International guidelines on the diagnosis and treatment of NUT carcinoma

Abstract

Nuclear protein in testis (NUT) carcinoma (NC) represents a rare, clinically aggressive cancer defined by pathognomonic NUT Midline Carcinoma Family Member 1 (NUTM1) gene fusions, with bromodomain and extraterminal domain (BET) protein 4 (BRD4)-NUTM1 being the predominant oncogenic driver. Since its description in 1991, gradual advances have clarified the pathologic mechanisms of NC and its diagnostic methods; however, NC treatment remains a significant challenge. Moreover, diagnostic and treatment approaches for this cancer require further validation and standardization. These guidelines were developed by the Chinese Alliance of Research for NC (ChARN) based on current evidence in the literature and incorporate consensus-based input from multiple international experts. They provide comprehensive guidance on NC diagnosis and treatment, covering epidemiology, pathogenesis, diagnostic methods, therapeutic strategies, BET-inhibitor toxicity, palliative care, and prognostic assessment during follow-up. They also emphasize the importance of multidisciplinary team collaboration in NC treatment and recommend prioritizing enrollment in prospective clinical trials for patients. Current mainstays of treatment include surgical resection, radiotherapy, and medical treatment (chemotherapy, targeted therapy, and immunotherapy), although no standard treatment protocol exists. Future research directions include improving diagnostic efficiency, exploring new therapeutic strategies (such as highly selective BET inhibitors, BET-inhibitor combinations, and PROTAC technologies), and recommending basket trials as a research approach for patients with NUTM1 gene fusions.

Graphical abstract

Public summary

• •BRD4:NUTM1 is the fusion oncogene driver in ∼70% of NUT carcinomas.
• •NUT IHC is recommended as an initial screening for malignant tumors exhibiting poorly differentiated phenotype.
• •Existing chemotherapy regimens and BET inhibitors have limited efficacy in NUT carcinoma.
• •For patients with advanced NUT carcinoma, rational clinical trial combinations should be considered.

Introduction

Nuclear protein in testis (NUT) carcinoma (NC) is a rare yet highly aggressive malignancy defined by NUT Midline Carcinoma Family Member 1 (NUTM1) gene rearrangements (15q14). In 1991, two independent research teams reported NC cases characterized by the t(15;19) translocation.^1,2 In vitro studies by French et al.³ led to the pivotal discovery of NC in 2003 as a distinct disease entity driven by the fusion of bromodomain and extraterminal domain (BET) protein 4 (BRD4) and NUTM1. In 2004, the World Health Organization (WHO) classified tumors with t(15;19) translocation as a thymic malignancy and designated it “NUT midline carcinoma,” due to its predominant occurrence in midline organs.⁴ However, subsequent reports revealed NC’s emergence in numerous non-midline organs, leading to its reclassification as the independent entity “NUT carcinoma of the thorax” by the WHO in 2015.⁵ NC exhibits rapid progression and profound resistance to conventional radiotherapy and chemotherapy, with a dismal median survival of 6.5–10.6 months.^6,7,8 Notably, the rarity of NC has hindered large-scale prospective randomized trials, limiting advancements in therapeutic paradigms. Beyond NC, NUTM1 fusions have been identified in NUT sarcomas, hematologic malignancies,⁹ poroma and porocarcinoma,¹⁰ and NUT adnexal carcinoma.¹¹ These findings challenge traditional disease classifications and highlight the potential of basket trials targeting NUTM1-altered malignancies as a novel strategy to explore effective therapies for NC. While the 2021 International Symposium on NC delivered foundational consensus guidelines,¹² subsequent advances in molecular diagnostics and targeted therapies mandate contemporary refinement of evidence-based NC management standards.

The Chinese Alliance of Research for NC (ChARN) was established in 2024 with the support of the International Association for the Study of Lung Cancer Rare Tumors Committee and now comprises 112 member institutions. This collaborative network focuses on advancing research specific to NC in China, encompassing investigations into pathogenesis, diagnostic methodologies, therapeutic strategies, and mechanisms of drug resistance. A cornerstone of ChARN’s efforts is a dual-database system: a retrospective database consolidating NC cases from 2000 to 2024, and a prospective database launched in late 2024 that systematically collects data across multiple organ systems (e.g., lung, pancreas, bone, nasal cavity, parotid gland, nervous system, genitourinary tract, and soft tissues). These databases underpin China’s first real-world study on NC, aiming to align domestic research with global standards. ChARN leverages these datasets to publish clinical NC profiles from China, foster collaborations with academic institutions for mechanistic studies, diagnostic enterprises for biomarker discovery, and pharmaceutical companies for preclinical and clinical drug development. Such multidisciplinary integration is critical given the absence of an international consensus or guidelines for NC management, particularly when adopting and evaluating novel diagnostic technologies and therapeutic approaches. To address this gap, ChARN convened domestic NC experts to assess existing evidence through rigorous discussions, culminating in the development of this guideline.¹³ The initiative seeks to translate emerging diagnostic and therapeutic advancements into tangible clinical benefits, ultimately improving patient survival outcomes.

Incidence and epidemiology

Key point 1: NC predominantly arises in the thoracic and head/neck regions, although non-midline organ origins are also documented. Routine physical screening for patients with NC is not recommended due to the low incidence rates, which occur across young age predominance and show no significant sex predilection. The etiology of NC remains unclear. (Recommendation grade: recommended).

Incidence

The precise incidence of NC remains undefined. Data from Western Australia (1989–2014) estimate an annual incidence of 0.41 per million in children aged 0–16 years¹⁴ A United States (US)-based study utilizing next-generation sequencing (NGS) data suggests NC accounts for 0.06% of all solid tumors.¹⁵ However, a 2022 structured literature analysis of 310 NC cases published since 1991 revealed that 49% of reported cases originated from the US.¹⁶ This not only highlights a significant global under-recognition of this disease and substantial publication bias but also suggests that current incidence rates may be underestimated. Concurrently, with a misdiagnosis rate as high as ∼60%,¹⁷ the true incidence of NC requires validation through large-scale population studies. Given its low prevalence and prohibitive cost-benefit ratio, population-based screening for NC is not recommended.

Epidemiology

NC can occur at any age (reported range from neonates to 84 years)^18,19 but is predominantly observed in adolescents and young adults. A systematic review of 119 NC cases found that nearly 50% of patients were adolescents or young adults aged 14–30 years.²⁰ Two large retrospective analyses of patients with NC reported a median age at diagnosis of approximately 23 years,^6,20 whereas the median age for thoracic-origin patients with NC is around 30 years.²¹ No significant sex-based difference in overall NC incidence has been identified^6,20; however, a male predominance has been noted among NC cases of pulmonary origin.²² Epidemiologic studies have found no definitive associations between NC and smoking, Epstein-Barr virus, human papillomavirus (HPV), or other risk factors.^21,23,24,25 To date, there is no direct etiologic evidence linking specific causative agents to NC.

Pathogenesis

Key point 2: The cellular origin of NC remains elusive. The NUTM1 fusion gene is the primary pathogenic driver of NC, with BRD4 being the most frequent fusion partner. The BRD4::NUTM1 fusion represents the predominant pathogenic variant in NC. Although additional NUTM1 fusion partners have been identified in recent years, the clinical relevance and mechanistic contributions of these novel fusion events to NC pathogenesis remain incompletely characterized.

Due to BRD4::NUTM1 fusion being the predominant variant in NC, it is recommended to focus more attention on the detection of BRD4::NUTM1 variant in undifferentiated carcinoma in clinical practice. (Recommendation grade: strongly recommended).

