Low‐grade myxoid spindle cell neoplasm with novel gene fusions involving MAP3K3 and MAP3K8 kinases: a report of two cases

Herein, we report novel gene fusions SCD::MAP3K3 and MAP3K8::PRLR, identified in two morphologically low-grade myxoid predominant hypocellular spindle cell neoplasms. Gene fusions involving MAP3K3 and MAP3K8 genes have been previously identified in spitz melanocytic neoplasms, and mesotheliomas1-5; however, they have not been identified in any soft tissue tumours so far. Case 1, a 71-year-old man presented with a slow-growing right thigh mass. Magnetic resonance imaging (MRI) demonstrated an 8.2 cm homogeneous lobulated intramuscular lesion of the right medial thigh, which was hypointense on T1 and extremely hyperintense on T2 (Figure 1A,B). A needle biopsy followed by subsequent resection was performed. The histologic sections revealed proliferation of bland spindle cells in a prominent background of myxoid stroma with scattered focally hyalinized thin-walled blood vessels. The majority of spindle cells showed elongated nuclei with ill-defined eosinophilic cytoplasm and perinuclear vacuolation (Figure 1C–E). Scattered lipoblast-like cells were also evident. The lesional cells were patchy positive for desmin, SMA, and CD34 (Figure 1F), while negative for S100, SOX10, CD30, MUC4, STAT6, ERG, ALK, myogenin, cytokeratin AE1/AE3, CK5/6, MyoD1, EMA, caldesmon, smooth muscle heavy chain, HMB45, and Melan A (Figure S1). RB showed retained nuclear expression. Fluorescence in situ hybridization (FISH) studies for MDM2 gene amplification, and DDIT3 and PLAG1 gene rearrangements were negative. The neoplasm had a low-grade overall in morphology, with no evidence of mitotic activity, necrosis, or significant cytologic atypia. Next-generation sequencing (NGS) studies by Oncomine Comprehensive assay (a 161-gene panel) on Case 1 were negative for any mutation or copy number alterations, while whole transcriptomic sequencing (RNAseq) identified a novel in-frame gene fusion SCD::MAP3K3 with 5′ breakpoint at exon 5 (chr10:102116521, NM_005063) and 3′ breakpoint at exon 2 (chr17:61710041, NM_203351) (hg19) (Figure 3A). Case 2, a 63-year-old female with a past medical history of malignant melanoma in situ on the left cheek, presented with a slow-growing left upper neck mass with itching, pain, and mild erythema of the overlying skin. The patient has had this mass for the last 5 years. Computed tomography (CT) imaging of the neck with contrast enhancement demonstrated a 4.0 × 3.0 × 2.3 cm subcutaneous mass in the left submandibular region, superficial to the platysma, which had grown from 1 to 4 cm. There was a mild contrast enhancement. A surgical resection was performed. The histologic sections demonstrated a well-circumscribed, relatively hypocellular neoplasm composed of bland elongated spindle to stellate shaped cells with scattered collagen fibres, focally hyalinized thin-walled vessels (Figure 2), and variable background chronic inflammation. No atypia, increased mitotic activity, or necrosis was identified. No perinuclear vacuolation were identified in this case. The lesional cells were focally positive for CD34, SMA, and CD68, while negative for S-100, SOX10, MUC4, desmin, cytokeratin AE1/AE3, p40, STAT6, ERG, myogenin, Melan-A (Figure 2E,F). RB showed retained nuclear expression. MDM2 FISH studies were negative. RNAseq studies identified novel in-frame gene fusion MAP3K8::PRLR with 5′ breakpoint at exon 8 (chr10:30748430, NM_005204) and 3′ breakpoint at exon 8 (chr5:35068980, NM_000949) (hg19) (Figure 3B). Based on the morphologic features and immunohistochemical stains, a differential diagnoses of intramuscular myxoma with cellular features, adipocytic tumours, angiofibroma of soft tissue, or low-grade myxofibrosarcoma were considered. Although both cases showed a predominant myxoid pattern with overall low-grade histologic features, which is reminiscent of myxoma, there are several interesting morphologic findings in these two unique cases, which do not quite fit with the diagnosis of myxoma. First, the tumour showed relatively increased cellularity, at least focally, in contrast to the expected paucicellularity seen in myxomas. Second, a major subset of spindle cells showed elongated nuclei with perinuclear vacuolization (Case 