Detection of NTRK gene fusions in sarcomas: a comparative study of Pan-TRK immunohistochemistry, FISH, and RNA-Based NGS

Background. Fusion involving NTRK1, NTRK2 and NTRK3 are oncogenic driver occurring in several adult and pediatric tumor types. In sarcomas they are mostly found in infantile fibrosarcoma, inflammatory (IFS), inflammatory myofibroblastic tumor (IMT) and in the so-called “NTRK-rearranged spindle cell neoplasm” entity described in the current WHO (2020) classification, including lipofibromatosis-like neural tumor, fibrosarcoma-like and malignant peripheral nerve sheath tumor-like spindle cell neoplasms.

Methods. We retrospectively reviewed 92 soft tissue and bone sarcomas diagnosed at the Rizzoli Institute between 2019 and 2023, in which pan-TRK IHC was performed. 17 tumours showed positive staining and were further assessed using FISH for NTRK1, NTRK2, and NTRK3 rearrangements. A subset of 12 cases underwent RNA-based NGS for fusion detection.

Results. In total, we collected data from 17 patients who tested positive for pan-TRK antibody and compared pan-TRK IHC and molecular testing for the detection of NTRK rearrangement in sarcomas. FISH analysis detected NTRK rearrangements in 4/17 cases (23.5%), while NGS confirmed NTRK fusions in 3/12 cases (25%). All NTRK fusion-positive cases confirmed by both FISH and NGS showed diffuse pan-TRK staining. Two additional cases exhibited pan-TRK diffuse positivity but were NTRK wild-type by NGS and harboured BCOR::MAML1 and EWSR1::NACC2 fusions, respectively. 1 case with focal positivity by immunohistochemistry and NTRK rearrangement by FISH was not confirmed by NGS.

Conclusions. Pan-TRK IHC can be considered an initial screening tool to identify sarcomas potentially harbouring NTRK fusions. However, the presence of diffuse pan-TRK immunoreactivity in NTRK-wild-type tumours highlights the need for cautious interpretation of IHC results. FISH may represent a useful intermediate diagnostic tool when NGS is unavailable but requires cautious interpretation particularly in cases with atypical or isolated signals. NGS-based molecular confirmation remains essential for the definitive identification of NTRK fusions.

Introduction

The tropomyosin receptor kinase (TRK) family comprises transmembrane proteins – TRKA, TRKB, and TRKC – that are physiologically expressed in neural tissues and play a crucial role in neuronal development and neurophysiological functioning. These receptors are encoded by the NTRK1, NTRK2, and NTRK3 genes, located on chromosomes 1q23.1, 9q21.33, and 15q25.3, respectively. Each TRK receptor consists of three main domains: an extracellular ligand-binding domain, a single transmembrane region, and an intracellular domain containing the conserved tyrosine kinase (TK) region, which is essential for signal transduction ^1,2.

Wild-type TRK receptors are activated through homo-dimerization and transphosphorylation, processes that are mediated by the binding of neurotrophin ligands such as nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), and neurotrophin-3 (NT-3). The dimerization of TRK leads to the activation of multiple downstream signal transduction pathways, including the MAPK pathway, the PI3K-AKT pathway, and the phospholipase C-gamma (PLCγ) pathway. These pathways regulate key cellular functions such as proliferation, differentiation, and survival in both normal and neoplastic neuronal cells ³.

NTRK genes are implicated in oncogenic fusions that result in ligand-independent constitutive activation of TRK signaling. Such fusions have been identified in a wide spectrum of pediatric and adult tumors. Among mesenchymal neoplasms, they are mostly found in infantile fibrosarcoma (IFS), inflammatory myofibroblastic tumors (IMTs), and in a provisional tumour category introduced in the 2020 WHO Classification of Soft Tissue and Bone Tumors: “spindle cell neoplasm with NTRK fusion.” This latter category encompasses lipofibromatosis-like neural tumors (LPFNTs), lipofibromatosis-like-neural tumors (LPFNTs) fibrosarcoma-like and malignant peripheral nerve sheath-like tumors ^4-7.

The prevalence of NTRK fusions in both pediatric and adult sarcomas is reported to be less than 1% ^8,9. To date, various types of NTRK fusions have been described in sarcomas, including LMNA::NTRK1 and EML4::NTRK3. The LMNA::NTRK1 fusion has been identified in low-grade sarcomas with myopericytoma-like features and in uterine spindle cell sarcomas, whereas the EML4::NTRK3 fusion has been reported in infantile fibrosarcoma ^{5, 10}.

The ETV6::NTRK3 fusion, resulting from the t(12;15)(p13;q25) translocation, is one of the earliest identified and most well-characterized NTRK rearrangements. It represents the canonical fusion for infantile fibrosarcoma, occurring in approximately 90% of cases ⁷.