Cellular origin

The precise cellular origin of NC remains undefined. Clonal fusions of the NUTM1 gene constitute the primary pathogenic driver of NC.³ These fusion events may occur during infancy without requiring cumulative mutations in other genes.¹⁹ Based on the oncogenic parallels between NUTM1 fusions and translocation-associated hematologic malignancies, French et al.²⁴ hypothesized that NC likely arises from a rare, migratory stem-cell-like progenitor population. The expression of the stemness marker SRY-box transcription factor 2 (SOX2) in NC cell lines provides indirect support for this hypothesis.²⁶ However, after analyzing whole-genome and transcriptome sequencing data from three NC cases and the Ty-82 cell line, Lee et al.²⁷ proposed that these rearrangements likely result from a single catastrophic event in otherwise healthy dividing cells. The chromoplexy-like events are likely to be involved in the pathogenesis of NC, resulting in complex DNA fusion breakpoints.²⁷ Hakun and Gu²⁸ further speculated that NC may originate from a rare epithelial cell subset within midline tissues. Recently, two genetically engineered mouse models (GEMMs) of NC have revealed that NC can arise within any of the three germ cell layers, although in both models the vast majority arose from epithelial tissues, predominantly squamous type.^29,30 The discovery of squamous in situ lesions of NC in one of these GEMMs is perhaps the most direct evidence that NC can arise from epithelial cells of squamous lineage.²⁹

Pathogenic fusion genes and downstream mechanisms

BRD4::NUTM1 fusion gene and its pathogenic mechanism

The BRD4 gene is the most frequent fusion partner of NUTM1 in NC. In 2003, French et al.³ first reported that the t(15;19) translocation generates the BRD4:NUTM1 fusion gene. Subsequent large-scale analyses revealed that approximately 70% of patients with NC harbor this fusion.²⁴ The breakpoints for this fusion typically occur upstream of exon 3 in NUTM1,³¹ although variants with alternative breakpoints have also been documented. For example, Stirnweiss et al.³² described a case of NC with the BRD4:NUT (exon15:exon2Δnt1–585) variant, which exhibited extremely poor prognosis (survival of only 79 days post complete resection). In contrast, Kato et al.³³ reported a pulmonary NC patient with the BRD4:NUTM1 (exon10:exon4) variant who survived for 14 years. Cellular studies further demonstrated that the BRD4:NUTM1 (exon11:exon2) variant confers heightened sensitivity to bromodomain inhibitors.³⁴

Current research on NC pathogenesis predominantly focuses on the BRD4::NUTM1 fusion gene. Recently, two independent teams validated the oncogenic potential of the BRD4::NUTM1 fusion using genetically engineered mouse models, confirming its ability to drive NC tumorigenesis.^29,30 This fusion gene was proposed to drive NC tumorigenesis via chromatin binding, disrupting squamous epithelial differentiation, and inhibiting cell-cycle arrest.³⁵ NUTM1 also mediates widespread histone acetylation. As its most frequent fusion partner, the BET protein exhibits broad chromatin accessibility at enhancer regions.³⁶ BRD4, a member of the BET family, consists of two bromodomains, an extraterminal (ET) domain and a carboxyl-terminal domain (CTD).³⁶ Functioning as an epigenetic “reader,” BRD4 recognizes and binds acetylated lysine residues on histones H3 and H4 via its bromodomains while serving as a scaffold for other factors.^37,38 Through its ET domain, BRD4 forms complexes with proteins including BRD3, NSD3, and ZNF, and further recruits and activates the positive transcription elongation factor (pTEFb) via its CTD.^36,39 Activated pTEFb phosphorylates RNA polymerase II, stimulating transcriptional elongation.⁴⁰ The BRD4-NUT fusion protein acts as an epigenetic super-enhancer, leveraging NUT to bind DNA and recruit the histone acetyltransferase p300. This interaction generates nuclear hyperacetylated megadomains, which drive the sustained expression of oncogenic transcription factors such as MYC, p63, and SOX2, thereby maintaining NC cells in an undifferentiated proliferative state.^41,42 Among these, MYC is a critical downstream target of NUTM1-BRD4, which is essential for preserving the undifferentiated and proliferative phenotype.⁴³ A schematic diagram of these pathologic mechanisms is provided in Figure 1.

Figure 1.

Pathogenic mechanisms of NUT-fusion oncoprotein

Other fusion genes

Of those with NUTM1 fusions, approximately 15% of patients with NC harbor BRD3 as the fusion partner, while 6% have NSD3 fusion variants. BRD3 and BRD4 share high structural homology.⁶ Other rare fusion types include ZNF532::NUTM1,^44,45 ZNF592::NUTM1,⁴⁶ and BRD2::NUTM1.⁴⁷ Mechanistic studies reveal that BRD3, NSD3, ZNF532, and ZNF592 interact with BRD4, serving as critical components of the BRD4::NUT oncogenic complex. These fusion proteins retain the functional capacity to recruit BRD4 and drive chromatin remodeling, mirroring the oncogenic activity of the canonical BRD4::NUT complex.^31,45 Clinically, distinct fusion partners correlate with divergent prognoses and therapeutic responses. Retrospective data indicate that non-thoracic patients with NC with BRD3 or NSD3 fusions exhibit significantly prolonged overall survival (OS) (median OS: 36.5 months).⁶ Notably, the BRD2::NUTM1 fusion cases may demonstrate heightened chemosensitivity and reduced tumor aggressiveness.⁴⁷

With the widespread clinical application of NGS, novel NUTM1 fusion partners have been increasingly identified in solid tumors, including NSMCE2,⁴⁸ BCORL1,⁴⁹ MYXD1,⁴⁹ MYXD4,^15,49,50 MGA,^15,51 and CIC.^52,53 Tumors harboring these fusions exhibit variable histopathologic features: some of these fusion genes do not exhibit typical NC histology and should be referred to as non-NC NUT fusions (such as MGA or CIC). However, the biological relevance of these fusion partners to NC pathogenesis remains incompletely characterized. Notably, NGS analyses reveal that tumors with such fusions lack additional oncogenic driver mutations or tumor-suppressor gene-inactivation events and often display a low tumor mutation burden,^6,27,54 suggesting that NUTM1 fusions may act as the sole driver in these malignancies. Indeed, the observation that NC can rapidly form in GEMMs upon formation of the BRD4-NUTM1 fusion indicates that BRD4-NUT can alone drive malignant transformation and progression of NC. In addition to the NUTM1 fusion, NGS studies have further demonstrated that mutations in DNA-repair genes are the most frequent somatic alterations in NC.³⁴ DNA- and RNA-based NGS testing of 50 patients with NC showed that 26% of patients had co-mutations of epigenetic or cell-cycle pathways, and there was no difference in the type of co-mutation between different fusion partners.⁵⁵ Recent investigations highlight the critical role of Enhancer of Zeste Homolog 2 (EZH2) in maintaining NC proliferation and blocking differentiation. Mechanistically, EZH2 and BRD4::NUT regulate distinct genomic programs in NC cells, with combined inhibition of both pathways synergistically downregulating proliferation-associated genes and inducing expression of tumor-suppressor genes, namely CKDN2A.⁵⁶

Diagnosis

Clinical diagnosis

Key point 3: The clinical symptoms and signs of patients with NC lack specificity. Notably, NC is not strictly confined to midline organs. Although the thoracic cavity and head/neck region are the most common primary sites of NC, the clinical manifestations are non-specific and primarily related to rapidly growing tumors at these anatomical locations.

Therefore, if rapidly progressing tumor masses are found in midline regions such as the head/neck and thoracic cavity on imaging, NC should be given priority consideration and confirmed by timely biopsy. (Recommendation grade: recommended).

