Main

Nodular fasciitis is a self-limited myofibroblastic/fibroblastic tumor, which usually presents as a small but rapidly growing mass.1, 2, 3, 4 Nodular fasciitis can occur in any age, with peak incidence between 20 and 40 years of age, and is found equally in males and females.4, 5 It is most commonly identified in the upper extremities, head, neck, and trunk, but can be seen almost anywhere,6, 7 including superficial presentations.8 Excision is typically curative and nodular fasciitis may spontaneously regress over time in the absence of treatment.1, 2, 3, 9, 10, 11, 12

Histologically, these tumors are composed of plump fibroblasts and myofibroblasts13 that show a ‘tissue culture’ appearance2 and are arranged in short, irregular bundles and fascicles. The cells contain oval nuclei with prominent nucleoli,1, 2 but there is typically lack of pleomorphism/nuclear atypia and necrosis.1 The overall appearance can vary, as these tumors can be cellular, hyalinized, or show a variably myxoid matrix.1, 2, 3, 4, 14 In addition, microcystic change, erythrocyte extravasation, lymphocytes, histiocytes, and multinucleated giant cells may be seen.1, 2, 3 Because of its rapid growth, cellularity, prominent mitotic activity,2 and locally infiltrative growth pattern,1, 3 nodular fasciitis can be misdiagnosed as a sarcoma.6, 15

Previously, nodular fasciitis was speculated to be a reactive process because 10–15% of patients report a history of trauma.14 However, Erickson-Johnson et al16 identified a balanced rearrangement involving the USP6 gene in 44/48 cases, supporting the idea that nodular fasciitis represents a transient, often self-resolving neoplasm driven by overexpression of USP6. 5′ RACE PCR and reverse transcription-PCR (RT-PCR) identified two fusion transcript types involving MYH9 exon 1 and either USP6 exon 1 or exon 2. However, break-apart fluorescence in situ hybridization (FISH) for MYH9 showed balanced rearrangement in only 31/48 cases. Additional studies using FISH for USP6 rearrangement have shown a sensitivity of 83–93%17, 18, 19, 20 and a specificity of 100%.18, 20 In contrast, RT-PCR yielded only 53–74% sensitivity.17, 19 These findings suggest that other gene partners may be involved in a subset of these tumors.

We examined a large series of nodular fasciitis cases to characterize alternative fusion genes and transcript variants. Samples were initially screened by FISH and RT-PCR. We then used a next-generation sequencing assay, the FusionPlex Sarcoma Panel (ArcherDx, Boulder, CO, USA), to identify novel fusion events in these samples. This assay performs targeted RNA sequencing using anchored multiplex PCR, which utilizes unidirectional gene-specific primers to detect fusion genes by targeting one of the fusion partners—USP6 for this present series.21 Anchored multiplex PCR technology is particularly suitable for the identification of fusion genes when there is alternative splicing and the possibility of multiple partner genes, as in the case of nodular fasciitis. The FusionPlex Sarcoma Panel targets 127 exons in 26 genes for agnostic detection of gene fusions without a priori knowledge of the partner gene. We sought to determine whether this assay would detect fusion transcripts not identified by our RT-PCR assay (which is focused on the known MYH9USP6 fusion transcripts).

Materials and methods

Tumor Samples

With institutional review board approval, we searched the pathology archives of three institutions (University of Texas MD Anderson Cancer Center, Texas Children’s Hospital, and Leiden University) and identified 26 cases of nodular fasciitis diagnosed between 2002 and 2013, with available formalin-fixed, paraffin-embedded or frozen (case 5) tissue (Table 1). Cases were reviewed by bone and soft tissue pathologists (WLW, AJL, ED, SB, and JVMGB). Patient age and sex, tumor location, and available clinical information were obtained. One case (case 15) had histological features in keeping with nodular fasciitis but was located in the hand/finger and had a differential diagnosis of cellular fibroma. Seven cases were previously published in a study on superficial/dermal nodular fasciitis,8 but were further characterized and re-analyzed for this present study.