These fusions result from either intrachromosomal or interchromosomal rearrangements involving NTRK genes and lead to the in-frame juxtaposition of the 3 region of an NTRK gene (encoding the C-terminal tyrosine kinase domain) with the 5′ sequence of a fusion partner gene (typically encoding an N-terminal dimerization/ oligomerization domain). This fusion configuration enables constitutive activation or overexpression of the kinase domain, thereby contributing to oncogenic potential.

The partner gene generally contributes the 5′ promoter region, driving aberrant expression of the chimeric transcript ^1,11,12. NTRK gene rearrangements have recently emerged as actionable targets for molecular therapy with novel compounds that selectively inhibit the constitutively active fusion proteins generated by these genetic alterations ¹¹.

Entrectinib and larotrectinib are the first tumour-agnostic tyrosine kinase (TK) inhibitors to receive FDA and EMA approval for the treatment of unresectable, locally advanced, or metastatic solid tumors harboring an NTRK gene fusion. Larotrectinib is the most specific of the two, selectively targeting TRKA, TRKB, and TRKC, whereas entrectinib additionally inhibits ROS1 and ALK kinases. These agents have demonstrated antitumour efficacy irrespective of the NTRK fusion type, tumour histology, or anatomical site of origin, in both adult and pediatric populations.

Given this broad therapeutic applicability, it is crucial to implement diagnostic strategies that can be routinely used to screen tumors for the presence of NTRK fusions ^13,14.

The detection of NTRK gene fusions for patient selection in targeted therapy can be performed on clinical specimens using various methodologies. Current international guidelines from ASCO and ESMO recommend immunohistochemistry (IHC) as an initial screening tool, followed by confirmation with RNA-based next-generation sequencing (NGS), which is considered the most sensitive and specific approach ^15,16. Another available method for the detection of NTRK rearrangement is fluorescence in situ hybridization (FISH) ^16,17.

In the present study, we reviewed a series of soft tissue and bone sarcomas that were evaluated for pan-TRK expression by immunohistochemistry (IHC) and for NTRK1, NTRK2, and NTRK3 rearrangements by FISH. The objective was to assess and compare the diagnostic utility of IHC and FISH in identifying NTRK gene rearrangements within this group of tumors. Furthermore, we correlated IHC and FISH positivity with the presence of NTRK fusions confirmed by next-generation sequencing (NGS).

Materials and methods

PATIENTS

Following approval by the institutional ethics committee (protocol code: CE AVEC: 59/2024 Oss/IOR; approval date: 14 February 2024), we retrospectively analyzed a series of 92 soft tissue and bone sarcomas diagnosed at the Rizzoli Institute over a 5-year period (2019-2023), in which pan-TRK immunohistochemistry had been performed for diagnostic or predictive purposes. The study cohort included tumors with spindle cell morphology not clearly suggestive of a specific diagnostic entity and clinically aggressive mesenchymal tumors, such as dedifferentiated chondrosarcomas in which pan-TRK testing was requested for predictive purposes.

PAN-TRK IMMUNOHISTOCHEMISTRY (IHC)

All tumor specimens were fixed in 10% buffered formalin and routinely processed for histological evaluation. Pan-TRK immunohistochemical staining for TRKA, TRKB, and TRKC expression was performed using a commercially available assay (rabbit monoclonal antibody, clone EPR17341, RTU 28 μg/ml, Roche, Ventana, Tucson, AZ, USA). This antibody targets a homologous region near the C-terminus that is preserved in both fusion and wild-type TRK proteins.

Formalin-fixed paraffin-embedded (FFPE) tissue blocks were sectioned at 3 μm thickness, mounted on slides, and incubated at 58 °C for 2 hours. All staining procedures were carried out on the BenchMark Ultra automated staining platform according to the manufacturer’s instructions (Ventana Medical Systems, Tucson, AZ, USA), using the OptiView DAB Detection Kit, as previously described by Benini et al. ¹⁸.

Non-neoplastic brain tissue was included as a positive control in each run. The antibody was incubated for 16 minutes at 36 °C, following antigen retrieval with Cell Conditioning 1 (CC1, pH 8) for 88 minutes at 100 °C.

All stained slides were independently evaluated and scored by two pathologists (MG and AR). Immunoreactivity was assessed using a semi-quantitative scale as detailed in Table I ¹⁹.

Cases were considered positive if more than 1% of tumor cells showed staining at any intensity above background, in any subcellular localization pattern, including membranous, cytoplasmic, nuclear or combinations thereof ¹⁷.

FLUORESCENCE IN SITU HYBRIDIZATION (FISH)

Fluorescence in situ hybridization (FISH) for the detection of NTRK1, NTRK2, and NTRK3 gene rearrangements was performed on interphase nuclei using three commercially available break-apart probe kits: ZytoLight SPEC NTRK1 Dual Colour Break-Apart Probe, ZytoLight SPEC NTRK2 Dual Colour Break-Apart Probe, and ZytoLight SPEC NTRK3 Dual Colour Break-Apart Probe (ZytoVision, Bremerhaven, Germany). Each probe is designed to label the 3′ region with ZyOrange and the 5′ region with ZyGreen, enabling detection of gene rearrangements involving NTRK1, NTRK2, and NTRK3, respectively.

