Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2022 Dec 12;82(4):596–607. doi: 10.1111/his.14842

NTRK‐rearranged spindle cell neoplasms are ubiquitous tumours of myofibroblastic lineage with a distinct methylation class

Arnault Tauziède‐Espariat 1,2,✉, Mathilde Duchesne 3, Jessica Baud 4, Mégane Le Quang 4,5, Dorian Bochaton 6, Rihab Azmani 7, Sabrina Croce 4, Isabelle Hostein 4, Carole Kesrouani 8, Delphine Guillemot 6, Gaëlle Pierron 6,9, Franck Bourdeaut 10,11, Liesbeth Cardoen 12, Lauren Hasty 1, Emmanuèle Lechapt 1, Alice Métais 1, Fabrice Chrétien 1, Stéphanie Puget 13, Pascale Varlet 1,2, François Le Loarer 3,5,14
PMCID: PMC10108022  PMID: 36413100

Abstract

Aims

NTRK gene fusions have been described in a wide variety of central nervous system (CNS) and soft tissue tumours, including the provisional tumour type ‘spindle cell neoplasm, NTRK‐rearranged’ (SCN–NTRK), added to the 2020 World Health Organisation Classification of Soft Tissue Tumours. Because of histopathological and molecular overlaps with other soft tissue entities, controversy remains concerning the lineage and terminology of SCN–NTRK.

Methods and results

This study included 16 mesenchymal tumours displaying kinase gene fusions (NTRK fusions and one MET fusion) initially diagnosed as infantile fibrosarcomas (IFS), SCN–NTRK and adult‐type fibrosarcomas from the soft tissue, viscera and CNS. We used immunohistochemistry, DNA methylation profiling, whole RNA‐sequencing and ultrastructural analysis to characterise them. Unsupervised t‐distributed stochastic neighbour embedding analysis showed that 11 cases (two CNS tumours and nine extra‐CNS) formed a unique and new methylation cluster, while all tumours but one, initially diagnosed as IFS, clustered in a distinct methylation class. All the tumours except one formed a single cluster within the hierarchical clustering of whole RNA‐sequencing data. Tumours from the novel methylation class co‐expressed CD34 and S100, had variable histopathological grades and frequently displayed a CDKN2A deletion. Ultrastructural analyses evidenced a myofibroblastic differentiation.

Conclusions

Our findings confirm that SCN‐NTRK share similar features in adults and children and in all locations combine an infiltrative pattern, distinct epigenetic and transcriptomic profiles, and ultrastructural evidence of a myofibroblastic lineage. Further studies may support the use of new terminology to better describe their myofibroblastic nature.

Keywords: central nervous system, DNA methylation profile, myofibroblastic, NTRK, soft tissue

graphic file with name HIS-82-596-g006.jpg

Introduction

NTRK1/2/3 gene fusions are involved in a wide variety of tumour types in different organs. 1 There is a striking diversity in soft tissue, as NTRK fusions have been reported in inflammatory myofibroblastic tumours (IMT), 2 infantile fibrosarcomas (IFS) 3 and the ‘spindle cell neoplasm with NTRK fusion’ (SCN–NTRK). 4 This provisional category, included in the 2020 World Health Organisation (WHO) Classification of Soft Tissue Tumours, is associated with a wide morphological spectrum [IFS, IMT, lipofibromatosis‐like neural tumour and malignant peripheral nerve sheath tumour (MPNST)]. 4 SCN–NTRK harbour mainly NTRK1/2/3 gene fusions (75 reported cases to date). 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 Alternative fusions (implicating ALK, BRAF, MET, RAF1, RET and ROS1 genes) have also been reported, and their exact relationship to SCN–NTRK remains unknown. 8 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 The histological lineage of these ubiquitous tumours [mainly located in soft tissue, but also in viscera and even in the central nervous system (CNS)] has not been elucidated to date. 11 , 20 While the WHO Classification of soft tissue tumours has classified them within tumours of uncertain differentiation, SCN–NTRK were listed within the myofibroblastic/fibroblastic category in the 2022 WHO classification of paediatric tumours. 26 , 27 Recently, the 2021 WHO classification of CNS tumours has been enriched with a wide variety of new tumour types identified thanks to DNA‐methylation profiling, 28 , 29 , 30 representing a combination of both somatically acquired DNA methylation changes and a signature reflecting the cell of origin. 31 , 32 A DNA‐methylation based classifier has been created for sarcomas but does not yet cover SCN–NTRK. 33 To identify the lineage of SCN–NTRK and further characterise their histopathological/molecular features, we analysed a cohort of NTRK‐fused mesenchymal tumours using histopathology, ultrastructural analyses, whole RNA‐sequencing and DNA‐methylation profiling.

Materials and methods

Sample Collection And Clinical Data

Tumour samples were provided by the consultation archive database (1982–2021) from the Department of Neuropathology at Sainte‐Anne Hospital, NetSarc+ and Institut Bergonie. Fifteen mesenchymal tumours with a proven NTRK fusion were selected for this study. Additionally, one CNS case suspected to be SCN–NTRK harbouring a TFG::MET fusion was also included. Patient characteristics and clinical data were retrieved from hospital records and included sex, age at presentation and medical history.