Table 1 Clinical–pathologic features

USP6 FISH

Break-apart FISH was performed on interphase nuclei for detection of USP6 rearrangement. A customized probe was developed using direct-labeled bacterial artificial chromosome DNA clones, as previously described.8 Briefly, clones RP11-794K10 and RP11-124C16 (telomeric to USP6) were both labeled using SpectrumGreen dUTP (Abbott Molecular; Abbott Park, IL, USA). Clones RP11-1022O13 and RP11-457I18 (centromeric to USP6) were labeled with SpectrumOrange dUTP (Abbott Molecular). All bacterial artificial chromosome clones were obtained from the BAC-PAC Resource at Children’s Hospital Oakland Research Institute (http://bacpac.chori.org/).

The probe mixture consisted of 0.5 μl of each labeled DNA combined with 1 μl of nuclease-free water and 5 μl LSI/WCP Hybridization Buffer (Abbott Molecular). Four-micrometer-thick tumor sections were prepared for hybridization using the Paraffin Pretreatment Reagent Kit II (Abbott Molecular), per manufacturer recommendations; however, protease treatment was increased to 32 min. Hybridization was performed at 37 °C for up to 72 h in a ThermoBrite System (Abbott Molecular). Post hybridization, slides were washed in 2 × saline-sodium citrate buffer (SSC) with 0.3% IGEPAL at 70 °C for 2 min, followed by two additional washes with 2 × SSC with 0.3% IGEPAL at room temperature for 1 min each. Vectashield Mounting Media with 4,6-diamino-2-phenylindole (Vector Laboratories, Burlingame, CA, USA) was used for counter stain and for anti-fading effects. FISH analysis was performed with a multi-filtered fluorescence microscope (Olympus BX51: Olympus; Center Valley, PA, USA) and the CytoVision imaging system (version 7.4, Leica Microsystems Inc.; Buffalo Grove, IL, USA), following standard procedures. FISH signals were scored by two pathologists (DL-T and NRP), and 200 interphase nuclei were scored for fused or split orange and green signals. Cases in which 20% of interphase nuclei had split signals were considered positive for USP6 rearrangement (Figure 1d).

Figure 1
figure 1

(a) Schematic of alternative MYH9USP6 fusion types with corresponding design of RT-PCR primers. (b) Schematic of the resulting fusion transcripts, showing the juxtaposition of MYH9 exon 1 (encoding the 5′-UTR) to either exon 1 or 2 of USP6. Both fusion transcripts contain the entire coding sequence of USP6. Fusion type 1 may be detected with both brown (RS) and blue (RL) reverse primers, while fusion type 2 can only be detected using the blue (RL) reverse primer. (c) Case 5, electropherograms showing the type 1 (top) and type 2 (bottom) fusion transcripts with in-frame fusion of MYH9 to USP6. (d) Case 2, break-apart FISH for USP6 showing rearrangement with splitting of the green and orange signals.

RT-PCR for MYH9USP6 Fusion Transcript

RT-PCR was performed as previously described.8 Briefly, total RNA was extracted from formalin-fixed, paraffin-embedded tissue scrolls using the Ambion RecoverAll Total Nucleic Acid Isolation Kit (Ambion, Life Technologies, ThermoFisher; Austin, TX, USA), as per the manufacturer’s instructions. Reverse transcription of RNA to cDNA was performed using the Invitrogen Super Script III Reverse Transcriptase Kit (Invitrogen, Life Technologies; Carlsbad, CA, USA), according to the manufacturer’s instructions, with random hexamers as primers. One forward primer (MYH9-For: 5′-GCACGGAAGGCTAAGCAAG-3′) and two reverse primers (USP6-RS: 5′-GGATGTGGATGTGAACTGCG-3′ and USP6-RL: 5′-CGGTGTCCCTTGTCATACTTC-3′) were used. Primers were designed to detect the two known fusion types16 (Figure 1a–c). PCR was performed with an annealing temperature of 55 °C.