For each case, representative 3-μm-thick FFPE tissue sections were prepared. Slides were incubated overnight at 60 °C, deparaffinized in xylene, and dehydrated in a graded ethanol series. Pretreatment was performed using the Histology FISH Accessory Kit (Dako, Glostrup, Denmark), according to the manufacturer’s protocol, as previously described ²¹.

A serial hematoxylin-eosin-stained section was prepared from each tumour, and areas of viable, non-necrotic neoplastic tissue were marked by a pathologist to guide the hybridization procedure.

To interpret FISH results, at least 100 non-overlapping tumor nuclei exhibiting strong, well-delineated signals were visually examined under an Olympus BX41 fluorescence microscope (Tokyo, Japan) at 100× magnification using dual-spectrum filters (Spectrum Green and Spectrum Orange). A case was considered positive for NTRK rearrangement based on the identification of one of the following two signal patterns: (1) A classic break-apart pattern, defined by one intact fusion signal along with two split signals (3′ orange and 5′ green) separated by a distance of at least three signal diameters. (2) An atypical pattern, consisting of one fusion signal and a single isolated 3′ orange signal in the absence of a corresponding green signal. Isolated orange signals are indicative of either 5′ NTRK region deletion or unbalanced translocations involving this chromosomal segment. (3) Genomic aberration due to small deletion, duplications or inversion might result in inconspicuous patterns. These signal patterns should be further investigated.

To minimize subjectivity, each case was independently evaluated in a double-blinded manner by at least two experienced analysts. Any discrepancies in signal interpretation were subsequently resolved through consensus review.

Tumors were defined as FISH-positive if at least 10% of tumor cells demonstrated either of these rearrangement patterns.

Copy number alterations for NTRK1, NTRK2, and NTRK3 genes were also assessed. Copy number gain was defined as a mean of 3-5 signals in ≥10% of analysed nuclei. Images were acquired using a Colour View III CCD camera (Olympus) and processed with CytoVision imaging software, version 7.5 (Leica Biosystems, Richmond, USA).

ANCHORED MULTIPLEX PCR (AMP)-BASED TARGET NGS

Twelve samples were analyzed using an RNA-based anchored multiplex sequencing assay, as previously described ¹⁸. Briefly, total RNA was extracted from FFPE specimens using the RNeasy FFPE Tissue Kit (Qiagen GmbH, Hilden, Germany), following the manufacturer’s protocol. Sections measuring 6-8 μm in thickness were cut from representative paraffin blocks for RNA extraction.

RNA quantity was determined using a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). Libraries were prepared using the Archer FusionPlex Sarcoma v2 panel, which includes 659 unidirectional gene-specific primers (GSPs) targeting 63 genes known to be involved in oncogenic fusions in sarcomas.

For library construction, 200 ng of total RNA was used as input, in accordance with the manufacturer’s instructions. Initial RNA quality was assessed using the Archer PreSeq RNA Quality Control (QC) assay. cDNA libraries were generated only if the sample met the threshold for acceptable quality (PreSeq qPCR crossing point ≤ 30). In this assay, the quality of first-strand cDNA synthesis is used as a direct measure of RNA integrity. Briefly, prior to library preparation, all samples undergo real-time PCR-based quality control using VCP primers (Archer® PreSeq™ RNA QC Assay), as required by the Archer protocol for assessing RNA quality. Only samples that meet the VCP threshold proceed to library construction. Specifically, a Ct value > 30 (PreSeq qPCR crossing point ≤ 30) indicates low RNA quality, and such samples are deemed insufficient for testing. After sequencing, the Archer Analysis software performs additional quality control assessments within the analysis pipeline. Key quality metrics, including Fusion QC, are provided based on a pool of Gene-Specific Primers (GSPs) targeting four control gene transcripts: CHMP2A, GPI, RAB7A, and VCP Known to be moderately expressed across most tissue type. The Fusion QC score represents the average number of unique RNA start sites calculated per control GSP2. A Fusion QC score greater than 10 is required to ensure adequate RNA quality for reliable assay results. Notably, the Fusion QC score increases as the Ct value decreases, reflecting higher RNA integrity. All purification steps during library preparation were performed using Agencourt AMPure XP beads (Beckman Coulter, Brea, CA, USA). Purified libraries were quantified using the Ion Library TaqMan Quantitation Kit (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA) and subsequently pooled to achieve equimolar concentrations of 25 pM.

The prepared libraries (anchored multiplex PCR amplicons) were loaded onto an Ion 510 chip using the Ion Chef System (ThermoFisher Scientific, Waltham, MA, USA) and sequenced on the Ion Torrent S5 next-generation sequencing platform.

Sequencing reads were aligned to the human reference genome (GRCh37/hg19), and data files (BAM format) were analysed using the Archer Analysis software (Web Application v7.1.0.38; Pipeline Version 7.1.0-14; ArcherDx, Invitae, San Francisco, CA) under default settings.