Parents/guardians gave written informed consent for the retrospective analysis of clinical data according to the Institutional Review Board and before inclusion into ongoing protocols. All soft tissue cases are recorded in the database of the French expert sarcoma network (RRePS/Netsarc+), which is approved by the National Committee for Protection of Personal Data (CNIL, no. 910390). The study was also approved by the research board of Institut Bergonie. The analyses of tissue were performed in accordance with local ethics regulations. Parents/guardians gave written informed consent for the retrospective analysis of clinical data according to the Institutional Review Board and before inclusion into ongoing protocols. The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Histopathological Review And Immunohistochemistry

The central pathology review was performed conjointly by two neuropathologists (ATE and PV) and soft tissue pathologists (FLL, MLQ). A representative paraffin block was selected for each case. Unstained 3‐μm‐thick slides of formalin‐fixed, paraffin‐embedded (FFPE) tissues were obtained and submitted for immunostaining (the list of antibodies is detailed in Supporting information, Table S1). External positive and negative controls were used for all antibodies.

Whole RNA‐sequencing And Bioinformatic Analysis

Total RNA was extracted from the FFPE blocks and library prepared using the TruSeq RNA Access Library Prep Kit (Illumina, San Diego, CA, USA) and sequenced with an Illumina NextSeq 500 automat (Illumina), as previously described. 34 Gene expression was measured using counts of sequencing reads with the software HtSeq version 0.6.0. 35 Gene count data were normalised using the VOOM method. 36 Unsupervised clustering was performed using agglomerative hierarchical clustering with distance criteria (1‐Pearson_correlation) and linkage criteria using the function hclust available in R. 37 Consensus clustering was performed with the package ConsensusClusteringPlus from BioConductor. 38 The fusion and cluster analyses were performed as previously described. 34 To achieve clustering, we used a set of control samples that encompassed similar histotypes to those included in the methylation analysis: undifferentiated pleomorphic sarcomas (n = 12), MPNST (four), atypical neurofibromatous neoplasms of uncertain biological potential (two), meningiomas (five), dermatofibrosarcoma protuberans (nine), infantile hemispheric glioma with BRAF fusion (one), IFS (one), solitary fibrous tumour (SFT) (three), low‐grade fibromyxoid sarcomas (four), fibromatosis with CTNNB1 mutations (seven) and angiomatoid fibrous histiocytomas (AFH) (four) (Supporting information, Table S1).

DNA Methylation Array Processing And Copy Number Profiling

Genomic DNA was extracted from fresh‐frozen or FFPE tissue samples. DNA methylation profiling for all samples was performed using the Infinium MethylationEPIC (850 k) BeadChip (Illumina) or Infinium HumanMethylation450 (450 k) BeadChip array (Illumina), as previously described. 28 The resulting matrix was used as input for a t‐SNE analysis (t‐distributed stochastic neighbour embedding; Rtsne package version 0.15) with a perplexity set to 10, maximum iteration to 500, resulting in two more easily interpretable dimensions (using a control set of 241 well‐characterised reference tumours representing a wide variety of tumour types from the Brain Tumour and Sarcoma Classifiers of Heidelberg). 28 , 33

Ultrastructural Analyses

A representative section was first selected for each case from FFPE tissues stained with haematoxylin phloxin saffron (HPS). Then, tissues were deparaffinised and fixed for 1 h in glutaraldehyde. After the dehydration processing, the tissues were embedded in Epon. Semi‐thin sections (1‐μm‐thick slides) were stained with toluidine blue. Ultrathin sections (90 nm) were stained with lead citrate and uranyl acetate, then observed under an electronic microscope (JEOL JEM 1400 Flash). Analysis was performed at the Pathology Department of Limoges University Hospital by one neuropathologist (MD).

Results

Histopathological Findings

Histopathological results are detailed in Supporting information, Table S1. Considering the age of patients, the central histopathological review classified the 16 tumours as SCN–NTRK (n = 8), IFS (n = 5) and adult‐type fibrosarcomas (n = 3). SCN–NTRK were highly infiltrative, whatever the organ concerned (Figure 1A,B). The tumours were composed of spindle cells arranged in intersecting fascicles, with a variable cellular density (low to high, Figure 1C,D and Supporting information, Table S1). The tumour cells presented a pale scant cytoplasm with indistinct cell borders and fusiform nuclei with inconspicuous nucleoli (Figure 1C,D). Focally, we observed nuclear atypia with hyperchromasia and pleomorphism and an elevated mitotic index (up to 37 mitoses per 1.7 mm2) in some cases (Figure 1E). Necrosis was only observed in one recurrent case (Figure 1F). All cases presented areas of fibrous stroma and five of eight (62%) exhibited a perivascular hyalinisation forming collagen rings (Figure 1G). A myxoid stroma was present in four of eight cases (50%). Three cases (37%) presented inflammatory infiltrates composed as foci of lymphocytes and plasma cells. Immunohistochemical analysis showed that all tumours co‐expressed S100 and CD34 (Figure 1H,I), ranging from focal to diffuse staining. Smooth muscle actin was only focally expressed in two of eight cases (25%). SOX10 was not expressed. The proliferation index (Ki67) ranged from 1 to 50%, with a median of 2.5%. For one case (case 16), paired samples of the first surgery and the recurrence were available (see below for details): the first tumour presented a low cellular density, without necrosis and low mitotic and proliferative indexes, whereas the recurrence exhibited signs of malignancy (necrosis, elevated mitotic and proliferative indices and anisokaryosis).

Figure 1.