PCR products were resolved on 2% agarose gel by electrophoresis with ethidium bromide staining for visualization. Bidirectional Sanger sequencing of the PCR products was performed to confirm amplicon identity. Analysis was performed using web-based alignment tools and gene databases (http://blast.ncbi.nlm.nih.gov/Blast.cgi, http://www.ncbi.nlm.nih.gov/gene) with mRNA reference sequences for MYH9 (NM_002473.5) and USP6 (NM_004505.3).

Amplification of the housekeeping gene beta-actin was used as an RNA integrity control. Positive controls for USP6MYH9 fusion consisted of clonally amplified plasmids derived from sequence-confirmed cDNA from positive patient samples. In addition, negative controls were tested with each set of reactions.

Targeted RNA Sequencing by Anchored Multiplex PCR

Total RNA was extracted from nine formalin-fixed, paraffin-embedded nodular fasciitis specimens and quantified using the Qubit RNA assay (ThermoFisher Scientific; Waltham, MA, USA). Next-generation sequencing libraries were prepared with anchored multiplex PCR-based methodology as previously described,21 using the FusionPlex Sarcoma Panel. Libraries were generated according to the manufacturer’s protocol using 100 ng of total RNA. Amplifiable cDNA quality was assessed using the PreSeq RNA QC assay. Final libraries were quantified using the KAPA Library Quantification Kit (Kapa Biosystems, Roche; Wilmington, MA, USA) for Illumina Platforms. Illumina paired-end indexed libraries were sequenced 8-plex on a MiSeq (2 × 150 bp, v2 chemistry) and data analyzed on a vendor-provided virtual-machine-based analysis pipeline with custom-developed output scripts. On average, sequencing with the Archer FusionPlex Sarcoma Panel generated a total of 1.07 million paired-end reads with ~28 980 unique RNA reads per sample. Sequencing quality was assessed by % purity-filtered reads, % bases >Q30, total number of reads, total number of unique RNA reads, and percent of aligned reads with high mapping quality. Eight of nine cases yielded optimal sequencing metrics, with one case, case 10, showing low unique RNA read counts (12 751 unique RNA reads). Therefore, eight out of nine cases proceeded forward to analytic processing and fusion detection. All detected fusion genes were confirmed using RT-PCR, followed by Sanger sequencing and gel electrophoresis, as described above but with specifically designed PCR primers.

Results

Clinical and Pathologic Features of Samples

Clinical and pathological parameters are summarized in Table 1. Mean age was 31.5 years (range=6–65), with a male to female ratio of 2:1. The most common tumor sites included the upper extremities and head and neck, with additional locations including pelvis and pelvic lymph node. One case was initially diagnosed as sarcoma. Tumors ranged in size from 1.0 to 11.0 cm, with the largest sample located within the psoas muscle (case 1). In all, 10 samples were biopsies, while 16 were excision specimens (Table 2). Histologically, the samples showed a combination of various patterns consistent with nodular fasciitis, including hypercellular, sclerotic, myxoid degeneration, and microcystic change (Figure 2). Notable cases include the 11.0 cm tumor, which is larger than usual for nodular fasciitis (case 1), and another case located in the tendon of the hand (case 15). In both cases, the histological findings were consistent with the diagnosis of nodular fasciitis (Figure 3). Follow-up information was available for 12 cases, with a mean follow-up time of 12.5 (1–33) months. In 10 cases, the tumor was excised without recurrence. In the remaining 2 cases, the tumor was not excised, showing regression in 1 case and lack of significant growth in the other.

Table 2 Summary of molecular results
Figure 2
figure 2

Representative histological features of nodular fasciitis cases seen in this series. (a) Vaguely circumscribed tumor involving skeletal muscle (× 1, H&E). (b) Cellular areas with tissue culture-like pattern and myxoid areas. (c) Myofibroblastic cytomorphology with extravasated red blood cells and lymphocytes. (d) Mitotic figures can be present. (e) Some tumors had prominent myxoid ‘pseudocystic’ areas. (f) Others had thick hyalinized bundles reminiscent of keloid collagen.