A systematic review of 119 patients with NC demonstrated that the most common primary sites are the thoracic cavity (42 cases, 35.3%), followed by the head/neck region (40 cases, 33.6%).²⁰ Other rare primary sites include the salivary glands,⁵⁷ lacrimal sac,⁵⁸ bladder,⁵⁹ thyroid,⁶⁰ liver and pancreas,¹⁹ ovary,⁶¹ brain, stomach, kidney, soft tissue, bone,⁴⁹ pelvis,⁶² adrenal gland,⁶³ and skin,⁶⁴ indicating that NC is not confined to the traditionally defined midline organs. Symptoms of NC vary depending on the primary site, and most patients are diagnosed at advanced stages with non-specific manifestations attributable to rapid tumor growth. For thoracic-origin NC, common symptoms include cough (76%), wheezing (35%), chest tightness (35%), dyspnea (35%), chest pain (35%), shoulder pain (35%), back pain (35%), hemoptysis or blood-tinged sputum (29%), and fever (18%).⁶⁵ Head/neck-origin NC predominantly arises in the nasal cavity/paranasal sinuses (73.4%), followed by the salivary glands (11%), larynx (7.3%), pharynx (4.6%), and oral cavity (3.7%).⁶⁶ These patients present with mass-related symptoms such as rhinorrhea, epistaxis, nasal obstruction, proptosis, visual impairment, dysphagia, or pain. Non-specific systemic symptoms such as fever and weight loss are occasionally observed.⁶⁷ There are no classic paraneoplastic or other presenting symptom that would help in early diagnosis.

Laboratory testing

Patients with NC lack specific laboratory biomarkers. Limited studies suggest that elevated levels of lactate dehydrogenase (LDH), C-reactive protein (CRP), and alpha-fetoprotein (AFP) may be observed in patients with NC. In a cohort of 35 patients with NC reported by Kloker et al.,⁶⁸ the median LDH and CRP levels at initial diagnosis were significantly higher than normal ranges, with even greater elevations observed in metastatic NC cases. Additionally, two case reports documented patients with NC with markedly elevated AFP levels who were initially misdiagnosed with mediastinal seminoma.^69,70

Imaging diagnosis

Key point 4: Comprehensive imaging evaluation targeting common metastatic sites should be performed for NC based on the primary tumor location. Contrast-enhanced computed tomography (CT) is recommended for thoracic and abdominal lesions, while contrast-enhanced magnetic resonance imaging (MRI) is preferred for head/neck lesions. Positron emission tomography (PET)-CT should be recommended in patients with non-metastatic or oligometastatic disease that could benefit from complete surgical resection (Recommendation grade: recommended).

The diagnosis of NC through imaging requires multimodal techniques to screen for metastatic sites, yet its imaging features remain non-specific. Most patients with NC present with advanced-stage disease involving multiple metastases at diagnosis. Comprehensive imaging evaluation using CT, MRI, and PET-CT is essential to identify common metastatic sites and confirm diagnosis. The imaging features for NC are not specific to this malignancy alone, with primary lesions often presenting as aggressive masses demonstrating strong invasiveness into adjacent structures.⁷¹ Thoracic NC typically manifests as mediastinal masses contiguous with enlarged hilar and mediastinal lymph nodes, while bone is the most common extrathoracic metastatic site.²¹ Notably, PET-CT has been reported to detect multifocal bone metastases in patients with NC, whereas bone scintigraphy may yield false-negative results,⁷¹ suggesting limited sensitivity of bone scans for osseous metastases in NC. Beyond bone, the liver, breast, retroperitoneum, soft tissues, and adrenal glands are also frequent metastatic sites.⁷² During imaging examinations, it is critical to meticulously evaluate the aforementioned sites for the presence of metastatic lesions. Due to the rarity of NC, no standardized staging system specific to NC has been established. Current clinical practice typically adopts the TNM staging system based on the primary tumor location.

Pathologic diagnosis

Key point 5: NC is a poorly differentiated epithelial malignancy that lacks pathognomonic histologic features. It is recommended to perform initial screening with NUT immunohistochemistry (IHC) for poorly differentiated malignancies by IHC, followed by confirmation using fluorescence in situ hybridization (FISH) or NGS to detect NUTM1 fusion variants. While NC may exhibit some characteristic histologic features such as focal abrupt keratinization, these findings are non-specific and overlap with other poorly differentiated tumors such as HPV-associated basaloid squamous cell carcinoma.

To reduce the risk of underdiagnosis, NUT IHC is recommended as an initial screening for malignant tumors exhibiting poorly differentiated phenotype, especially in thoracic and head/neck carcinoma. FISH confirmation or NUTM1 fusion partner analysis via deep sequencing (DNA ± RNA) can be helpful when feasible and is mandatory if the histology is not consistent with a poorly differentiated carcinoma. (Recommendation grade: recommended).

Morphologic characteristics

NC typically consists of a monotonous tumor cell population that lacks distinctive architectural patterns in most cases. The neoplastic cells exhibit high homogeneity with undifferentiated or poorly differentiated features, characterized by high nuclear grade and elevated mitotic activity. Focal abrupt keratinization, manifesting as sharply demarcated nests of small round cells with abundant keratinized cytoplasm and pyknotic nuclei without transitional zones, serves as a hallmark morphologic feature. However, this finding is observed in only one-third of NC cases⁶ and overlaps with basaloid squamous cell carcinoma and HPV-associated oropharyngeal squamous cell carcinoma.^73,74 Another common feature is the presence of clear peritumoral halos surrounding neoplastic cells, which may represent “fried-egg-like cells” containing abundant clear cytoplasm.³¹ NC frequently mimics the morphologic appearance of other malignancies, including squamous cell carcinomas, high-grade neuroendocrine carcinomas, other poorly differentiated carcinomas, and mesenchymal small round blue cell tumors.^75,76,77 Notably, NC arising from salivary glands or soft tissues may exhibit prominent myoepithelial differentiation,^49,52,78 further complicating diagnosis based solely on histomorphology. Figures S1–S5 provide microscopy images of H&E-stained NC tissue.

Immunohistochemistry

The diagnosis of NC relies on raising clinical suspicion and detecting NUT protein overexpression using the C52/B1 NUT-specific monoclonal antibody and then NUTM1 rearrangements via FISH or sequencing methods.⁵⁹ According to the WHO classification criteria, diffuse nuclear speckled staining (defined as >50% tumor nuclei positivity) is considered sufficient to diagnose NC.⁵ NUT IHC demonstrates 87% sensitivity, 100% specificity, 99% negative predictive value, and 100% positive predictive value.⁵⁹ When NUT staining shows focal positivity (<50% tumor nuclei) or negativity despite high clinical suspicion, FISH or molecular testing is mandatory to confirm NUTM1 rearrangements. One potential diagnostic pitfall in NUT IHC is in germ cell tumors, which can express NUT protein in a small (5%–10%) subset of cells.

Apart from NUT protein expression, the NC IHC profile lacks specificity and often demonstrates squamous differentiation characteristics. NC classically may express epithelial markers including cytokeratin, CK5/6, P63, P40, EMA, MUC1, SOX2, and MYC.^{21,26,43,79,80} Neuroendocrine markers, including synaptophysin, chromogranin, and CD56, are also observed in some NC tumors, particularly in cases originating from the lung or thyroid, where TTF-1 and PAX8 expression, respectively has mainly been documented.^21,81,82,83 Similarly, SOX2 and MYC are not specific to NC,^84,85,86,87 and some cases may lack squamous differentiation with negative expression of p40 and p63.^88,89 CD34 is commonly expressed in hematopoietic stem cells, vascular endothelial cells, and soft-tissue tumors and is frequently observed in NC,⁹⁰ while PD-L1 expression varies among patients.^60,91 Other rare IHC markers are occasionally positive in NC, including Ewing sarcoma markers (FLI1, CD99, and cytoplasmic glycogen⁹²), melanoma marker (PRAME⁸³), germ cell tumor markers (CD30, PLAP, and SALL4⁹³), muscle markers, and hematolymphoid markers (CD30,⁹³ CD43, CD138, and CD45⁹⁴). In summary, IHC markers demonstrate significant heterogeneity and non-specificity in NC, necessitating cautious interpretation. Figure S6 shows NUT IHC staining.