Tumors were considered NTRK1-3 fusion-positive when an in-frame fusion involving the kinase domain of NTRK was identified.

STATISTICAL ANALYSIS

Data were analyzed descriptively. Results are reported as absolute numbers and percentages. Concordance between pan-TRK immunohistochemistry, FISH, and NGS was assessed descriptively (observed agreement among the cases tested). Due to the retrospective design and stepwise application of molecular testing only to Pan-TRK-positive cases, formal calculations of sensitivity, specificity, or predictive values were not possible.

Results

PATHOLOGICAL DIAGNOSIS

The majority of cases were classified as malignant spindle cell mesenchymal neoplasms, arranged in fascicles to herringbone pattern, and lacking specific morphological and immunohistochemical differentiating features. Exceptions included cases 10, 12, and 13, which showed pleomorphic morphology with immunohistochemical expression of myogenic markers; case 5, characterized by pleomorphic features with pleomorphic lipoblasts; four malignant peripheral nerve sheath tumor (MPNST) showing loss of H3K27me3 expression; and case 11, showed MDM2 amplification. These pleomorphic cases belonged to the subgroup of tumors selected for predictive purposes and were molecularly investigated based on clinical indications related to tumor aggressiveness rather than diagnostic uncertainty.

PAN-TRK IMMUNOHISTOCHEMISTRY

We retrieved 92 patients with a diagnosis of soft tissue and bone sarcoma in which Pan-TRK expression was tested by immunohistochemistry (IHC). Of these, 17 patients exhibited positive pan-TRK IHC results. Staining patterns was diffuse (5/17), focal (11/17) and rare positive cells (1/17). Regarding subcellular localisation, 16 samples showed cytoplasmic TRK protein expression; among these, five also demonstrated additional membranous expression.

One sample (case 9) showed both cytoplasmic and nuclear membrane expression, while another sample (case 7) showed only membranous expression (Fig. 1). Notably, isolated cytoplasmic staining was the most common pattern occurring in 10 out of 17 cases. Immunohistochemical patterns and the distribution of subcellular localization for TRK protein are summarized in Table II. It is also worth noting that two cases (case 3 and case 4) had been previously reported ^20,21.

FISH ANALYSIS

Interphase FISH analysis for NTRK1, NTRK2, NTRK3 was performed in all 17 cases; however, one case (case 13) was not evaluable due to poor signal quality and was therefore classified as not adequate (NA).

NTRK1 rearrangements were identified in three cases (cases 8, 9, and 16) and an NTRK3 rearrangement in one case (case 4); all positive cases met the criterion of > 10% tumor cells exhibiting break-apart signals.

In case 8 and 9, NTRK1 FISH analysis revealed the presence of one single orange signal mapping to the 3’ region of NTRK1, combined with a loss of the green signal corresponding to the 5′ region (Fig. 1). One case (case 16) showed NTRK1 split-apart signals approximately in 15% of analyzed tumor cells along with copy number gain. Regarding the NTRK3-positive tumor (case 4), a classic break apart signal was observed in approximatively the 50% of tumor cells, with a separation distance of at least three signal diameters together with an increase of the 3’ region of NTRK3. Copy number gain involving NTRK1, NTRK2, and NTRK3 was detected in case 11. Sample 15 had insufficient tumor tissue for a complete FISH analysis. No NTRK2 rearrangements were identified in this series.

Overall, among the pan-TRK IHC-positive cases, concordant FISH results were observed in 4 of 17 cases (23.5%).

NGS FUSION DESCRIPTION

NGS RNA analysis was successfully performed in 12 patients. The remaining five cases were not analyzed by NGS due to lack of available biological material. (Tab. I). An EML4::NTRK3 fusion was detected in the case (case 4) that exhibited an NTRK3 rearrangement by FISH, whereas an LMNA::NTRK1 fusion was identified in two patients (cases 8 and 9) with NTRK1 rearrangements confirmed by FISH (Fig. 2). The EML4::NTRK3 fusion detected in this analysis involved exon 2 of EML4 (NM_019063.3) and exon 14 of NTRK3 (NM_002530.3).

The genomic coordinates of the RNA fusion junction localised to Chr2:42472827 and Chr15:88576276, respectively. Notably, both the coiled-coil domain of the EML4 gene and the tyrosine kinase domain at the C-terminus of NTRK3 were preserved. In the intrachromosomal LMNA::NTRK1 fusion, exon 2 of LMNA (NM_005572.3) was joined to exon 11 of NTRK1 (NM_002529.3) in both patients (cases 8 and 9). In these two LMNA::NTRK1 fusion-positive cases, the genomic coordinates of the RNA fusion junction were localized to Chr1:156100564 for LMNA and Chr1:156844698 for NTRK. The resulting fusion transcript retained a portion of the alpha-helical rod domain of LMNA and the kinase domain of NTRK1.