Figure 1

Open in a new tab

Histopathological and ultrastructural features of tumours from the new methylation class. (A) Infiltration of skeletal muscle by tumour cells (case 13, HPS). (B) Meningothelial cells (arrowheads) included in the proliferation (case 1, HPS). (C) Spindle cell proliferation associated with an abundant myxoid stroma (case 1, HPS). (D) Another area showing tumour cells densely arranged in fascicles (case 1, HPS). (E) Cellular proliferation of spindle and epithelioid cells showing a brisk mitotic activity (arrowheads) (case 16, HPS). (F) Area of necrosis (case 16). (G) Perivascular collagen rings (case 1). (H) Strong expression of S100 (case 16). (I) Diffuse expression of CD34 (case 16). Scale bars represent 100 μm (A,B) and 50 μm (C–I). HPS, haematoxylin phloxin saffron. [Color figure can be viewed at wileyonlinelibrary.com]

All IFS cases affected infants (all aged less than 1 year) and displayed a fascicular architecture. Tumour cells harboured monotonous medium‐sized nuclei with clear chromatin along with a pale eosinophilic cytoplasm. The accompanying stroma was limited to small fibrous foci in all cases. Large dilated vessels with haemorrhage were seen in three cases (three of five). Tumours expressed CD34 in one of four cases.

Adult‐type fibrosarcomas affected adults with a mean age of 55 years (ages ranged from 48 to 66 years). All cases displayed a fascicular architecture. Tumour cells were arranged in densely cellular fascicles. Large vessels with haemorrhage were present in all cases. Tumour cells harboured ovoid nuclei and a majority of nuclei were monotonous, but anisokaryosis was also present focally in all cases taking the form of hyperchromatism and/or larger multilobulated nuclei. The stroma was heterogeneous and absent in some areas or collagenous in some foci. The mean mitotic index was 10/10 high‐power fields.

Whole RNA‐Sequencing Findings

In eight cases, NTRK1 was fused with the following 5′ partners: TPM3 (four of eight), IRF2BP2 (two of eight) and LMNA and TPR (one case each). NTRK3 was only found fused in 5′ with ETV6 (seven of seven). One case was associated with a TFG::MET fusion. Unsupervised RNA‐sequencing cluster analysis showed that 15 of the 16 cases clustered together, independently of their location and their histopathological subtype (Figure 2). Based on this clustering, these 15 cases were subdivided into three subgroups: one of four tumours clustering with IFS; two of two tumours with a TPM3::NTRK1 fusion and three of nine tumours harbouring NTRK1/3 fusions (eight) and the case MET‐fused. The last case (no. 15), presenting an ETV6::NTRK3 fusion, segregated with an ALK‐fused IMT (Figure 2).

Figure 2.

Figure 2

Open in a new tab

Unsupervised RNA‐sequencing clustering analysis. The control cohort included: ALK‐fused inflammatory myofibroblastic tumours (IMT), angiomatoid fibrous histiocytomas (AFH), desmoid‐type fibromatosis, undifferentiated pleomorphic sarcomas (UPS), dermatofibrosarcoma protuberans (DFSP), low‐grade fibromyxoid sarcomas (LGFMS), meningiomas, solitary fibrous tumours (SFT), malignant peripheral nerve sheath tumours (MPNST) and atypical neurofibromatous neoplasms of uncertain biological potential (ANNUBP). [Color figure can be viewed at wileyonlinelibrary.com]

DNA Methylation Findings

Using DNA methylation‐based brain tumour and sarcoma classifiers (latest versions: 12.5/12.2, which include more than 96 000 samples), nine of 16 were significantly classified (calibrated scores for DNA methylation class ≥ 0.9) as sarcoma (MPNST‐like) (six) and IFS (three). Then, on t‐SNE analysis performed in order to visualise actual distances and similarities between the different tumours, our cohort segregated into three different groups (Figure 3A). The largest fraction (11 of 16, 69%) of tumours formed a novel cluster clearly distinct from the reference entities. Four tumours (four of 16, 25%) were classified within the IFS cluster. The last tumour (case 15) was classified among AFH (one of 16, 6%). In light of the copy number data, we identified a CDKN2A deletion in six of 11 (55%) tumours belonging the new methylation class (MC). Interestingly, no other tumour from our cohort presented this alteration. This deletion was confirmed by a FISH (fluorescence in‐situ hybridisation) analysis of the CDKN2A locus showing a homozygous deletion (Figure 3B). No NF1 deletion was seen (none of 16).

Figure 3.

Figure 3

Open in a new tab

Epigenetic and copy number variation features. (A) t‐Distributed stochastic neighbour embedding (t‐SNE) analysis of DNA methylation profiles from the 16 investigated tumours alongside selected reference samples. Reference DNA methylation diagnoses: five angiomatoid fibrous histiocytomas, five atypical fibroxanthomas, 36 dermatofibrosarcomas protuberans, 13 desmoid‐type fibromatoses, 14 infantile fibrosarcomas, nine infantile hemispheric gliomas, seven inflammatory myofibroblastic tumours, eight low‐grade fibromyxoid sarcomas, 25 malignant peripheral nerve sheath tumours, 67 meningiomas, seven MPNST‐like, nine pleomorphic dermal sarcomas, six sclerosing epithelioid sarcomas, 26 solitary fibrous tumours, nine solitary fibrous tumours/haemangiopericytomas. (B) Illustrative copy number variation of case 1 showing a 9p deletion confirmed by FISH analysis of CDKN2A locus showing a homozygous deletion (no orange signal with green signals representing the control centromeric of the chromosome 9). FISH, fluorescence in‐situ hybridisation. [Color figure can be viewed at wileyonlinelibrary.com]

Integrated Diagnoses

Accordingly, eight cases, which included adult cases and one infant (cases 1, 4, 6, 8, 11, 13, 14 and 16), were classified as SCN–NTRK by histopathology and clustered together using DNA‐methylation profiling and RNA‐sequencing. They presented as infiltrative spindle cells with a coexpression of S100 protein and CD34 and harboured NTRK1 (seven) or a MET (one) fusion. Accordingly, four cases, all infants (cases 2, 3, 7 and 10) were classified as IFS by histopathology and clustered together using DNA‐methylation profiling and RNA‐sequencing. They presented typical histopathological features of IFS and harboured the classical ETV6::NTRK3 fusion. The three cases (cases 5, 9 and 12) initially diagnosed as adult‐type fibrosarcomas clustered with the subgroup of SCN–NTRK by DNA‐methylation profiling and RNA‐sequencing despite histopathological features of malignancy. They presented coexpression of S100 and CD34, and were associated with NTRK3 (two) and NTRK1 (one) fusions. The last case (case 15) concerned an infantile tumour classified as IFS by histopathology with the classical ETV6::NTRK3 fusion, but clustered with AFH (by RNA‐sequencing) and IMT (by DNA‐methylation profiling). Based on the WHO classification and because of discrepancies between histopathology, transcriptomic and epigenetic results, no integrated diagnosis was made.