Molecular diagnosis

The detection of NUTM1 rearrangements can be performed using FISH with split probes targeting the NUTM1 breakpoint at 15q14, RT-PCR, or NGS. Interpretation of FISH results is subject to some degree of subjectivity, whereas RT-PCR offers higher objectivity and sensitivity compared to FISH. However, both FISH and RT-PCR are limited by probe design and cannot detect unknown NUTM1 breakpoints.^59,95,96 Figures S7 and S8 contain FISH results for NC. In cases where NUT IHC is positive but FISH is negative, further confirmation should be pursued using FISH with full-length NUTM1 probes or NGS.³⁶

DNA-based NGS has identified more NC cases, but its sensitivity is constrained by insufficient coverage of intronic regions, which may fail to capture all NUTM1 and BRD4 breakpoints.³¹ RNA-based NGS enhances fusion gene detection rates and better reflects true protein-functional fusion events,⁹⁷ having identified multiple NUTM1 fusion variants that were undetectable by DNA-based NGS.^47,98,99,100 Given that NC has significant histologic and immunophenotypic heterogeneity, which often mimics other tumors, DNA or RNA sequencing to identify specific NUTM1 fusion partners has become critical when clinical and histopathologic findings are discordant. It is recommended to prioritize NGS RNA panels or transcriptome sequencing because NGS DNA panels often do not adequately cover the large intronic regions within which NUTM1 breakpoints occur. RNA-based NGS panels demonstrate higher sensitivity, particularly when DNA-based NGS yields negative results despite positive IHC results. Concurrent extraction of DNA and RNA from formalin-fixed paraffin-embedded sections for parallel sequencing is advised when feasible. A comparison of the advantages and limitations of current NUTM1 fusion diagnostic methods is summarized in Table 1.

Table 1.

Advantages and disadvantages of detection methods for NUTM1 fusions

Method	Sample requirements	Rare fusion detection	Efficiency	Sensitivity	Specificity	Affordability	Turnaround time
IHC	low	no	high	high	high	high	quick
FISH	low	no	high	high	high	moderate	quick
RT-PCR	low	no	high	high	high	moderate	moderate
DNA-based NGS	moderate	yes	moderate	medium to low	high	low	slow
RNA-based NGS	high	yes	moderate	high	high	low	slow

Diagnostic value of distinction of NC from its mimickers

Due to the non-specific histologic and immunohistochemical features associated with NC, the differential diagnostic considerations can vary depending on the primary site. For the most common thoracic NC cases, distinction is required from primary pulmonary squamous cell carcinoma, large cell carcinoma, and small cell carcinoma; if presenting as a small round cell tumor, it should be distinguished from Ewing sarcoma, CIC-rearranged sarcoma, and epithelioid sarcoma. Additionally, differentiation is necessary from hematopoietic/lymphoid malignancies, malignant melanoma, germ cell tumors, and thoracic SMARCA4- or SMARCB1-deficient undifferentiated tumor.⁷⁹ For extrathoracic primary NC, differential diagnoses include synovial sarcoma, GLI1-rearranged tumors, undifferentiated carcinoma, and desmoplastic small round cell tumor.⁸⁰ At the molecular level, NUTM1 gene rearrangements are not exclusive to NC, as they have also been detected in some skin tumors with apocrine and eccrine differentiation (e.g., poroma and porocarcinoma),¹⁰ hematopoietic/lymphoid malignancies,⁹ and sarcomas.

The relationship between different fusion variants and NC remains incompletely understood (see the “pathogenic fusion genes” section) and requires further investigation. Furthermore, NC must be distinguished from NUT sarcoma and CIC::NUTM1 sarcoma, which have specific features summarized in Table 2. Given the rarity of NC, which limits the diagnostic experience for most pathologists, consultation with experts at large medical centers is recommended. For cases with unclear fusion variants, submission to the Molecular Tumor Board (MTB) for further discussion is advised.

Table 2.

Comparison of NUTM1 carcinomas, NUTM1-rearranged sarcomas, and CIC::NUTM1 sarcomas

Characteristics	NC	Sarcomas with NUTM1 rearrangement (non-CIC)	CIC::NUTM1 sarcomas
Age	adolescents/young adults and middle-aged	broad age range (children to adults)	younger adults (20–40 years)
Primary sites	midline structures: mediastinum, head/neck, lungs	soft tissues (extremities, trunk)	soft tissues (trunk, extremities, retroperitoneum)
Histology	undifferentiated cell sheets with abrupt squamous differentiation/keratinization	small round blue or spindle cells; variable differentiation	round/epithelioid cells, necrosis, high mitotic activity
Immunohistochemistry	NUT+ (diffuse nuclear staining with stippled pattern), cytokeratin+, p40/p63+ (squamous markers), CD34−	NUT+, variable CD99+, epithelial markers (e.g., cytokeratin) usually negative	NUT+, WT1 (nuclear)+, ETV4+, may express CD99
Fusion genes	NUTM1 fused with BRD4 (70%), BRD3 (15%), or other partners (e.g., NSD3, ZNF532)	NUTM1 with non-BRD partners (e.g., ZNF532, MGA family [MXD1–4, MNT])	CIC fused with NUTM1 (rare)
Genetic alterations	chromosome 10q monosomy commonly observed	complex karyotypes; no specific pattern	CIC wild-type allele loss; 19p13.2 alterations
Symptoms	rapidly growing mass, pain, airway obstruction, dysphagia	painless or painful mass, localized swelling	aggressive mass, possible systemic symptoms (e.g., fever)
Metastasis	early distant spread (lungs, bones)	frequent lung/liver metastases	high metastatic potential (lungs, brain)
Prognosis	extremely poor (median survival: 6–12 months)	poor (similar to other high-grade sarcomas)	very poor (resistant to conventional therapy)

Treatment

There is currently no standard treatment regimen recommended for patients with NC. It is recommended to adopt a multi-disciplinary treatment model that involves specialists from medical/surgical oncology, molecular pathology, and radiology to create comprehensive diagnostic and therapeutic plans based on the TNM staging for the primary tumor site. Concurrently, patients should be encouraged to participate in prospective clinical trials. Establishing a unified institutional database to incorporate patient information for real-world sample repositories is recommended to enhance the understanding of this disease.

Surgical treatment

Key point 6: A multidisciplinary model should be adopted to evaluate surgical options. For resectable patients with NC, radical surgical resection should be performed while preserving critical functions of the primary site whenever possible. Complete surgical resection is an independent positive prognostic factor for prolonging survival in patients with NC.

Given the complexity of NC diagnosis and treatment, it is recommended to evaluate the patient performance, surgical time point and methods, and so forth through a multidisciplinary team in order to establish the optimal surgical decision. (Recommendation grade: recommended).