Each NTRK fusion event was predicted to be in-frame according to Archer software. The frame status was further confirmed by manual inspection, comparing the sequences to the consensus coding sequence (CCDS) available in the NCBI database (https://www.ncbi.nlm.nih.gov/).

Patient 3 harbored a BCOR::MAML1 fusion with breakpoints located in exon 15 of the BCOR gene and exon 1 of MAML1. This rearrangement has previously been described by Cocchi et al ²¹. In patient 6, Archer software identified a novel in-frame EWSR1::NACC2 fusion, involving exon 7 of EWSR1 and exon 3 of NACC2. No gene fusions were identified in seven cases, while the remaining five cases were not tested. With regard to immunohistochemical correlation, all tumors harboring NTRK fusions showed diffuse pan-TRK immunoreactivity; however, diffuse staining was also observed in two cases with non-NTRK gene fusions (BCOR::MAML1 and EWSR1::NACC2).

In addition, LMNA::NTRK1 fusion-positive tumors showed distinct subcellular localization patterns, with predominant cell membrane expression in case 8 and nuclear membrane expression in case 9. The single case harbouring an EML4::NTRK3 fusion (case 4) showed membranous staining in addition to cytoplasmic staining. Tumors harboring BCOR::MAML1 and EWSR1::NACC2 fusions exhibited combined cytoplasmic and membranous pan-TRK expression. Overall, NTRK fusions were confirmed by NGS in 3 of the 12 cases analyzed (25%), all of which showed diffuse pan-TRK immunoreactivity providing a descriptive measure of IHC-NGS agreement within the tested subset. However, diffuse pan-TRK staining was also observed in two additional cases that were NTRK wild-type by NGS and harbored non-NTRK gene fusions. No NTRK rearrangements were identified among cases displaying focal pan-TRK staining or rare positive cells.

Figure 3 provides a concise overview of the diagnostic workflow, integrating results from multiple methodologies and outlining the final diagnosis.

Discussion

In the most recent edition of the WHO Classification of Soft Tissue and Bone Tumors, NTRK-rearranged spindle cell neoplasm (excluding infantile fibrosarcoma, IFS) has been recognized as an emerging entity.

More recently, it has been demonstrated that NTRK gene fusions represent actionable therapeutic targets in cancer, with promising results observed using highly selective tyrosine kinase inhibitors (TKIs) such as larotrectinib and entrectinib ^13,14. The choice of methodology for detecting NTRK fusions must consider both the histological features of the tumor and the resources available within the diagnostic laboratory.

In the context of sarcomas, some morphological and immunohistochemical criteria are suggestive for NTRK-rearranged spindle cell neoplasms, as outlined in the WHO Classification of Soft Tissue and Bone Tumors (5th edition), in the chapter on NTRK-rearranged spindle cell neoplasm ²². Following international guidelines (ESMO, ASCO) we employed pan-TRK immunohistochemistry (IHC) as an initial screening tool to identify sarcomas potentially harbouring NTRK fusions ¹⁵. Pan-TRK IHC is considered an appropriate first-line assay to triage cases for subsequent molecular confirmation.

The FDA- and CE-IVD-approved clone EPR17341 is currently the most widely used antibody for pan-TRK detection. It targets a conserved C-terminal domain shared by TRKA, TRKB, and TRKC proteins, which is retained in both wild-type and fusion variants ^1,14,16. Previous large immunohistochemical studies have reported highly variable sensitivity and specificity of pan-TRK expression across a wide range of solid tumors ¹⁷.

With respect to sarcomas, low specificity has been reported, particularly in tumors exhibiting neural or myogenic differentiation, as wild-type TRK proteins are physiologically expressed in neural and smooth muscle tissues ¹⁶. Additional false-positive staining can occur in non-NTRK-fused tumors, including leiomyosarcomas and BCOR-rearranged sarcomas ^23,24.

False pan-TRK IHC positivity has been reported to be associated with increased expression of NTRK genes, particularly NTRK3, which may be upregulated in sarcomas independently of gene rearrangements ²⁴. In contrast, false-negative pan-TRK immunohistochemical results have been reported to be relatively uncommon, as shown by Hechtman et al. ¹⁷. The diagnostic performance of the pan-TRK antibody clone EPR17341 has been evaluated in large series, showing high sensitivity for NTRK fusion-positive tumors but variable specificity, particularly in mesenchymal neoplasms ²⁵.

In our cohort, five pan-TRKpositive tumors were not assessed by NGS due to the unavailability of biological material, a limitation related to the retrospective study design, therefore, reliable estimation of the overall sensitivity of the pan-TRK assay cannot be determined.

As described in recently published studies, different subcellular staining patterns may be observed with pan-TRK immunohistochemistry, all of which are considered positive¹⁷. Notably, subcellular staining patterns have been shown to correlate with the fusion partner, as the latter can influence protein localization, leading to atypical subcellular distribution in contrast to the membrane-associated expression of wild-type TRK proteins ¹⁷.

In our series, cytoplasmic staining represented the predominant pattern. Cell membrane and nuclear membrane staining were also observed in a subset of cases.