Figure 3
figure 3

(a) Case 1. Cross section of nodular fasciitis, which was unusually large, measuring 11.0 cm in longest dimension (specimen sectioned perpendicular to longest dimension). This tumor had classic histological features (b) and was positive for RRBP1USP6 fusion transcripts. The patient had no evidence of disease at 33 months’ follow-up. (c) Case 12. Low power of a 3.0 cm tumor involving the tendons of the finger with histological features of nodular fasciitis (d). This tumor was found to harbor SPARCUSP6 fusion transcripts. The patient had no evidence of disease at 7 months’ follow-up.

USP6 FISH

FISH was successfully performed in 20 cases (Table 2), yielding positive results for USP 6 locus rearrangement in all cases. Case 1 showed a variant pattern consisting of two copies of fused signals and one extra orange signal, suggestive of copy gain with an unbalanced translocation and retention of the 3′-portion of USP6. For the remainder of cases, testing was not performed or was not successful due to insufficient tissue or inadequate quality due to pre-analytical variables.

RT-PCR for MYH9USP6 Fusion Transcript

RT-PCR was successfully performed in 20 cases and fusion transcripts were identified in an overlapping set of 13/20 samples (Table 2). Five cases showed only fusion type 1, while 8 cases showed the presence of both the type 1 and type 2 transcripts, likely as a result of alternative splicing.16, 22 All RT-PCR-positive cases were positive by FISH, when results were available. RT-PCR was also positive in 5 cases in which there was insufficient or unavailable tissue for FISH. In contrast, 2 cases were insufficient for RT-PCR testing, but yielded positive results by FISH. Of the 7 negative RT-PCR cases, 6 were positive by FISH, while FISH was not performed in the remaining case.

Targeted RNA Sequencing by Anchored Multiplex PCR

Paired-end sequencing of target-enriched RNA libraries prepared from eight nodular fasciitis cases using the FusionPlex Sarcoma kit yielded an average of 1.07 × 106 paired-end reads and 28 980 unique RNA reads per sample. Analysis of FASTQ files using the Archer Analysis pipeline for fusion gene detection revealed USP6 fusions in eight of eight (100%) cases, including a control case that was also positive by USP6 FISH and MYH9USP6 RT-PCR. In the remaining seven cases, USP6 fusions were detected with novel 5′-gene partners for nodular fasciitis, including RRBP1 (NM_001042576.1), CALU (NM_001219.4), CTNNB1 (NM_001098210.1), MIR22HG (NR_028502.1), SPARC (NM_003118.3), THBS2 (NM_003247.3), and COL6A2 (NM_001849.3). FusionPlex data are summarized in Table 3. All novel fusions were confirmed using RT-PCR, followed by Sanger sequencing. RT-PCR and FusionPlex data were re-analyzed using public databases and the fusion transcripts identified in each case are summarized in Table 4.

Table 3 Summary of FusionPlex results
Table 4 Fusion types

Analysis of the novel fusion transcripts reveals that they are structurally identical to the MYH9–USP6 fusion (Figure 4). Each case shows fusion of the 5′-UTR of the upstream gene to either exon 1 (5′-UTR) or exon 2 (one base upstream of the ATG translation initiation codon) of USP6. Some cases show multiple fusion types, which show the same USP6 fusion sites as the type 1 and type 2 MYH9 fusions (Table 4). The fusions result in promoter swap, such that the promoter and associated regulatory sequences of the 5′-partner gene are fused to the complete coding region of USP6. Activation of the 5′-gene promoter should therefore drive USP6 expression. In the FusionPlex assay, each of the fusion genes was supported by moderate-to-high unique split read counts (range: 92–1220 reads), with the fusion-to-wild-type fraction ranging from 20 to 100%, suggestive of elevated USP6 expression from the rearranged loci.

Figure 4
figure 4

Schematic of fusion transcripts. The top pane shows the two MYH9USP6 fusion types, both of which include the entire coding sequence of USP6 (transcription start site indicated by an arrow). The TBC and USP (UBP) domains of the USP6 protein are also indicated. In the remaining panes, the other novel fusion transcripts identified in this study are shown. Case 12 shows the inclusion of a cryptic exon from CALU, while cases 13, 16, and 21 show the presence of two fusion transcripts, with identical USP6 fusion sites as those seen in MYH9 fusions.