1), which is unusual for myxoma. Third, a somewhat increased or prominent branching vascular pattern was noted in the myxoid stroma, unlike the characteristically sparse vasculature usually described in myxomas. That being said, although GNAS mutation was negative by NGS (Case 1), the possibility of these cases being an unusual morphologic variant of GNAS-negative cellular myxoma with novel gene fusion remains a consideration. We considered adipocytic tumours, including atypical spindle cell lipomatous tumour and lipoblastoma-like tumour, given the perinuclear vacuolization. However, the lack of cytologic atypia and the absence of plexiform-like vasculature make these diagnoses unlikely. Angiofibroma of soft tissue was considered in the differential diagnosis due to the increased branching vascular pattern noted within the myxoid stroma of both cases; however, the absence of NCOA2 or the rarely reported GAB1::ABL1 gene fusions characteristic of angiofibroma of soft tissue argued against such a diagnosis.6 Additionally, the bland nature of the tumour cells basically ruled out a low-grade myxofibrosarcoma. Other entities that were initially considered in the differential diagnosis based on the tumour morphology, but were essentially excluded via ancillary studies included dedifferentiated liposarcoma, myxoid liposarcoma, peripheral nerve sheath tumours, low-grade fibromyxoid sarcoma, and myxoid variant of solitary fibrous tumour. Overall, both lesions showed circumscribed borders, with low-grade morphology and no cytologic atypia, and overall, did not quite fit in any of the current diagnostic entities. Gene fusions involving MAP3K3 and MAP3K8 lead to constitutive activation of the MAP-kinase pathway and have been previously identified in a subset of melanocytic Spitz tumours, and rarely in mesothelioma. In the described MAP3K3 fusion-positive mesothelioma, the fusion product maintained both the PB1 and serine/threonine kinase catalytic domains, as was seen in our case, whereas in the reported spitzoid neoplasm, only the serine/threonine kinase domain was maintained. The PB1 domain is needed for oligomerization and the formation of macromolecular complexes for cellular signalling. Additionally, MAP3K3 was the 3′ partner in both cases.2, 4 Pathogenic MAP3K8 fusions are more well-described, most commonly being found in a melanocytic lesion.1, 2 The fusions predominantly involve MAP3K8 as the 5′ partner, with a breakpoint at exon 8, maintaining the serine/threonine kinase domain, as was seen in our case.1, 2 To further evaluate the functional impact of the gene fusions, we examined the expression levels of MAP3K3 and MAP3K8 compared to other soft-tissue tumours; Cases 1 & 2 displayed elevated expression in MAP3K3 and MAP3K8, respectively (Figure S2). Adequate follow-up is not available, since both cases were very recently excised. Additionally, the exact nature or behaviour of these neoplasms is uncertain. Additional studies of more cases are required to make a definite classification and prognostication. Azfar Neyaz: prepared the draft and figures, Mana Mohebnasab: provided molecular data and edited the article, Elan Hahn: provided molecular data and edited the article, Joel Rosenbaum: performed the bioinformatic analysis, Da Gama Lobo Lucas: provided radiologic data and figures, Karunamurthy: provided molecular data, Ivy John: reviewed cases, Kurt R. Weiss: reviewed Case 1, Karen Schoedel: reviewed Case 1, Simion I Chiosea: provided Case 2, and edited the final article, Rana Naous: designed the study, provided Case 1, reviewed cases and edited the final article, All authors approved the final article. No funding was received for this study by any author. No competing interest for any author. The data that support the findings of this study are available from the corresponding author upon reasonable request. Figure S1. Additional immunohistochemical stains of case 1: SMA, and desmin showed patchy positive in the lesional cells, while S100 was negative. Figure S2. Normalized RNA expression levels of our two cases compared to all historical soft tissue tumours having undergone whole transcriptome RNA sequencing at our institution. Case 1 is indicated in blue and Case 2 is indicated in orange. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.