Ultrastructral Findings

Ultrastructural analyses were interpretable in 11 cases (Supporting information, Table S1). All tumours, including those from the new MC, harboured features suggestive of a myofibroblastic differentiation; namely, indentation of nuclei (Figure 4A), presence of abundant myofilament bundles with core densities within the cytoplasm (Figure 4B), either with a subplasmalemmal location or loosely arranged in the cytoplasm, and the rough endoplasmic reticulum (Figure 4A,C). However, their exact location could not be determined due to artefacts related to the FFPE pre‐analytical conditions of the samples. Finally, extracellular collagen was present in the matrix (Figure 4D).

Figure 4.

Figure 4

Open in a new tab

Ultrastructural features of tumours from the new methylation class. (A) Presence of indented nucleus, intracytoplasmic myofilaments with core densities (MF) and rough endoplasmic reticulum (*) in tumour cells. (B) Presence of rough endoplasmic reticulum (*) and intracytoplasmic subplasmalemmal bundles of myofilaments (arrow) whose plasma membrane can be estimated (arrowheads). (C) Intracytoplasmic myofilaments with core densities (MF) and rough endoplasmic reticulum (*) in tumour cells. (D) Presence of collagen fibres.

Correlation With Clinical Data And Patients’ Outcome

The patients’ clinical data are summarised in Table 1. Tumour locations of the new MC were ubiquitous, including the following anatomical sites: soft tissue (six), viscera (three) and CNS (two). The median age at diagnosis was 30.5 years (ranging from 1 to 66) with a female predominance (82%, nine of 11). Outcome data were available for nine of 11 patients (median follow‐up of 21 months); two died of their disease (24 and 49 months after the initial diagnosis). Interestingly, one case (case 16) associated with a TFG::MET fusion had initially benefited from a targeted therapy with crizotinib, until drug toxicity led to the interruption of the treatment (Figure 5). Despite other treatments, the subsequent imaging revealed a local and metastatic progression, and the patient died 49 months after the initial diagnosis.

Table 1.

Clinical and molecular data of cases from our cohort

Case Sex Age (years) Histological diagnosis Location Fusion transcript/RNA‐sequencing cluster DNA methylation classification with the sarcoma classifier v12.2 version (calibrated score) t‐SNE analysis methylation cluster Chr. 9p status FO
2 M 0 Infantile fibrosarcoma Chest wall (soft tissue) ETV6(ex5)‐NTRK3(ex15)/infantile fibrosarcoma Infantile fibrosarcoma (0.99) Infantile fibrosarcoma Balanced ANED (84 months)
3 M 0 Infantile fibrosarcoma Presacral (soft tissue) ETV6(ex5)‐NTRK3(ex14)/infantile fibrosarcoma Infantile fibrosarcoma (0.99) Infantile fibrosarcoma Balanced ANED (14 months)
7 F 1 Infantile fibrosarcoma Paravertebral (soft tissue) ETV6(ex5)‐NTRK3(ex15)/infantile fibrosarcoma Infantile fibrosarcoma (0.99) Infantile fibrosarcoma Balanced NA
10 F 0 Infantile fibrosarcoma Pelvis (soft tissue) ETV6(ex5)‐NTRK3(ex15)/infantile fibrosarcoma Infantile fibrosarcoma (0.85) Infantile fibrosarcoma Balanced ANED (65 months)
1 M 17 Spindle cell neoplasm with NTRK fusion Brain TPM3(ex7)‐NTRK1(ex12)/new cluster Sarcoma (MPNST‐like) (0.72) New methylation class Balanced AWD (13 months)
4 F 29 Spindle cell neoplasm with NTRK fusion Uterus TPM3(e3)‐NTRK1(e10)/new cluster Sarcoma (MPNST‐like) (0.99) New methylation class Balanced ANED (11 months)
5 F 48 Adult type fibrosarcoma Pancreas ETV6(ex4)‐NTRK3(ex13)/new cluster Infantile fibrosarcoma (0.39) New methylation class Deleted ANED (12 months)
6 F 56 Spindle cell neoplasm with NTRK fusion Finger (cutaneous) TPM3(ex9,7,8)‐NTRK1(ex9,10)/new cluster Sarcoma (MPNST‐like) (0.99) New methylation class Balanced ANED (21 months)
8 F 36 Spindle cell neoplasm with NTRK fusion Foot sole (cutaneous) TPR(ex21)‐NTRK1(ex3,2)/new cluster Sarcoma (MPNST‐like) (0.26) New methylation class Balanced ANED (50 months)
9 F 66 Adult type fibrosarcoma Ankle (cutaneous) IRF2BP2(ex2)‐NTRK1(ex9,10) /new cluster Sarcoma (MPNST‐like) (0.99) New methylation class Deleted DOD (24 months)
11 M 27 Spindle cell neoplasm with NTRK fusion Neck (fascia) TPM3(ex8,7)‐NTRK1(ex9,10)/new cluster Sarcoma (MPNST‐like) (0.90) New methylation class Balanced ANED (34 months)
12 F 50 Adult type fibrosarcoma Stomach ETV6(ex4)‐NTRK3(ex13)/new cluster Sarcoma (MPNST‐like) (0.81) New methylation class Deleted AWD (12 months)
13 F 32 Spindle cell neoplasm with NTRK fusion Palm (soft tissue) IRF2BP2(ex1)‐NTRK1(ex9)/new cluster Sarcoma (MPNST‐like) (0.99) New methylation class Deleted NA
14 F 21 Spindle cell neoplasm with NTRK fusion Thorax (soft tissue) LMNA(ex2)‐NTRK1(ex11)/new cluster Sarcoma (MPNST‐like) (0.95) New methylation class Deleted NA
16 F 1 Spindle cell neoplasm with NTRK fusion Meninge TFG(ex4)‐MET(ex15) /new cluster Dermatofibrosarcoma protuberans (0.83) New methylation class Deleted DOD (49 months)
15 F 1 Infantile fibrosarcoma Axillary region (soft tissue) ETV6(ex5)‐NTRK3(ex14)/angiomatoid fibrous histiocytoma Angiomatoid fibrous histiocytoma (0.74) Inflammatory myofibroblastic tumour Balanced ANED (108 months)
Open in a new tab