Seven of the eight cases, including the MYH9–USP6 fusion, show inter-chromosomal rearrangements, with one case (case 14) characterized by an intra-chromosomal rearrangement between MIR22HG and USP6, which are located 3.4 MB apart on chromosome 17. The MIR22HG–USP6 gene fusion within case 14 is the only fusion identified in our study that occurs between a non-coding RNA gene and USP6. Furthermore, an additional case (case 12) has an insert within the fusion transcript, which maps to a portion of intron 1 of the 5′-gene (CALU). Case 1, which was 11 cm in greatest dimension, showed RRBP1USP6 fusion, confirming the diagnosis of nodular fasciitis. No significant association was identified between the fusion variants and either clinical behavior or histologic features of the tumors.

Discussion

USP6 is a hominoid-specific gene primarily expressed in testicular tissue that is located on chromosome 17p13 and is part of a family of de-ubiquitinating enzymes.23 Rearrangements involving the USP6 gene were first identified and characterized in aneurysmal bone cysts by Oliveira et al24, 25 Subsequently, multiple fusion partners for USP6 were discovered in these tumors.26 Similar to nodular fasciitis, aneurysmal bone cyst is a benign tumor that grows rapidly. It was demonstrated that these tumors arise as a result of a promoter-swapping mechanism that drives transcriptional upregulation of USP6 (refs 26, 27) (a mechanism of transformation seen in many tumor types). Subsequently, MYH9– USP6 fusion was identified in nodular fasciitis and it was shown that increased expression of USP6 could induce formation of a tumor histologically and clinically similar to nodular fasciitis in xenografts.16

In our study, FISH, RT-PCR, and targeted RNA sequencing assays were developed and utilized for the detection of USP6 rearrangement in nodular fasciitis. For clinical laboratories, several factors must be considered when implementing various methodologies to detect genomic events associated with tumors. For the detection of fusion transcripts, RT-PCR can be an ideal method when there are a limited number of fusion variants. However, RNA extracted from formalin-fixed, paraffin-embedded tissue is often fragmented to <300 bases in length and this fragmentation increases with storage time28 and more harsh conditions. Assays that are not appropriately designed may result in false negatives.

On the basis of the previously described fusion types,16 we designed an RT-PCR assay to detect both types 1 and 2. Several cases in our study harbored both fusion transcripts in the same sample. Alternative splicing has been well documented in fusion genes associated with bone and soft tissue tumors,22, 29, 30 and this likely explains the findings in our cases and in similar cases in the literature16 (Figure 1). Such an assay is highly specific, but it cannot detect other fusion partners. Moreover, if there is alternative splicing, amplicons that are larger or have different fusion sites may not be detected.

A break-apart FISH assay is especially useful when there is a single gene fused with a variety of partner genes, as in the case of aneurysmal bone cyst or, now, nodular fasciitis. USP6 rearrangement has also been identified in a variety of other tumors, including cellular fibroma of tendon sheath31 and giant cell reparative granulomas of the hands and feet (a subset of aneurysmal bone cyst),32 allowing for the implementation of a single assay across these diseases. Occasionally, unusual patterns may be identified and it can be difficult to determine the clinical significance. The pattern identified in case 1 was also described in one of six cases analyzed by Papp et al,33 consisting of two fused signals and one extra 3′-signal. In our experience, normal cases may show split signals in a low percentage of cells. When interpreting results, it is important to adhere to strict criteria regarding number of cells scored, distance between the orange and green signals, and validated cutoff for percentage of split cells in positive cases.