For resectable locally advanced NC, complete surgical resection is associated with improved survival rates and OS.^8,101 In a case series of 118 patients with NC undergoing various therapeutic approaches, the 11 long-term survivors all presented initially with non-metastatic NC, and all underwent radical surgery.¹⁰² A retrospective analysis of 40 head and neck NC patients¹⁰¹ demonstrated that over half received surgical intervention. Initial surgical treatment (with or without postoperative chemoradiotherapy) and negative surgical margins were significant independent predictors of improved OS: the 2-year OS rate reached 50% in surgically treated patients versus 7% in non-surgical patients (p = 0.003); patients with negative margins achieved an 80% 2-year OS rate, compared to 44% for those with gross total resection but positive margins and 37% for those undergoing debulking surgery.¹⁰¹ These results suggest that even incomplete resection may confer survival benefits in head and neck NC, although further large-scale validation is required. A literature review of 50 thoracic NC patients²² revealed that only ∼20% were eligible for surgery. Yuan et al.²² evaluated 55 patients among 35 literature articles and found that 11 patients had surgical resection of the primary lesion, of whom seven patients undergoing surgery alone had a mean OS of 3.54 months. For the other four patients who received neoadjuvant therapy and postoperative adjuvant chemoradiotherapy, three of them were still alive at the time of reporting, and one of them had an 8-month OS. Patients receiving surgery alone had a median OS of 3 months, while those undergoing surgery combined with neoadjuvant/adjuvant therapy achieved longer survival.²²

Given the highly aggressive nature and complexity of NC, surgical intervention should ideally be performed at comprehensive centers with multidisciplinary expertise in oncologic surgery. Preoperative multidisciplinary discussions are critical in evaluating the potential integration of neoadjuvant or adjuvant therapies.⁸⁰ The primary surgical objective is to achieve complete tumor resection. The surgical approach should follow standard protocols for carcinomas of the same anatomic site, encompassing resection of the primary tumor, adjacent involved tissues, and regional lymph node dissection to minimize recurrence risk. Functional preservation of speech, swallowing, and respiration should be prioritized while ensuring negative margins. For head and neck NC cases, regional lymph node dissection is recommended even in clinically N0 patients.⁸⁰ In thoracic NC cases, minimally invasive techniques such as endoscopic or robot-assisted approaches may be employed to reduce postoperative recovery time and complications, provided they are deemed feasible by experienced surgeons. In cases of positive or close margins, postoperative salvage therapy should be initiated promptly, guided by the pTNM staging principles established for other primary solid tumors at the corresponding anatomic site.⁸⁰

Radiotherapy

Key point 7: Radiotherapy can be utilized across all NC treatment stages. For non-resectable patients with NC with good performance status, concurrent chemoradiotherapy targeting primary lesions in the head/neck or thorax, as well as radiotherapy applied at various treatment phases, positively impacts prognosis. However, there is currently no standardized radiation dose or field for NC. Concurrent chemotherapy regimens should align with those used for other solid tumors at corresponding anatomic sites. Advanced radiotherapy techniques are strongly encouraged to escalate tumor dose delivery while minimizing damage to healthy organs.

For unresectable NC patients with good performance status, radiotherapy can be utilized in all treatment stages of patients with NC; the radiation dose and field for other tumors in this region can be used as reference. (Recommendation grade: recommended).

During multimodal treatment for NC, radiotherapy is often administered as preoperative neoadjuvant therapy or postoperative adjuvant therapy, typically combined with surgery or chemotherapy. However, there is no standardized radiation dose or modality for these approaches. During treatment, the integration of radiotherapy⁶ and radiation doses exceeding 50 Gy²⁰ have been associated with improved prognosis. A retrospective study of 119 patients with NC demonstrated that those receiving radiotherapy achieved 1-year and 5-year OS of 37.7% and 11%, respectively, compared to 13% and 0% in non-radiotherapy patients. The survival benefit of radiotherapy was more pronounced in head and neck patients with NC but limited in primary mediastinal NC cases.²⁰ Similar to other adult epithelial-derived tumors, radiation doses should be adjusted based on local tumor infiltration (e.g., margin status and perineural and vascular invasion).⁸⁰

Current literature reports radical radiotherapy doses ranging from 50 to 70 Gy, adjuvant doses from 24 to 66 Gy, and palliative doses of approximately 20–30 Gy,²⁰ although no unified recommendations exist. The European NC guidelines recommend delivering 65–70 Gy to the gross tumor volume and involved lymph nodes, with 50–54 Gy prophylactic irradiation to elective nodal regions.⁸⁰ For head and neck patients with NC with clinically N0 status (pathologically unconfirmed), systematic nodal irradiation should be considered, including bilateral retropharyngeal and levels II–V lymph nodes.⁸⁰ However, this recommendation is based on insufficient (low) evidence and requires further validation through large-scale real-world data.

Concurrent chemotherapy (typically platinum-based) regimens for radical radiotherapy should follow tumor-site-specific protocols.¹⁰² Limited case reports suggest there is potential benefit from combining radiotherapy with anti-angiogenic agents such as anlotinib.¹⁰³ or apatinib.¹⁰⁴ Advanced radiotherapy techniques—including three-dimensional conformal radiotherapy, intensity-modulated radiotherapy, and proton radiotherapy—are encouraged to minimize toxic effects in healthy tissue. For example, an 84-year-old man with head and neck NC completed a 7-week proton radiotherapy course with only mild adverse effects (skin burns and xerosis) and achieved tumor control with an OS exceeding 8 months.¹⁸ In contrast, adolescent patients with NC (7–16 years old) undergoing concurrent chemoradiation had poor outcomes, with a median OS of 8 months (range: 4.5–28.8 months) in five cases. In addition, radiotherapy combined with immunotherapy may serve as a potentially promising therapy option. Ng et al.¹⁰⁵ reported an elderly patient with metastatic sinonasal NC who was treated with pembrolizumab plus palliative radiotherapy and survived for more than 2 years. Haebe et al.¹⁰⁶ also found a patient with advanced metastatic nasopharynx NC who obtained partial response (PR) from the combination of pembrolizumab and chemoradiotherapy. In summary, radiotherapy should be implemented as part of a comprehensive treatment strategy after thorough evaluation of patient fitness and toxicity tolerance. Participation in clinical trials is strongly encouraged to further define existing protocols and explore novel therapeutic approaches.

Pharmacotherapy

Systemic chemotherapy

Key point 8: Existing chemotherapy regimens provide limited clinical benefit for patients with NC. Current regimens demonstrate response rates below 40%, with ifosfamide-based therapies showing marginal advantages in objective response rate (ORR) and progression-free survival (PFS) for non-metastatic NC. However, chemotherapy regimens have yet to improve OS for the patients with NC.

Due to the absence of effective systemic therapeutic approaches and while chemotherapy is still helpful in improving the ORR and PFS of NC, it is recommended that eligible patients with NC can be administered ifosfamide-based chemotherapy. (Recommendation grade: recommended).

NC exhibits poor chemosensitivity. Chemotherapeutic agents that have been used in case reports and case series include anthracyclines, platinum-based agents, alkylating agents, vincristine, gemcitabine, vinorelbine, and taxanes, often used in combination.^{67,92,107,108,109,110} Response rates across all regimens does not exceed 40%,^8,67,111 with transient efficacy and no consensus on optimal regimens. Ifosfamide-based protocols, particularly the Scandinavian Sarcoma Group IX (SSG IX) regimen for Ewing sarcoma, have achieved long-term complete remission (CR) or cure in four pediatric NC cases,^110,112,113 although this efficacy is not universal.¹¹² A retrospective study of 118 patients with NC demonstrated numerically higher, though not statistically significant, 1-year PFS rates with ifosfamide-based therapy compared to platinum-based regimens (59% vs. 37%, p = 0.3), although OS remained comparable. Notably, among 56 patients with advanced NC (80% with thoracic primaries), ifosfamide-treated cases (n = 8) showed superior ORR over platinum-treated cases (n = 36) (75% vs. 31%).¹⁰² While ifosfamide-based therapy demonstrates marginal ORR and PFS benefits in non-metastatic disease, this does not translate into OS improvements in metastatic disease. Thus, conventional chemotherapy remains suboptimal for NC, underscoring the urgent need for novel therapeutic approaches. In addition, patient age, ECOG performance status, and local approval and cost coverage should inform the decision between platinum- or ifosfamide-based regimens and the potential addition of immunotherapy.