For patients to be eligible for targeted therapies, confirmation of the fusion event through molecular techniques is required. FISH is a rapid and accessible method for the detection of NTRK rearrangements and may be employed for further selection of cases for RNA-based NGS analysis, particularly when upfront NGS testing is unavailable.

However, several interpretative pitfalls may arise with FISH, especially in the presence of subtly split signals or isolated 3’ signals which are more common in intrachromosomal rearrangements. These patterns can be challenging to interpret due to their borderline nature, especially when they involve subtle separation of the green and orange signals. Notably, isolated 3’ signals do not necessarily correlate with an active fusion event and may instead reflect proximal deletions near the NTRK1 breakpoint region or unbalanced rearrangements.

Such cases require careful interpretation and, whenever possible, additional molecular validation such as NGS, to confirm true structural rearrangements and their biological relevance. Additional challenges including signal copy number gains, may further complicate interpretation. Given the potential for such ambiguous results, intra-observer reproducibility is essential. In this study, each case was independently reviewed by two experienced analysts to ensure consistency and accuracy in FISH interpretation, thereby minimizing the risk of misinterpretation associated with atypical signal patterns. In our series, all tumors underwent FISH analysis.

Notably, in our series, the two cases showing isolated 3′ NTRK1 signals were those in which an NTRK1 fusion was subsequently confirmed by molecular analysis, underscoring the diagnostic relevance of these borderline FISH patterns.

Overall, NTRK rearrangements were identified in only four cases (23.5%), with NTRK1 rearrangements observed in three cases and an NTRK3 rearrangement detected in one case.

These findings support the notion that NTRK rearrangements occur in only a small subset of mesenchymal tumors exhibiting pan-TRK positivity.

Beyond the detection of rearrangements, our analysis also revealed increased copy number of NTRK genes in two cases. However, as centromeric copy number was not assessed, the presence of polysomy cannot be excluded. Further studies are needed to clarify the biological and clinical significance of this observation.

Regarding NGS analysis, two cases with NTRK1 rearrangement identified by FISH tested positive for an LMNA::NTRK1 fusion by NGS. In both cases, the LMNA::NTRK1 fusion was in-frame, involving exon 2 of LMNA and exon 11 of NTRK1. This represents one of the shortest LMNA::NTRK1 fusion transcripts reported in the literature. The fusion results from an intrachromosomal rearrangement involving loci on chromosome 1q22 (LMNA) and chromosome 1q23.1 (NTRK1). Exons 1-2 of LMNA encode the N-terminal globular domain and part of the alpha-helical rod domain, suggesting a potential role in the functional activation of the resulting chimeric protein.

In sarcomas, LMNA::NTRK1 represents the most frequently reported genetic alteration in lipofibromatosis-like neural tumors.

In our experience, one tumor harbouring an LMNA::NTRK1 fusion exhibited nuclear membrane immunostaining. This pattern appears inconsistent with the nature of the fusion, which involves the loss of the genomic region encoding the nuclear localization signal (NLS). Notably, this staining pattern was not observed in the second case with LMNA::NTRK1 fusion, which instead demonstrated cell membrane immunoreactivity. Of particular interest, one case displaying focal pan-TRK positivity and an NTRK1 rearrangement by FISH was negative by NGS.

This finding may reflect structural rearrangements detected by FISH that do not generate a stable or expressed fusion transcript. This phenomenon can be explained by several biological mechanisms, including short-range intrachromosomal rearrangements or microdeletions. Alternatively, low transcript abundance or breakpoints not covered by the targeted RNA panel may have impaired NGS detection. RNA degradation in FFPE tissue is a well-known limitation of fusion assays; however, as detailed in the Methods section, the likelihood that degradation contributed to this NGS-negative result is minimized in our Archer FusionPlex workflow through specific RNA quality-control steps.

This case highlights that such discordances, although uncommon, can occur and support the use of complementary diagnostic approaches.

This raises the possibility that not all aberrant NTRK FISH patterns correspond to true fusion events, highlighting the need for molecular confirmation in selected cases.

In sarcomas, in addition to LMNA, NTRK1 forms fusions with various gene partners, including TPM3 and TPR. All of these fusions result from intrachromosomal rearrangements involving chromosome 1q23.1 (NTRK1 locus) and either chromosome 1q21.3 (TPM3), 1q22 (LMNA), or 1q31.1 (TPR). These rearrangements can occur via interstitial deletion, as in LMNA::NTRK1, or through chromosomal inversion, as in TPM3::NTRK1 or TPR::NTRK1⁴.

The NTRK3 FISH-positive case in our cohort corresponded to an EML4::NTRK3 fusion, which was previously reported by Palmerini et al ²⁰. This fusion involves exon 2 of EML4 joined to exon 14 of NTRK3, encompassing the coiled-coil domain of EML4 and the tyrosine kinase (TK) domain of NTRK3. This fusion has been reported in congenital fibrosarcoma (CFS), cervical sarcoma, and spindle cell sarcoma ^26,27.