The discrepancy between FISH and RT-PCR results in our study suggested the possibility of alterative fusion types or partner genes. Alternative fusion partners have already been described in aneurysmal bone cyst and Guo et al34 identified PPP6R3USP6 amplification in a case they termed ‘malignant nodular fasciitis.’ Anchored multiplex PCR is a powerful method for identifying alternative fusion partners, although it is not completely unbiased, as one partner gene must be targeted. During the composition of this manuscript, several other groups published studies using the Archer FusionPlex Sarcoma Panel to characterize USP6 rearrangement partners in aneurysmal bone cyst35 and cellular fibroma of tendon sheath.31

Of 26 total positive cases in our study of nodular fasciitis, 7 showed novel fusion partners, expanding the spectrum of genes known to be fused with USP6 in this tumor. In addition to encoding a ubiquitin-specific protease, the USP6 protein also contains a TBC (Tre-2/Bub2/Cdc16) domain implicated in GTPase regulation and trafficking.23 The two MYH9USP6 fusion types show retention of both the TBC and UBP (or USP, ubiquitin-specific peptidase) domains.20 Similar to MYH9 fusion, some of the novel fusions showed alternative splicing, with the simultaneous presence of both fusion transcripts (as detected by next-generation sequencing and/or RT-PCR; Table 4). In all cases, the entire coding sequence of USP6 is retained. USP6 functions in remodeling of the cytoskeleton and extracellular matrix, inflammatory response, cell signaling, cellular trafficking, and protein turnover,23, 27, 36, 37, 38, 39, 40, 41, 42, 43, 44 but its specific role in tumorigenesis is not entirely clear. More recently, Quick et al45 reported that Jak1–STAT3 signaling has an essential role in USP6-related tumorigenesis, while Madan et al46 suggested a role for Wnt signaling.

In the majority of nodular fasciitis cases, USP6 fuses to MYH9, which is located on chromosome 22q12.3-q13 and encodes for a non-muscle myosin protein involved in cytokinesis, cell motility, maintenance of cell shape, and disassembly of the actin network in crawling cells.16 In the body, it is normally present at low levels, except in fibroblasts, endothelial cells, macrophages, leukocytes, and certain renal cells. Mutations have been associated with macrothrombocytopenias,47 and MYH9 has been reported to be fused to ALK in a case of anaplastic large cell lymphoma.48

We identified several novel fusion partners for USP6 in nodular fasciitis. COL6A2 encodes for the collagen type VI alpha 2 chain. Mutations in COL6A2 are associated with inherited muscular dystrophies.49 Collagen VI is enriched in the pericellular matrix of tendon fibroblasts and has an important role in tendon repair.49

THBS2 encodes for a member of the thrombospondin family of extracellular proteins, which function in cell-to-cell and cell-to-matrix communication.50 Thrombospondin-2 (TBS-2) has also been shown to inhibit angiogenesis.50 In Ewing sarcoma, the EWSR1FLI fusion gene downregulates TBS-2.51

Case 12 showed CALUUSP6 fusion, with the fusion transcript containing an insertion that maps to CALU intron 1. The inclusion of a cryptic exonic sequence in the fusion transcript occurs commonly in other tumors, and we previously showed this is a recurrent event in epithelioid hemangioendotheliomas.22 This may be related to alterations in splice sites or splicing enhancers and silencers as a result of chromosomal translocation.22 Calumenin, encoded by CALU, is a calcium-binding protein belonging to the CREC protein family.52 Different isoforms are localized to different parts of the secretory pathway, including the endoplasmic reticulum, Golgi apparatus, and extracellular medium. Calumenin is thought to have a role in inhibiting tumor metastasis, and its expression levels have been reported to be altered in cancer of the head and neck, endometrium, liver, pancreas, lung, and colon.53 CALU expression was demonstrated to be increased in the tumoral stroma (cancer-associated fibroblasts) and epithelium of colorectal cancer samples.54

Activating mutations in CTNNB1 have been associated with desmoid fibromatosis,55 indicating a role in fibroblastic tumors. CTNNB1 encodes beta-catenin, which interacts with APC and has a well-characterized role in the Wnt signaling pathway in most cell types and cellular adhesion in epithelial cells. Furthermore, CTNNB1PLAG1 fusion in pleomorphic adenomas of the salivary gland has been shown to result in activation of PLAG1 by a promoter-swapping mechanism.56, 57 Of note, CTNNB1USP6 fusion was recently identified in a single case of aneurysmal bone cyst,35 another member of the biological spectrum of USP6-induced neoplasms.20