ANED, Alive with no evidence of disease; AWD, Alive with disease; Chr., Chromosome; DOD, Died of disease; F, Female; FO, Follow‐up; M, Male; MPNST, Malignant peripheral nerve sheath tumour; t‐SNE, t‐distributed stochastic neighbour embedding; NA, Not applicable.

Figure 5.

Figure 5

Open in a new tab

Radiological and histopathological features of case 16. (A) Axial T2‐weighted image showing a voluminous frontotemporal mass with a solid component and central hyperintense areas suggestive of cystic–necrotic changes and without peritumoral oedema. (B) Axial T2‐weighted coronal contrast‐enhanced T1‐weighted MR image showing a local relapse with enhancing nodules (arrows). (C) Sagittal contrast‐enhanced T1‐weighted image of the spinal cord showing a spinal metastasis at the level of the medullary cone (arrowhead). (D) Histopathology of the primary tumour showing a low cellular proliferation composed of spindle cells without atypia (HPS). (E) Histopathology of the recurrent tumour showing a high cellular proliferation with frequent areas of necrosis (HPS). (F) The tumour was composed of spindle cells with marked atypia and several mitoses (arrowheads) (HPS). (G) Time‐line of the tumour progression and treatment choices as shown (LP, local progression; M, metastases). Scale bars represent 250 μm (E) and 50 μm (D,F). HPS, haematoxylin phloxin saffron. [Color figure can be viewed at wileyonlinelibrary.com]

Discussion

As per the last WHO classification of soft issue tumours, the lineage of SCN–NTRK remains unknown. Because of their S100 immunoexpression and spindle cell pattern, a potential neural differentiation was initially suggested. 5 , 6 , 8 , 10 , 13 , 14 , 15 , 21 , 22 , 39 In this study, we show that SCN–NTRK share a common DNA‐methylation profile known to be correlated with a distinct cell of origin. According to t‐SNE analysis, they were distinct from ‘MPNST‐like sarcomas’. ‘MPNST‐like sarcomas’ have only been described in the DNA‐methylation classifier under DKFZ (as para‐spinal tumours occurring in adults aged 50 years, with NF1 alteration and no NTRK fusion, but retained H3K27me3) and are different from NTRK‐fused tumours. 40 Our results are in line with the literature data suggesting a fibroblastic/myofibroblastic differentiation in SCN–NTRK. 41 , 42 , 43 Presence of intercellular collagen and rough endoplasmic reticulum and actin myofilaments were observed in the subplasmalemmal area (in a minority of cases) or arranged in a more loosely attributed pattern (possibly due to the pre‐analytical conditions of FFPE samples and because those tumours may not necessarily retain the characteristics of a mature non‐tumoral myofibroblast). 41 , 42 Similarly to other myofibroblastic tumours, 27 , 43 , 44 , 45 , 46 , 47 myogenic markers (particularly smooth muscle actin and desmin) are heterogeneously expressed in SCN–NTRK. 6 , 9 , 10 , 11 , 13 , 14 , 15 , 18 , 20 , 21 , 23 , 25 SCN–NTRK may be confused with other entities due to the diversity of morphologies. Infantile forms of SCN–NTRK (20% of reported cases 6 , 9 , 10 , 11 , 17 , 20 , 21 , 22 ) may share histopathological and genetic features (ETV6::NTRK3 fusion, as in two current cases, or ALK fusions 5 , 22 ) with IFS and IMT. Reciprocally, few cases of IFS and IMT have been reported with NTRK fusions. 48 , 49 , 50 , 51 Herein, SCN–NTRK form a MC clearly distinct from IMT and IFS, showing the interest of DNA‐methylation profiling in difficult cases. However, a grey zone emerges with one case from our series (case 15) presenting histopathologically and genetically (ETV6::NTRK3 fusion) as an IFS, but did not classify within this MC or as SCN–NTRK when analysed by DNA‐methylation profiling. This case was classified as an AFH, despite being clinically and morphologically typical of IFS and despite harbouring the typical ETV6::NTRK3 fusion of IFS. This result could not be explained and illustrates the limitations of molecular classification that may occasionally further complicate the matter rather than solving it. Additional series are needed to characterise in detail the histomolecular landscape of infantile mesenchymal tumours with NTRK fusions. In doubtful cases, perivascular collagen rings is a common feature of these tumours, 16 reminiscent of a NTRK‐fusion in this morphological setting. Additionally, the presence of a CDKN2A deletion also seems common in SCN–NTRK, as 9p deletions were previously found in 33% of cases (two of six) in a precedent series. 52 The DNA‐methylation profiling correlates more with the cell origin than the tumours’ genotype, as was evidenced in CNS tumours with, for example, tumours sharing MAPK alterations or ROS1 fusions, and were clearly distinct by epigenetic profiles. 28 , 53 The two current CNS cases were clearly distinct from infantile hemispheric gliomas sharing NTRK fusions. Our isolated case with MET fusion clustered among NTRK‐fused tumours, which may suggest the representation of a genetic variant of the same entity or may emerge later with additional cases as a distinct entity. TFG::MET fusion has been previously described in two cases of the soft tissue and the retroperitoneum presenting similar clinical (infants) and histopathological features (spindle cell tumours infiltrating the surrounding tissue with a high mitotic index for one of them). 21 , 22 Further studies are needed to confirm that other kinase fusions (ALK, BRAF, RAF1, ROS1) are part of the same MC of SCN–NTRK. In the CNS, SCN–NTRK is probably underdiagnosed, and further cases of CNS SCN–NTRK are needed to confirm or not their place in the next WHO Classification of CNS tumours. The outcome of extra‐CNS SCN–NTRK seems to be favourable, with a median follow‐up of 24 months (ranging from 1 month to 54 years). 5 , 6 , 8 , 9 , 10 , 11 , 12 , 13 , 15 , 16 , 18 , 20 , 21 , 25 Recurrences are relatively frequent (35% of cases) and mostly local, 5 , 13 , 16 , 17 , 22 , 25 but distant metastases have been reported. 16 , 17 , 22 However, only 5% of patients 16 , 20 die from their disease (median overall survival of 24 months). As illustrated here, targeted therapy (with crizotinib or entrectinib) may be an interesting alternative therapeutic option (for patients with unresectable or recurring tumours). 8 , 25