Targeted therapy

Key point 9: The efficacy of BET-inhibitor monotherapy in patients with NC is below 30%, with most patients experiencing response durations of only 2–3 months. High incidences of gastrointestinal toxicity and severe thrombocytopenia are observed.

Due to low efficacy and gastrointestinal toxicity as well as severe thrombocytopenia of BET-inhibitor monotherapy in patients with NC, it is recommended that patients with NC can be treated with a BET inhibitor in combination with other regimens. (Recommendation grade: recommended).

BET inhibitors

BET inhibitors directly bind to the dual bromodomains of BRD4, displacing BRD4::NUTM1 from chromatin, thereby downregulating the expression of oncogenic driver genes such as MYC,^42,43 TP63,⁴² and SOX2.²⁶ This leads NC cells to experience rapid differentiation, growth arrest, and apoptosis.¹¹⁴ Several BET inhibitors have undergone clinical investigation for hematologic malignancies and solid tumors (Table 3). There are no BET inhibitors currently approved by the US Food and Drug Administration (FDA) or European Medicines Agency (EMA) (and the Chinese National Medical Products Administration [NMPA]).

Table 3.

Summary of published BET inhibitor clinical trials

BET inhibitor	Targets	Phase	No. of patients with NC	No. of PRs	PFS (months)	Fusion protein status in patients with PR	Status
Birabresib (MK-8628/OTX015)115	BRD2/3/4/T	IB	10	3	1.2–8.4	unknown: 3	completed
Molibresib (GSK525762)116	BRD2/3/4/T	I–II, part 1	19	4	not reported	BRD3: 3; unknown: 1	completed
Molibresib (GSK525762)117	BRD2/3/4/T	I–II, part 2	12	3	not reported	unknown: 3	completed
BMS-986158118	BRD4	I/IIa	7	1	14	BRD4: 1	completed
RO6870810119	BRD2/3/4/T	I	8	2	3.1	NSD3: 1; BRD3: 1	completed
ODM207120	BRD4	I	4	0	–	–	completed
BI894999121	BRD4	I	42	3	1.7–1.9	unknown: 3	completed

Birabresib is an oral BET inhibitor targeting BRD2/3/4/T, demonstrating preclinical activity in NC and other tumor types. Among the first four patients with NC treated with birabresib via compassionate use at a daily dose of 80 mg, two achieved rapid remission, while the other two achieved disease stabilization.⁵⁴ In subsequent phase I clinical trials of birabresib for solid tumors involving ten patients with NC, three achieved PR with durations ranging from 1.4 to 8.4 months; however, the gene fusion status of these three patients was not reported.¹¹⁵

Molibresib (GSK525762) is a small-molecule inhibitor targeting BET-family proteins (BRD2, BRD3, BRD4, and BRDT). It eliminates chromatin recruitment of CDK9, a component of pTEFb, in a BRD2–4-dependent manner, thereby suppressing MYC and MYC-dependent transcriptional activity.¹²² In its phase I clinical trial, which included the largest cohort of patients with NC to date (n = 19), 13 patients received doses exceeding 60 mg daily. Among them, four achieved PR (two confirmed, including three with BRD3:NUTM1 fusions), eight achieved stable disease (SD), and four maintained PFS for up to 6 months.¹¹⁶ In a disease-specific expansion cohort of the same study, an additional 12 patients with NC were treated, with three achieving PR (one confirmed) and four achieving SD.¹¹⁷

BMS-986158 features a novel carboline-based structure that effectively inhibits BRD4 and C-MYC expression through favorable interactions in aqueous environments. Preclinical studies have demonstrated its potent cytotoxic effects in both hematologic malignancies and solid tumors.¹²³ In a phase I trial involving patients with various solid tumors, seven patients with NC were enrolled (six in cohort A and one in cohort C). Only one NC patient harboring a BRD4:NUTM1 fusion in cohort A achieved PR (receiving a dose-escalation regimen of 0.75, 1.25, 2.0, 3.0, and 4.5 mg orally for 5 days followed by 2 days off).¹¹⁸ Despite developing grade IV thrombocytopenia and grade II herpes zoster during treatment, the patient exhibited gradual symptom improvement and achieved a 58.1% tumor regression at 41 weeks, indicating deep remission. Unfortunately, the patient returned to their home country at 59 weeks and was lost to follow-up during the COVID-19 pandemic.⁹⁹ Subsequent NGS analysis revealed a BRD4_{c.1213-1G>T (splice)} mutation in this patient, which researchers hypothesized to reside on the wild-type BRD4 allele, potentially explaining the enhanced BET-inhibitor efficacy.⁹⁹

Distinct from oral BET inhibitors, RO6870810 is a subcutaneously administered, non-covalent small-molecule BET inhibitor sharing core structural similarities with JQ1 but optimized for improved solubility, metabolic stability, and reduced serotonin receptor interactions.¹¹⁹ In its phase I trial involving eight patients with NC, two achieved PR and four achieved SD, yielding an ORR of 25% and a median PFS of 94 days. The two PR cases included one with an NSD3:NUTM1 fusion (thoracic NC, response duration 797 days) and another with a BRD3:NUTM1 fusion.¹¹⁹

ODM-207 is a structurally distinct BET inhibitor (compared to JQ1) exhibiting minimal cross-reactivity with non-BET bromodomains. However, it failed to demonstrate efficacy in four enrolled patients with NC: three experienced rapid disease progression within days of enrollment, while one showed negligible plasma drug exposure.¹²⁰ Its toxicity profile mirrored that of previously reported BET inhibitors, with no novel adverse events observed.

BI894999 is a new type of oral BET inhibitor, which has demonstrated strong antitumor activity in clinical studies. In a phase I trial involving patients with various solid tumors, 42 patients with NC were enrolled. Among the 42 patients with NC, one achieved CR, two achieved PR, and 13 achieved SD. Patients with NC had a median PFS of 6.9–7.8 weeks and median OS of 6.6–15.4 weeks.¹²¹

In sum, across the agents trialed, the efficacy of BET-inhibitor monotherapy in patients with NC is approximately 20%–30%. Apart from BMS-986158 achieving efficacy in one BRD4:NUTM1 fusion case, other PRs occurred in patients with NSD3 or BRD3 fusions. Whether different fusion variants influence BET-inhibitor sensitivity requires further investigation. The transient efficacy of BET inhibitors in NC (typically 2–3 months) may relate to incomplete displacement of BRD4:NUTM1 from chromatin.¹²⁴ Clinically used BET inhibitors non-selectively target all BET proteins (BRD2, BRD3, BRD4, and BRDT), potentially contributing to their high toxicity profile. Notably, severe thrombocytopenia is a common adverse event, with grade 3–4 thrombocytopenia reported in 11%–43% of clinical trial participants. Preclinical studies suggest that BET inhibitors induce thrombocytopenia via suppression of BRD2/3/4-mediated GATA1 expression, positioning platelet decline as a potential pharmacodynamic biomarker for BET-inhibitor activity.

HDAC inhibitors

Histone deacetylases (HDACs) regulate chromatin structure and gene expression. HDAC inhibitors increase histone acetylation, reactivate cell-cycle control and differentiation genes, and counteract BRD4:NUTM1-mediated oncogenesis.