In our study, pan-TRK IHC for the EML4::NTRK3 case demonstrated both cytoplasmic and membranous staining of tumor cells. This staining pattern differs from that observed in other EML4::NTRK3-positive tumors, such as CFS, in which both cytoplasmic and nuclear localization have been described.

The presence of a complete NTRK tyrosine kinase (TK) domain without frameshift variants is essential for maintaining the oncogenic potential of NTRK gene fusions. All NTRK fusions identified in this study were predicted to be in-frame, with preservation of the TK domain. Notably, each NTRK fusion-positive cases demonstrated diffuse pan-TRK immunoreactivity.

However, we also report two cases displaying diffuse pan-TRK positivity that were NTRK wild-type by NGS. These tumors harboured BCOR::MAML1 and EWSR1::NACC2 fusions, respectively. In the context of BCOR-rearranged sarcomas, pan-TRK positivity has been described by Kao et al., and was attributed to NTRK3 upregulation ²⁴.

Consistenly, RNA expression analysis of both BCOR::MAML1 and EWSR1::NACC2 fusion-positive cases revealed increased relative expression of NTRK3, indicating that pan-TRK immunoreactivity in these tumors is driven by NTRK3 overexpression rather than by NTRK gene fusions (Supplementary Figure 1).

Limitations section

This study was conducted on a retrospective cohort, and only a subset of IHC-positive cases underwent NGS analysis due to clinical factors unrelated to assay performance. Although RNA quality from FFPE tissue is an inherent challenge in targeted RNA-based approaches, the workflow used in this study includes internal controls and quality metrics to ensure technical reliability. Nonetheless, rare fusion variants, very low transcript expression, or breakpoints not represented in the targeted design may theoretically escape detection. While these considerations did not account for the absence of NGS testing in our cohort, they remain relevant in routine diagnostic practice and should be considered when interpreting discrepancies between IHC, FISH, and NGS results.

Conclusion

In this study, we compared pan-TRK immunohistochemistry and molecular testing for the detection of NTRK rearrangements in sarcomas. FISH analysis performed on the 17 IHC-positive cases yielded positive results in four cases, while NGS confirmed NTRK fusions in three of these. Notably, some cases with diffuse pan-TRK staining were NTRK wild-type, highlighting the potential for false-positive IHC results, particularly in sarcomas harboring BCOR or other non-NTRK fusions. These findings suggest that, although FISH is less specific than NGS, it remains a valuable intermediate diagnostic tool in cases with pan-TRK positivity, particularly when access to NGS is limited or unavailable. Overall, pan-TRK IHC should be considered a screening assay in sarcomas and interpreted with caution, with molecular confirmation recommended to establish the presence of NTRK rearrangements. These results also underscore the need for careful interpretation of staining patterns and awareness of known IHC pitfall in mesenchymal tumors.

ACKNOWLEDGMENTS

The authors would like to thank Cristina Ghinelli for her assistance with graphic design and Monica Contoli for her support in retrieving archived histological slides. We are grateful to the Musculoskeletal Tumour Biobank (BIOTUM) - Biobanca dei Tumori Muscoloscheletrici-a member of the CRB-IOR-for providing the biological samples used in this study.

CONFLICTS OF INTEREST STATEMENT

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

FUNDING

This research received no external funding; institutional funding was used.

AUTHOR CONTRIBUTIONS

Conceptualization, S.C., S.B., and M.G.; methodology, G.G., E.S.; software, A.R., and G.M.; validation, G.G.; resources, A.R, G.M., and E.S.; data curation, S.C., G.G., M.M., A.P., S.B.; writing-original draft preparation, S.C., S.B., and M.G.; writing-review and editing, S.C., S.B., and M.G. All authors have read and agreed to the published version of the manuscript.

ETHICAL CONSIDERATION

This study was approved by Institutional Ethics Committee of Area Vasta Emilia Centro della Regione Emilia-Romagna (code: CE AVEC:59/2024 Oss/IOR).

The research was conducted ethically, with all study procedures being performed in accordance with the requirements of the World Medical Association’s Declaration of Helsinki.

Written informed consent was obtained from each participant/patient for study participation and data publication.

History

Received: May 29, 2025

Accepted: March 13, 2026

Figures and tables

Figure 1. (A, D, G, L): Hematoxylin and eosin (H&E) staining of sarcomas harbouring EML4::NTRK3 (case 4) (A), LMNA::NTRK1 (case 8) (D), BCOR::MAML1 (case 3) (G), and a fusion-negative sarcoma (case 7) (L). Original magnification: 20×. (B, E, H, M): Immunohistochemical staining with pan-TRK antibody showing diffuse cytoplasmic and membranous positivity in fusion-positive cases (cases 4, 8, 3) and focal membranous staining in the fusion-negative case (case 7). Original magnification: 20×. (C, F, I, N): Representative FISH images for NTRK1 and NTRK3. (C): NTRK3 FISH showing an atypical rearrangement pattern (white arrow) with one or more fusion signals and additional 5′ (green) or 3′ (red) signals in the EML4::NTRK3 fusion-positive tumour (case 4). (F): NTRK1 FISH demonstrating an atypical pattern (white arrow) with one fusion signal and a single 3′ (red) signal in the absence of the 5′ (green) signal, suggesting an interstitial deletion on chromosome 1q22 in the LMNA::NTRK1 fusion (case 8). (I, N): Typical wild-type FISH pattern with two fusion signals observed in the BCOR::MAML1 fusion-positive sarcoma (case 3) (I) and in the fusion-negative sarcoma (case 7) (N). Original magnification: 100×.