In summary, our study confirms that SCN–NTRK have a common myofibroblastic differentiation throughout the many affected organs and represent a new MC distinct from their differential histomolecular diagnoses, including IFS. Further series are needed to confirm these data and suggest a consensual nosology.

Conflicts of interest

The authors declare that they have no conflicts of interest directly related to the topic of this article.

Supporting information

Table S1. Materials and methods used in this work.

Click here for additional data file. (26.2KB, xlsx)

Acknowledgements

The authors are grateful to the laboratory technicians from GHU Paris Neurosciences, Hospital Sainte‐Anne and Institut Bergonie (Bordeaux), for their assistance and the clinicians and pathologists who provided material and clinical follow‐up. FLL's research is supported by grants from the Fondation Bergonie, the SIRIC Bordeaux Brio and the charity ‘Au fil d’Oriane’. ATE's research is supported by grants from the Foundation ‘Liv et Lumiere’. AT‐E, MD JB, MLeQ, DB, PV and FLeL contributed equally to this work.

References

  • 1. Uguen A, Csanyi‐Bastien M, Sabourin J‐C, Penault‐Llorca F, Adam J. How to test for NTRK gene fusions: a practical approach for pathologists. Ann. Pathol. 2021; 41; 387–398. [DOI] [PubMed] [Google Scholar]
  • 2. Antonescu CR, Suurmeijer AJH, Zhang L et al. Molecular characterization of inflammatory myofibroblastic tumors with frequent ALK and ROS1 gene fusions and rare novel RET rearrangement. Am. J. Surg. Pathol. 2015; 39; 957–967. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Bourgeois JM, Knezevich SR, Mathers JA, Sorensen PHB. Molecular detection of the ETV6‐NTRK3 gene fusion differentiates congenital fibrosarcoma from other childhood spindle cell tumors. Am. J. Surg. Pathol. 2000; 24; 937–946. [DOI] [PubMed] [Google Scholar]
  • 4. Kallen ME, Hornick JL. The 2020 WHO classification: what's new in soft tissue tumor pathology? Am. J. Surg. Pathol. 2021; 45; e1–e23. [DOI] [PubMed] [Google Scholar]
  • 5. Agaram NP, Zhang L, Sung Y‐S et al. Recurrent NTRK1 gene fusions define a novel subset of locally aggressive Lipofibromatosis‐like neural tumors. Am. J. Surg. Pathol. 2016; 40; 1407–1416. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Bartenstein DW, Coe TM, Gordon SC et al. Lipofibromatosis‐like neural tumor: case report of a unique infantile presentation. JAAD Case Rep. 2018; 4; 185–188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Brčić I, Godschachner TM, Bergovec M et al. Broadening the spectrum of NTRK rearranged mesenchymal tumors and usefulness of pan‐TRK immunohistochemistry for identification of NTRK fusions. Mod. Pathol. 2021; 34; 396–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Dupuis M, Shen Y, Curcio C et al. Successful treatment of lipofibromatosis‐like neural tumor of the lumbar spine with an NTRK‐fusion inhibitor. Clin. Sarcoma Res. 2020; 10; 14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Haller F, Knopf J, Ackermann A et al. Paediatric and adult soft tissue sarcomas with NTRK1 gene fusions: a subset of spindle cell sarcomas unified by a prominent myopericytic/haemangiopericytic pattern. J. Pathol. 2016; 238; 700–710. [DOI] [PubMed] [Google Scholar]
  • 10. Higaki‐Mori H, Hisaoka M, Yoshida Y, Ehara Y, Shindo M, Yamamoto O. Infantile lipofibromatosis‐like neural tumour investigated by a fusion gene detection assay. Acta Derm. Venereol. 2020; 100; adv00180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Kang J, Park JW, Won J‐K et al. Clinicopathological findings of pediatric NTRK fusion mesenchymal tumors. Diagn. Pathol. 2020; 15; 114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Kohsaka S, Saito T, Akaike K et al. Pediatric soft tissue tumor of the upper arm with LMNA‐NTRK1 fusion. Hum. Pathol. 2018; 72; 167–173. [DOI] [PubMed] [Google Scholar]
  • 13. Lao IW, Sun M, Zhao M, Yu L, Wang J. Lipofibromatosis‐like neural tumour: a clinicopathological study of ten additional cases of an emerging novel entity. Pathology 2018; 50; 519–523. [DOI] [PubMed] [Google Scholar]
  • 14. Malik F, Santiago T, Newman S, McCarville B, Pappo AS, Clay MR. An addition to the evolving spectrum of lipofibromatosis and lipofibromatosis‐like neural tumor: molecular findings in an unusual phenotype aid in accurate classification. Pathol. Res. Pract. 2020; 216; 152942. [DOI] [PubMed] [Google Scholar]