Vorinostat is an oral inhibitor targeting HDAC1, HDAC2, HDAC3 (class I), HDAC6, HDAC7 (class II), and HDAC11 (class IV). Case reports indicate that a 17-year-old thoracic NC patient treated with radiotherapy combined with vorinostat 300 mg daily achieved shrinkage of lesions outside the radiation field after continuing vorinostat 300 mg daily for 4 weeks post radiotherapy. The dose was subsequently reduced to 200 mg daily, but it was discontinued due to grade III thrombocytopenia. Efficacy evaluation revealed an OS of 10 months.¹²⁵ Another case involved an 11-year-old patient with salivary gland NC who received neoadjuvant chemotherapy combined with vorinostat, resulting in significant tumor regression. Surgical specimens indicated residual tumor activity was less than 5%, but the primary lesion progressed postoperatively after approximately 3 weeks.¹²⁶ These reports suggest the potential efficacy of vorinostat in patients with NC; however, no clinical trials specifically targeting patients with NC have been conducted to date. In NC xenograft models, the non-selective HDAC inhibitor panobinostat demonstrated tumor growth suppression comparable to that obtained with BET inhibitors, and combination therapy with BET inhibitor and panobinostat improved both survival and tumor growth inhibition in mice.¹²⁷ CUDC-907 is a novel small-molecule dual inhibitor of HDAC and phosphoinositide 3-kinase. Preclinical studies demonstrate its potent suppression of MYC expression and protein stability, exhibiting antitumor activity in multiple MYC-driven tumor models, including NC.¹²⁸ A 61-year-old NC patient enrolled in the CUDC-907 clinical trial (NCT02307240) received second-line treatment with CUDC-907 (60 mg daily for 5 days followed by 2 days off, 21-day cycles). The treatment was well tolerated, with manageable toxicities primarily including diarrhea and thrombocytopenia. This patient maintained SD for over 32 months and remains on active treatment.¹²⁹ However, detailed results from this trial have not yet been published (NCT02307240). There are no HDAC inhibitors currently approved by the FDA or EMA (and NMPA).

Immune checkpoint inhibitors

Historically, NC has been categorized as a “cold tumor,” due to its low tumor mutational burden. RNA sequencing data from 54 patients with NC showed that the degree of immune cell invasion in NC was significantly lower than that of lung and head/neck squamous cell carcinoma.⁵⁵ However, recent studies have reported high PD-L1 expression in patients with NC with thyroid, sinonasal, or pulmonary origins. Despite this, the efficacy of immunotherapy in these patients remains inconsistent,^{60,130,131,132} suggesting that PD-L1 expression levels lack predictive value for immunotherapy response in NC, and the effectiveness of immune monotherapy in patients with NC remains unclear. CD274 (encoding PD-L1) is a direct transcriptional target of BRD4-mediated regulation, as such PD-L1 upregulation may be due to aberrant transcriptional control rather than in response to activated T cells and local interferon production. Preclinical studies demonstrate synergistic effects between PD-1 inhibitors and BET inhibitors in murine models.^133,134 However, no clinical trials evaluating immune checkpoint inhibitors in NC have been conducted to date, with existing evidence limited to isolated case reports. Moreover, the addition of immune checkpoint inhibitors to first-line chemotherapy showed a trend toward improved OS in a case series.⁶⁸ Further research is required to clarify the role of immunotherapy in NC.

Toxicity management of BET inhibitors

BET-inhibitor-related toxicity management should involve a multidisciplinary team composed of oncologists, pharmacists, nurses, and specialists to formulate treatment plans, educate patients on potential adverse effects, and establish protocols for symptom recognition and reporting. The most common toxicities associated with first-generation pan-BET inhibitors in clinical trials include gastrointestinal disturbances and thrombocytopenia (summarized in Table S1). Regular follow-up assessments and monitoring of hematologic/biochemical parameters are essential to evaluate patient tolerance. For BET-inhibitor-induced thrombocytopenia, pharmacologic interventions include recombinant human thrombopoietin (rhTPO) (recommended when platelet counts fall below 75 × 10⁹/L at a dose of 300 U/kg once daily until recovery) and recombinant human interleukin-11 (rhIL-11) (recommended dose: 25–50 μg/kg subcutaneously once daily for 7–10 days until discontinuation criteria are met). For patients with inadequate response or intolerance to rhTPO or rhIL-11, TPO receptor agonists may be considered. Platelet transfusion remains the most immediate and effective intervention for severe thrombocytopenia with or without bleeding complications. Concurrent with pharmacologic management, patients should adopt lifestyle modifications such as avoiding trauma, hard foods, and excessive straining to minimize bleeding risks. Compared to pan-BET inhibitors, second-generation BD2-selective BET inhibitors demonstrate improved safety profiles; however, unique toxicities have been reported, including hERG potassium channel inhibition by low-dose ABBV-744, which may pose potential cardiovascular risks.^135,136

Palliative supportive care

Cancer pain management

Palliative supportive care is a critical component of NC treatment. As a rare and highly aggressive malignancy, its focus lies in alleviating pain, managing severe comorbidities, addressing emotional distress, and improving quality of life. Due to the large tumor burden in most patients with advanced NC, local compression and pain are common, necessitating prioritized cancer pain management. According to the WHO three-step analgesic ladder, pharmacologic therapy is the first-line approach: weak opioids (e.g., codeine or tramadol) are added for moderate pain, while strong opioids are used for severe pain. Adjuvant medications targeting neuropathic pain mechanisms—most commonly tricyclic antidepressants and anticonvulsants (e.g., gabapentin and pregabalin)—are also recommended.¹³⁷

Palliative therapy for metastatic lesions

For unresectable primary lesions, bone metastases, brain metastases, or liver metastases, localized radiotherapy can achieve symptomatic relief. For patients with mild meningeal irritation symptoms and imaging-confirmed multiple brain parenchymal enhancing lesions, treatment should follow guidelines for solid tumor brain metastases, employing whole-brain radiotherapy (WBRT) with simultaneous integrated boost to parenchymal lesions. A dose of 30 Gy/10 fractions for WBRT is recommended, with parenchymal doses adjusted based on lesion size and normal tissue tolerance, alongside hippocampal avoidance techniques.¹³⁸ Bone metastases may be managed with palliative radiotherapy or bone-modifying agents (no specific drug recommendations exist), which reduce skeletal-related events. Notably, weight-bearing bone metastases require a spinal stability assessment: high-risk lesions warrant surgical intervention, while stable lesions may be treated with radiotherapy alone.

Nutritional and psychological support

Nutritional status can also be assessed using nutritional risk screening. For patients with NC with malnutrition or nutritional risk, intensified dietary counseling to optimize oral intake is prioritized. If oral intake remains insufficient despite counseling, oral nutritional supplements (ONSs) should be initiated. ONSs improve nutritional status and enhance tolerance to radiotherapy or chemotherapy. Dietary strategies should emphasize high-protein intake, fresh vegetables, and fruits to ensure adequate vitamin and mineral supply. If ONSs cannot improve the patient’s nutritional status, fluid replacement support therapy should be provided as appropriate according to the patient’s specific condition.

Psychological support should focus on family involvement. Families are encouraged to communicate actively with patients, fostering confidence in cancer management. Healthcare providers and caregivers must monitor emotional fluctuations, offering reassurance and encouragement. Moderate exercise is recommended to enhance immunity and alleviate anxiety or depression. Establishing peer support groups helps reduce isolation and strengthen resilience through shared experiences. Persistent severe emotional distress warrants referral to professional psychological counseling for targeted interventions.

Follow-up and prognosis

Key point 10: Differences in primary tumor sites and NUTM1 fusion partners may be associated with heterogeneous prognoses among patients with NC. Radical surgery, radiotherapy, non-metastatic disease at diagnosis, and lower initial LDH levels may serve as positive predictors for long-term survival.

Therefore, it is recommended to perform radical surgery and radiotherapy in patients with NC and conduct routine clinical laboratory tests including LDH detection. (Recommendation grade: recommended).