Figure 2. Visualization of NTRK fusion transcripts detected by RNA-based next-generation sequencing using the Archer FusionPlex Sarcoma Panel. The diagram illustrates the genes and exons involved, genomic breakpoint locations, and predicted functional domains. Horizontal red arrows in panels (A) and (B) indicate the nucleotide locations of the gene-specific primers (GSPs) targeting NTRK3 and NTRK1, respectively, as used in the Archer FusionPlex assay. (A) Anchored multiplex PCR results showing the EML4::NTRK3 fusion transcript involving exon 2 of EML4 and exon 14 of NTRK3. The schematic representation shows the retention of the trimerisation domain (TD/coiled-coil domain) of EML4 and the complete tyrosine kinase domain (TK) of NTRK3. (B) Anchored multiplex PCR results showing the LMNA::NTRK1 fusion transcript involving exon 2 of LMNA and exon 11 of NTRK1. The resulting fusion retains part of the alpha-helical rod domain of LMNA and the complete TK domain of NTRK1.

Figure 3. Overview of the diagnostic workflow applied in the study.

Supplementary Figure 1. Heatmap showing the relative expression levels (log2 scale) of NTRK1, NTRK2, and NTRK3 genes across the sarcoma samples analyzed. The scale bar on the right represents log2-transformed expression values ranging from -6 (low expression) to +6 (high expression). Each row corresponds to a gene and each column to a sample. Increased NTRK3 expression is observed in BCOR::MAML1 fusion-positive case. The heatmap was generated using Archer Analysis software Pipeline Version: 7.1.0-14 Web Application Version: 7.1.0-38.

Table I. Semiquantitative scoring system used for interpreting pan-TRK immunohistochemical staining in tumor cells.

Percentage of stained tumor cells (%)	Interpretation
1-10	Rare Positive
11-49	Focal Positive
> 50	Diffuse Positive

Table II. Summary of histopathological, immunohistochemical and molecular data.

Case ID	Pathologic diagnosis	Pan TRK staining pattern	Pan-TRK subcellular pattern	NTRK FISH	FISH Pattern	Fusion identified by NGS sequencing
1	MPNST	Focal positive	C	Neg	-	No fusion detected
2	Spindle cell sarcoma	Rare positive	C	Neg	-	NT
3	MPNST	Diffuse positive	C+M	Neg	-	BCOR::MAML1
4	NTRK-rearranged spindle cell neoplasm	Diffuse positive	C+M	Pos NTRK3	Break-apart + extra 3’	EML4(e2)::NTRK3(e14)
5	Pleomorphic liposarcoma	Focal positive	C	Neg	-	No fusion detected
6	Spindle cell sarcoma	Diffuse positive	C+M	Neg	-	EWSR1(e7)::NACC2(e3)
7	Spindle cell sarcoma	Focal positive	M	Neg	-	No fusion detected
8	NTRK-rearranged spindle cell neoplasm	Diffuse positive	C+M	Pos NTRK1	Isolated 3’	LMNA(e2)::NTRK1(e11)
9	NTRK-rearranged spindle cell neoplasm	Diffuse positive	C+NM	Pos NTRK1	Isolated 3’	LMNA(e2)::NTRK1(e11)
10	Leiomiosarcoma	Focal positive	C	Neg	-	No fusion detected
11	Dedifferentiated liposarcoma	Focal positive	C	Neg	CNG NTRK1/NTRK2/NTRK3	NT
12	Leiomyosarcoma	Focal positive	C+M	Neg	-	NT
13	Leiomyosarcoma	Focal positive	C	NA	-	NT
14	MPNST	Focal positive	C	Neg	-	No fusion detected
15	Lipofibromatosis	Focal positive	C	Neg NTRK1	NTRK2/NTRK3 NT	No fusion detected
16	Spindle cell sarcoma	Focal positive	C	Pos NTRK1	CNG NTRK1	No fusion detected
17	MPNST	Focal positive	C	Neg	-	NT
MPNST, Malignant Peripheral Nerve Sheath Tumor; C, cytoplasm; M, cell membrane; NM, nuclear membrane, Pos, positive; Neg, negative; FISH, Fluorescence in situ hybridization; NA, not adequate; NT, not tested; CNG, copy number gain; NGS, Next Generation Sequencing; e, exon.