  • 15. Panse G, Reisenbichler E, Snuderl M, Wang WL, Laskin W, Jour G. LMNA‐NTRK1 rearranged mesenchymal tumor (lipofibromatosis‐like neural tumor) mimicking pigmented dermatofibrosarcoma protuberans. J. Cutan. Pathol. 2021; 48; 290–294. [DOI] [PubMed] [Google Scholar]
  • 16. Suurmeijer AJH, Dickson BC, Swanson D et al. A novel group of spindle cell tumors defined by S100 and CD34 co‐expression shows recurrent fusions involving RAF1, BRAF, and NTRK1/2 genes. Genes Chromosomes Cancer 2018; 57; 611–621. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Davis JL, Lockwood CM, Albert CM, Tsuchiya K, Hawkins DS, Rudzinski ER. Infantile NTRK‐associated mesenchymal tumors. Pediatr. Dev. Pathol. 2018; 21; 68–78. [DOI] [PubMed] [Google Scholar]
  • 18. Abs D, Landman S, Osio A, Lepesant P, Schneider P, Obadia D, Moguelet P, Farges C, Poirot B, Lehmann‐Che J, Lebbé C Spindle cell tumor with CD34 and S100 co‐expression and distinctive stromal and perivascular hyalinization showing EML4‐ALK fusion. J. Cutan. Pathol. 2020;48(7):896‐901. [DOI] [PubMed] [Google Scholar]
  • 19. Al‐Ibraheemi A, Folpe AL, Perez‐Atayde AR et al. Aberrant receptor tyrosine kinase signaling in lipofibromatosis: a clinicopathological and molecular genetic study of 20 cases. Mod. Pathol. 2019; 32; 423–434. [DOI] [PubMed] [Google Scholar]
  • 20. Antonescu CR, Dickson BC, Swanson D et al. Spindle cell tumors with RET gene fusions exhibit a morphologic Spectrum akin to tumors with NTRK gene fusions. Am. J. Surg. Pathol. 2019; 43; 1384–1391. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Flucke U, van Noesel MM, Wijnen M et al. TFG‐MET fusion in an infantile spindle cell sarcoma with neural features. Genes Chromosomes Cancer 2017; 56; 663–667. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Kao Y‐C, Suurmeijer AJH, Argani P et al. Soft tissue tumors characterized by a wide spectrum of kinase fusions share a lipofibromatosis‐like neural tumor pattern. Genes Chromosomes Cancer 2020; 59; 575–583. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Lopez‐Nunez O, Surrey LF, Alaggio R, Fritchie KJ, John I. Novel PPP1CB‐ALK fusion in spindle cell tumor defined by S100 and CD34 coexpression and distinctive stromal and perivascular hyalinization. Genes Chromosomes Cancer 2020; 59; 495–499. [DOI] [PubMed] [Google Scholar]
  • 24. Paton DJW, Wong D, Amanuel B, Cheah K, Ardakani NM. S100/CD34‐positive spindle cell mesenchymal neoplasm harboring KIAA1549‐BRAF fusion. Am. J. Dermatopathol. 2021; 43; 217–220. [DOI] [PubMed] [Google Scholar]
  • 25. Sheng S‐J, Li J‐M, Zou Y‐F et al. A low‐grade malignant soft tissue tumor with S100 and CD34 co‐expression showing novel CDC42SE2‐BRAF fusion with distinct features. Genes Chromosomes Cancer 2020; 59; 595–600. [DOI] [PubMed] [Google Scholar]
  • 26. Pfister SM, Reyes‐Múgica M, Chan JKC et al. A summary of the inaugural WHO classification of pediatric tumors: transitioning from the optical into the molecular era. Cancer Discov. 2022; 12; 331–355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Anderson WJ, Doyle LA. Updates from the 2020 World Health Organization classification of soft tissue and bone Tumours. Histopathology 2021; 78; 644–657. [DOI] [PubMed] [Google Scholar]
  • 28. Capper D, Jones DTW, Sill M et al. DNA methylation‐based classification of central nervous system tumours. Nature 2018; 555; 469–474. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Sturm D, Orr BA, Toprak UH et al. New brain tumor entities emerge from molecular classification of CNS‐PNETs. Cell 2016; 164; 1060–1072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Louis DN, Perry A, Wesseling P et al. The 2021 WHO classification of tumors of the central nervous system: a summary. Neuro Oncol. 2021; 23(8); 1231–1251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Hovestadt V, Jones DTW, Picelli S et al. Decoding the regulatory landscape of medulloblastoma using DNA methylation sequencing. Nature 2014; 510; 537–541. [DOI] [PubMed] [Google Scholar]
  • 32. Zhu T, Liu J, Beck S et al. A pan‐tissue DNA methylation atlas enables in silico decomposition of human tissue methylomes at cell‐type resolution. Nat. Methods 2022; 19; 296–306. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Koelsche C, Schrimpf D, Stichel D et al. Sarcoma classification by DNA methylation profiling. Nat. Commun. 2021; 12; 498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Perret R, Michal M, Carr RA et al. Superficial CD34‐positive fibroblastic tumor and PRDM10‐rearranged soft tissue tumor are overlapping entities: a comprehensive study of 20 cases. Histopathology 2021; 79; 810–825. [DOI] [PubMed] [Google Scholar]