Key point 11: Post-treatment follow-up for patients with NC should include thoracic and/or abdominal CT scans every 3–6 months, or more frequently based on symptoms, of areas of known and possible tumor location. (Recommendation grade: recommended).

The OS for patients with NC may vary depending on distinct driver fusion genes and primary sites. Chau et al.⁶ developed a prognostic model using survival tree regression from the NMC Registry database, stratifying patients with NC into three subgroups: (1) thoracic NC with the poorest prognosis (OS 4.4 months), (2) non-thoracic NC with BRD4 fusion partners showing intermediate prognosis (OS 10 months), and (3) non-thoracic NC with non-BRD4 fusion partners demonstrating the best prognosis (OS 36.5 months). Favorable predictors for improved OS include the absence of distant metastasis at diagnosis, surgery, or radiotherapy; complete surgical resection; and CR after first-line therapy.⁶ However, a separate retrospective analysis found that only chemotherapy and radiotherapy, not surgery, significantly prolonged OS.²⁰ Recent German cohort studies suggest that lower baseline LDH levels in non-metastatic patients may correlate with improved prognosis.⁶⁸

Disease progression monitoring should be guided by clinical signs and symptoms during follow-up. Given the high risk of distant metastases in patients with NC, thoracic and/or abdominal CT scans, brain MRI, and annual PET-CT or bone scans (if feasible) are recommended every 3–6 months during the first 2 years post treatment.

Future directions

Key point 12: Patients with unresectable NC should be prioritized for screening in drug clinical trials. (Recommendation grade: recommended).

Future directions for NC diagnosis and treatment should focus on enhancing diagnostic efficiency, clarifying therapeutic strategies, and expanding technological platforms to achieve comprehensive diagnosis and precision therapy. In terms of improving diagnostic efficiency, NC diagnosis relies on pathologic examination; however, its rarity and histologic heterogeneity often lead to misdiagnosis or underdiagnosis. Therefore, raising awareness among pathologists and clinicians is critical. IHC staining for NUT protein should be performed for any poorly differentiated or undifferentiated tumors. Advanced technological platforms are required for the comprehensive identification of suspected cases: in addition to IHC and FISH assays, RNA-based NGS can validate pathologic diagnoses and discover novel fusion partners. Integrated approaches combining whole-genome sequencing and transcriptome sequencing have become increasingly prevalent, which may facilitate a more thorough understanding of NC’s molecular characteristics. With technological advancements, more NUTM1 fusion partners may be revealed to elucidate downstream mechanisms that are crucial for precise diagnosis and targeted therapy. Based on the preceding sections, we provide a diagnostic and therapeutic flowchart for reference (Figure 2).

Figure 2.

NUT diagnostic and treatment flowchart

In terms of advancing therapeutic strategies, the most promising directions in drug development include: highly selective second-generation BET inhibitors, BET-inhibitor-based combination therapies, downstream target drug development, and novel therapeutic modalities. Second-generation BET inhibitors exhibit selective activity for either the BD1 or BD2 bromodomains, potentially reducing drug-related toxicities. Current investigational agents include the BD2 inhibitor ABBV-744,¹³⁹ BD1 inhibitor BI894999,¹⁴⁰ and BRD4 inhibitor NHWD870 HCL,¹⁴¹ some of which have entered clinical trials for patients with NC. Regarding BET-inhibitor combination strategies, the dual p300 BDi/BETi agent NEO2734 (EP31670) has demonstrated superior antitumor efficacy compared to chemotherapy or BET-inhibitor monotherapy in NC preclinical models¹⁴² and is under clinical evaluation in patients with NC (NCT05488548). The CDK4/6 inhibitor palbociclib has been used with JQ1 to suppress NC cell growth,¹⁴³ and a clinical trial combining the BET inhibitor ZEN3694 with the CDK4/6 inhibitor abemaciclib (CTEP 10509) is ongoing (NCT05372640). For downstream target exploration, CDK9 inhibitors upregulate MYC expression in NC cells, and BET inhibitor-CDK inhibitor combinations have displayed synergistic effects across multiple tumor models.¹⁴⁴

Furthermore, novel therapeutic approaches are being explored, including proteolysis-targeting chimeras (PROTACs), which enable rapid and effective degradation of BRD4 and exhibit significant potential in NC treatment.¹⁴⁵ The BRD4 PROTAC ARV-825 has successfully inhibited BRD4:NUTM1 3T3 cell proliferation in preclinical studies,¹⁴⁶ while RNK05047, a dual BRD4/HSP90-selective PROTAC degrader, is being evaluated in a phase I/II trial for advanced solid tumors and lymphomas.¹⁴⁷ Oncolytic virus therapy combined with BET inhibitors has also shown potential clinical benefits in NC, although confirmation is required.¹⁴⁸ In addition, the immune microenvironment also represents a future direction for optimizing immunotherapeutic approaches.^29,30 The recent development of genetically engineered mouse models for NC will facilitate preclinical testing of immunologic strategies. In summary, the efficacy of existing targeted agents (e.g., BET inhibitors and HDAC inhibitors) has reached a plateau due to toxicity limitations. Until safer and more potent drugs emerge, combining established targeted therapies with novel PROTAC-based approaches may represent an optimized strategy.

Additionally, strategies for designing clinical trials targeting patients with NC require further optimization. NC is an exceptionally rare malignancy, with low incidence and scattered cases, making traditional prospective randomized controlled trials challenging to conduct. Basket trials, which enroll patients with the same molecular characteristics across different tumor types, can effectively expand sample sizes and overcome the limitations of insufficient enrollment in rare disease research.¹⁴⁹ The NUTM1 fusion gene is not only present in NC but has also been identified in various other tumor types, including NUT sarcomas, hematologic malignancies, and sebaceous gland tumors. Although these tumors differ histologically, they share a common molecular feature, namely the NUTM1 fusion. Basket trials will enhance efficiency in evaluating novel NC treatment strategies, address sample-size limitations in this rare disease, accelerate clinical validation of targeted agents, and promote the development of precision medicine. Most importantly, basket trials will provide patients with a pathway for more effective treatment options.

Due to its rarity, there are persistent challenges affecting the diagnostic and therapeutic options available for patients with NC. Future directions in NC diagnosis and treatment should focus on improving diagnostic efficiency, expanding technological platforms, elucidating molecular mechanisms, and optimizing therapeutic strategies.

Funding and acknowledgments

This work was supported by the China Postdoctoral Science Foundation (grant number 2022M723207), the Medical Scientific Research Foundation of Zhejiang Province, China (grant number 2023KY666), Zhejiang Traditional Chinese Medicine Science Fund Project (grant number 2024ZL372), Qiantang Cross Fund Project (grant number 2023-16), the National Natural Science Foundation of China of Zhejiang Cancer Hospital Cultivation Project (grant number PY2023006), the Medical Scientific Research Foundation of Zhejiang Province, China (grant number 2024KY812), the Natural Science Foundation of Zhejiang Province (grant number Q24H160110), the National Natural Science Foundation of China (grant number NSFC82371851), the Science and Technology Foundation of Guizhou Province (Outstanding Young Scientists of Guizhou Province) (grant number Qiankeherencai-YQK(2023)021), the Science and Technology Foundation of Guizhou Province (grant number Qiankehejichu-ZK (2023) General 212), and the Science and Technology Foundation of Guizhou Province (grant number Qiankehechengguo-LC(2025)General 068). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author contributions

C.A.F., Yuanzhi Lu, Q.W., and C.X. participated in the design of these guidelines. Yu Zhang, Q.Z., Y. Hao, J. Luo, and Y.P. conceived the guidelines and participated in their design, and the other authors coordinated and helped to draft the guidelines. All authors read and approved the final manuscript.

Declaration of interests

The authors declare no competing interests.

Published Online: July 31, 2025