosis, cellular brous histiocytoma, solitary brous tumor, myositis ossi cans, in ammatory myo broblast tumor, et al.
Clinic-pathological features of 54 cases of spindle cell lesions mimicking NF were summarized in Table 2. Some of the cases displayed proliferation of broblast or myo broblast features, some showed mucinous stroma, some showed proliferation of collagen ber, and some showed dense spindle cellularity. All these features were mimicking different periods of NF, which have been described as mucinous, cellular and brous subtypes (Fig. 2).

Frequency of USP6 translocation in NF and in NF-like lesions
The results of FISH analysis of USP6 translocation was shown in Table 3. Of these 98 cases, 35 (35/98, 37.6%) were positive. Within the 40 cases of NF, 35 (35/40, 85.0%) of them were positive. All of other 58 cases that mimic NF were negativet. It is worth noting that all 9 cases that initially diagnosed as malignant tumors by the submitting pathologist were all positive for USP6 gene rearrangement. The microscopic images and FISH results of representative cases showed in Fig. 3.

Discussion
In this study, we reviewed a cohort of NF and its histological mimics, collected their clinico-pathological characteristics, and studied their USP6 gene rearrangement status by FISH. Clinico-pathological analysis showed that NF occurred over a wide age range, from 1 to 60 years old, and predominant in young to middle-aged adults between 20 and 50 years old. Most of the lesions grow rapidly and the lesion duration was short, with most cases less than 3 months. The greatest dimensions of NF were smaller than 3 cm in 73.3% of cases. The follow-up data showed that most NF patients only underwent complete resection, and no recurrence or metastasis happened. These results demonstrated important clinical features of NF: rapid growth, small nodule, and benign course.
Similar to previously reported series, the pathologic spectrum of NF is wide and the morphology can vary signi cantly between and within cases. However, in the case of complete resection, more or less mucoid degeneration, or microcapsule structure with lymphocyte in ltration can be seen in the background of spindle tumor cells in most cases, suggesting that this is a very important diagnostic clue. Mitotic gures are fairly common, but atypical mitoses are virtually never seen. The absence of atypical mitosis is one of the most important morphological features to distinguish nodular fasciitis from malignant tumors.
It is worth noting that NF is very easy to be misdiagnosed, especially over diagnosed as a sarcoma. In 1991, Montgomery EA and Meis JM reviewed 53 cases of NF and found that the misdiagnosis rate was 57%, of which 21% were over diagnosed as sarcomas. In 2018, Erber R and Agaimy A reported that of the 36 consult cases with diagnostic suggestions, only one third of NFs were correctly diagnosed by submitting pathologists, while one third was initially judged as malignant. Our data showed that only 9.6% of NF was correctly diagnosed by submitting pathologists. Over diagnosis as malignancy (ranging from low-grade to high-grade sarcoma) was rendered or favored in 35.4% of cases, which would cause catastrophic clinical consequences. The main reasons for over diagnosis attributed to not considering NF as a differential diagnosis, or mislead by features such as increased mitotic activity, high cellularity, high Ki-67 index, deep in ltrative location, prominent myxoid change, and good circumscription being not compatible with a diagnosis of NF. Since the histological features of NF is confusing, while immunohistochemistry staining is not helpful, sometimes even misleading, more objective and speci c auxiliary methods are urgently needed. This study was ethically approved by Sun Yat-sen University Cancer Center IRB, and has been granted exemption from requiring informed consent to participate from participants.

Consent for publication
Not applicable.
Availability of data and materials The datasets generated during and/or analyzed for this study are available from the corresponding author upon reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
Not applicable.
Authors' contributions ZYC designed the study. FYF and ZYC drafted the manuscript and contributed to data analysis. ZYC and RHL contributed to the diagnoses. ZY contributed to clinico-pathological data collection. LJP assisted in patient selection and gure preparation. ZZC contributed to FISH analysis. LM contributed to the patients' follow-up. All authors read and approved the nal manuscript. Table   Table 4 is not available with this version. Figure 1 Different histological features of NF (H&E, 400x). a. Plump, immature-appearing broblasts and myo broblasts arranged in short intersecting fascicles. b. Area with obvious mucoid degeneration, and cells were arranged in a haphazard fashion (called "tissue culture appearance" or "feathery pattern). c. Extravasated erythrocytes and scattered lymphocytes in ltrate. d. Collagen production of stroma.

Figure 2
Different histological features of spindle cell lesions mimicking NF (H&E, 400x). a. In ammatory myo broblast tumor showed myo broblasts arranged in Fascicles, with numerous in ammatory cells. b. low-grade myo broblast sarcoma showed fat spindle cells with abundant plasma arranged in fascicles. c. myositis ossi cans showed long spindle cell range in fascicles, with abundant eosinophilic plasma, and bony tissues with osteoblast around were seen. d. desmoid-type bromatosis showed broblasts and myo broblasts arranged in fascicles, with collagen rich stroma.