  • 35. Anders S, Pyl PT, Huber W. HTSeq‐‐a python framework to work with high‐throughput sequencing data. Bioinforma. Oxf. Engl. 2015; 31; 166–169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Pilson D, Decker KL. A language and environment for statistical computing. Ecology 2002; 83; 3097–3107. [Google Scholar]
  • 37. Law CW, Chen Y, Shi W, Smyth GK. Voom: precision weights unlock linear model analysis tools for RNA‐seq read counts. Genome Biol. 2014; 15; R29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Wilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics 2010; 26; 1572–1573. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Crumbach L, Descotes F, Bringuier P‐P et al. Lipofibromatosis‐like neural tumor: a case report and review of the literature. Am. J. Dermatopathol. 2020; 42; 881–884. [DOI] [PubMed] [Google Scholar]
  • 40. Lyskjaer I, De Noon S, Tirabosco R et al. DNA methylation‐based profiling of bone and soft tissue tumours: a validation study of the ‘DKFZ sarcoma classifier’. J. Pathol. Clin. Res. 2021; 7; 350–360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Eyden B. Electron microscopy in the study of myofibroblastic lesions. Semin. Diagn. Pathol. 2003; 20; 13–24. [PubMed] [Google Scholar]
  • 42. Eyden B, Banerjee SS, Shenjere P, Fisher C. The myofibroblast and its tumours. J. Clin. Pathol. 2009; 62; 236–249. [DOI] [PubMed] [Google Scholar]
  • 43. Mariño‐Enríquez A, Wang W‐L, Roy A et al. Epithelioid inflammatory myofibroblastic sarcoma: an aggressive intra‐abdominal variant of inflammatory myofibroblastic tumor with nuclear membrane or perinuclear ALK. Am. J. Surg. Pathol. 2011; 35; 135–144. [DOI] [PubMed] [Google Scholar]
  • 44. Bennett JA, Nardi V, Rouzbahman M, Morales‐Oyarvide V, Nielsen GP, Oliva E. Inflammatory myofibroblastic tumor of the uterus: a clinicopathological, immunohistochemical, and molecular analysis of 13 cases highlighting their broad morphologic spectrum. Mod. Pathol. 2017; 30; 1489–1503. [DOI] [PubMed] [Google Scholar]
  • 45. Lopez‐Nunez O, John I, Panasiti RN et al. Infantile inflammatory myofibroblastic tumors: clinicopathological and molecular characterization of 12 cases. Mod. Pathol. 2020; 33; 576–590. [DOI] [PubMed] [Google Scholar]
  • 46. Papke DJ, Al‐Ibraheemi A, Fletcher CDM. Plexiform myofibroblastoma: clinicopathologic analysis of 36 cases of a distinctive benign tumor of soft tissue affecting mainly children and young adults. Am. J. Surg. Pathol. 2020; 44; 1469–1478. [DOI] [PubMed] [Google Scholar]
  • 47. Mentzel T, Dry S, Katenkamp D, Fletcher CDM. Low‐grade myofibroblastic sarcoma: analysis of 18 cases in the spectrum of myofibroblastic tumors. Am. J. Surg. Pathol. 1998; 22; 1228–1238. [DOI] [PubMed] [Google Scholar]
  • 48. Huson SM, Staab T, Pereira M et al. Infantile fibrosarcoma with TPM3‐NTRK1 fusion in a boy with bloom syndrome. Fam Cancer 2020; 21(1); 85–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49. Caldwell KJ, De La Cuesta E, Morin C, Pappo A, Helmig S. A newborn with a large NTRK fusion positive infantile fibrosarcoma successfully treated with larotrectinib. Pediatr. Blood Cancer 2020; 67; e28330. [DOI] [PubMed] [Google Scholar]
  • 50. Wong V, Pavlick D, Brennan T et al. Evaluation of a congenital infantile Fibrosarcoma by comprehensive genomic profiling reveals an LMNA‐NTRK1 gene fusion responsive to Crizotinib. J. Natl. Cancer Inst. 2016; 108; djv307. [DOI] [PubMed] [Google Scholar]
  • 51. Mahajan P, Casanova M, Ferrari A, Fordham A, Trahair T, Venkatramani R. Inflammatory myofibroblastic tumor: molecular landscape, targeted therapeutics, and remaining challenges. Curr. Probl. Cancer 2021; 45; 100768. [DOI] [PubMed] [Google Scholar]
  • 52. Vargas AC, Ardakani NM, Wong DD et al. Chromosomal imbalances detected in NTRK‐rearranged sarcomas by the use of comparative genomic hybridisation. Histopathology 2021; 78; 932–942. [DOI] [PubMed] [Google Scholar]
  • 53. Sievers P, Stichel D, Sill M et al. GOPC:ROS1 and other ROS1 fusions represent a rare but recurrent drug target in a variety of glioma types. Acta Neuropathol. 2021; 142; 1065–1069. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1. Materials and methods used in this work.

Click here for additional data file. (26.2KB, xlsx)

Articles from Histopathology are provided here courtesy of Wiley

RESOURCES