MIR22HG is a long non-coding RNA that is downregulated in lung adenocarcinoma58 and has been shown to be a responder to chemical stress,59 microgravity,60 and hypoxia.61 It serves as host gene for miR-22, which functions as a tumor suppressor by post-transcriptional regulation of p21 and represses cancer progression by inducing cellular senescence.60

SPARC encodes a cysteine-rich acidic matrix-associated protein. It is a member of a family of proteins that modulate interactions between cells and their environment by regulating growth factor signaling and extracellular matrix assembly and deposition.62 SPARC functions as both a secreted glycoprotein and an intracellular and membrane-associated protein that regulates cellular apoptotic pathways.62 It shows increased expression in epithelial cells from tissues with a high turnover rate during tissue injury and inflammation, as well as during abnormal tissue growth associated with neoplasia.62 Interestingly, this fusion was found in a case occurring in the tendon of the hand (case 15). The histological features were in keeping with nodular fasciitis, but given the location, a differential diagnosis of cellular fibroma could also be entertained. Recently, Carter et al31 described USP6 gene rearrangements in six of nine cellular fibromas of tendon sheath by FISH. No MYH9 rearrangements or MYH9USP6 and CDH11USP6 fusion transcripts were detected by RT-PCR assays. Given the overlapping features, they postulate that some cellular fibromas of tendon sheath may in fact be tenosynovial nodular fasciitis. Our case also highlights the similarities between these entities, both histologically and molecularly, and now with a fusion partner of SPARC.

All of the fusion partners for USP6 likely provide promoters that lead to increased transcription over what would be seen if the USP6 promoter was left in place. Some of these fused promoters could represent genes that are activated specifically within the cellular context of the myofibroblasts that comprise nodular fasciitis—a conjecture that requires further investigation.

To our knowledge, we also report the largest nodular fasciitis case with molecular confirmation in the literature. Although most cases are small, tumors up to 9 cm have been previously described.63, 64 Case 1 in our study was 11 cm and showed an unusual FISH pattern with unbalanced translocation and RRBP1USP6 fusion by next-generation sequencing. The association between the phenotype (ie, large size) and fusion variant is not certain; however, the patient was noted to be alive with no evidence of disease (recurrences or metastases) 10 years post excision. Ribosome-binding protein 1 (RRBP1) encodes for a coiled-coil protein that has a role in microtubule binding and in the interaction between the endoplasmic reticulum and ribosomes.65 RRBP1 is highly expressed in cells such as fibroblasts and is necessary for secretory activity.66 Overexpression of RRBP1 is associated with poor prognosis in breast and colorectal cancer, and RRBP1ALK fusion has been identified in several cases of epithelioid inflammatory myofibroblastic sarcoma.65

As previously mentioned, Guo et al34 described a patient who had multiply recurrent nodular fasciitis over 10 years before developing a metastasis.34 This is the first reported case of nodular fasciitis with malignant behavior, and the tumor was found to harbor a novel fusion transcript PPP6R3USP6. This fusion was not seen in our series, and no differences were seen in the behavior and histological features in the tumors of various fusion transcripts and types.

In summary, we identified seven novel fusion partners for USP6 in nodular fasciitis, highlighting the importance of USP6 expression and promoter-swapping fusions in the etiology of this neoplasm. The partner genes normally have a role in the interaction between the cell and extracellular matrix, inflammation, myo/fibroblastic activity, or stress response. The association of the coding sequence of USP6 with the promoter of these genes results in increased expression of USP6 within the milieu that gives rise to nodular fasciitis. This may account for the histologic features of inflammation and myofibroblastic proliferation associated with these tumors. Molecular testing for USP6 rearrangement can be a useful ancillary tool, though pathologists need to be cognizant of the limitations of each technique to detect multiple partners. Recently, a similar study examining aneurysmal bone cysts by anchored multiplex PCR expanded the spectrum of associated partner genes35 and, together, this provides additional evidence for the neoplastic nature of USP6-associated tumors. Further studies must be performed to elucidate the biology and phenomenon of transient neoplasia inherent to nodular fasciitis.