Describing the Mutational Characteristics of Myxofibrosarcoma: An AACR Project GENIE Analysis

Discussion

This investigation described the mutational landscape of MFS by aggregating data from the AACR Project GENIE repository. We identified novel and potentially targetable mutations to create a more robust molecular profile for MFS.

Novel mutations. We identified several novel mutations that have not been previously associated with MFS. These include NOTCH3, ALOX12B, SDHA, ETV6, NCOA2 and SOS2. NOTCH3 is implicated in many different cancers and may function either as an oncogene or tumor suppressor depending on the cancer type, stage, and grade (17). It has known associations with pediatric soft tissue sarcomas and osteosarcomas, but our study is the first to suggest its involvement in MFS (18). Interestingly, mutant NOTCH3 also underlies the pathology of CADASIL – a rare, inherited small vessel disease of the brain (19). Through disrupting the normally stable beta-sheet structure, mutant NOTCH3 undermines protein stability, inhibiting the survival and signaling of vascular cells (19). Similar effects may be seen with NOTCH3 disrupting the signaling of other healthy cells throughout the body, leading to tumorigenesis.

ALOX12B mutations, with established roles in cervical, breast, and lung cancer, had yet to be linked to any sarcoma cases until the present study (20). SDHA, while probably best known for its involvement in gastrointestinal stromal tumors, was recently linked to tumorigenesis of fibrosarcomas; our data further describes its involvement in MFS specifically (21). Both ETV6 and NCOA2 have been observed as fusion genes in certain sarcomas (22-25). The ETV6-NTRK3 fusion is a defining feature of infantile fibrosarcoma, while the VGLL2-NCOA2 fusion is observed in pediatric sarcomagenesis, and the MAZ-NCOA2 fusion was just recently described for the first time in a case of small round cell sarcoma of infancy (22-25). Our study could not determine whether the observed ETV6 and NCOA2 mutations were also present as fusion genes, but to our knowledge, this series is the first of its kind to characterize their involvement in MFS. Lastly, we present the first documented account of SOS2 in both MFS specifically and sarcomas more broadly. Collectively, we have documented several novel mutations that will shape the future of both detailed molecular profiling and targeted therapeutics for MFS.

Genetic alterations stratified by sex. Consistent with previous studies, our study cohort consisted of slightly more men than women with MFS (3). Predictably, many of the more common and well-known genetic mutations associated with MFS appeared in both men and women with no significant differences between these two groups, including TP53, RB1, ATRX, CDKN2A, CDKN2B, and NF1 (14-16). Notably, our data also revealed several mutations exclusive to either men or women within our limited sample size. Within our cohort, MAP2K4 and ALOX12B mutations were unique to men, whereas SDHA mutations were unique to women. ETV6 mutations were more frequently observed in men compared to women. MAP2K4 mutations have been previously shown to occur somewhat frequently in all soft tissue sarcomas, including MFS (26). However, as detailed above, we are the first to describe alterations in ALOX12B, SDHA, and ETV6 as being linked to MFS. While the novelty of these mutations certainly invites investigation into new diagnostic and therapeutic targets, taken together, these data further suggest the potential for new targets on the basis of patient sex. However, given the low event counts and failure to survive FDR-correction, these findings should ultimately be considered exploratory pending further validation with a larger cohort.

Genetic alterations stratified by race. While MFS tumors are not known to favor any particular race or ethnicity, our study cohort was predominantly composed of White patients (76.7%), followed by Asian patients (6.4%), Black patients (3.5%), and the remaining being another race or unknown race. Comparison of White patients against non-White patients revealed no significant differences in the incidence of more common MFS-associated mutations, such as TP53, RB1, ATRX, CDKN2A, CDKN2B, and NF1. However, two mutations – NCOA2 and SOS2 – were observed exclusively in White patients. As previously mentioned, this is the first study to associate NCOA2 and SOS2 mutations with MFS to date. Their exclusivity to White patients within our limited data set further adds to the robust molecular profile of MFS from a population-based standpoint, which may be of particular interest in targeting therapies to specific patient populations.

In contrast, NKX2-1 mutations were observed exclusively in non-White patients. Prior research has demonstrated that NKX2-1 alterations are unique to MFS compared to other soft tissue sarcomas, such as leiomyosarcoma, de-differentiated liposarcoma, and undifferentiated pleomorphic sarcoma (15). Our study’s finding that these mutations are uniquely seen in non-White patients may be of additional utility in the targeting of future therapies to certain population groups. Importantly, due to the limited number of samples analyzed in our study and their failure to survive FDR-correction, it is plausible that these race-specific findings are due to chance. Thus, these results require additional validation using different sample pools, such as other gene-sequencing panels or data from various medical centers.

Comparative analysis between Black and Asian patients revealed an incidence of KMT2C mutations in three samples from Black patients compared with zero samples from Asian patients. KMT2C has been previously linked to DNA transcription and repair and interestingly seems to encourage proliferation of certain cancers, such as squamous cell carcinoma of the lung, while protecting against development of others, such as renal clear cell carcinoma (27, 28). Although KMT2C alterations are a marker for certain sarcomas, including osteosarcoma, their incidence in MFS has not been consistently described (29). Its novel occurrence in Black patients from our study’s cohort suggests it may have a yet uncharacterized role in MFS development or progression.

Cell cycle and checkpoint pathway alterations. While the landscape of MFS mutations is heterogeneous, mutations to cell cycle and checkpoint pathway alterations have been previously implicated as a leading genetic anomaly (14-16). Both mutations and deletions of TP53 and RB1 promote tumorigenesis by allowing for unrestricted movement through the cell cycle (14). Our results give further credence to the association of TP53 and RB1 mutations with higher-grade cases of MFS as well as a more rapid recurrence post-treatment (30, 31). Similarly, deletions of CDKN2A and CDKN2B are common, having been identified in 20.5% and 19.48% of our cohort’s tumor samples, respectively. Previous research has found the incidence of CDKN2 mutations to be between 16%-30% (15, 16). Collectively, these findings are consistent with existing literature, confirming that alterations to cell cycle and checkpoint pathway alterations have a crucial function in the development of MFS. However, as these mutations are non-specific to MFS, they cannot reliably be used for definitive molecular diagnosis.

NF1 and RAS-MAPK pathway. Mutations to NF1 were also commonly observed in this study. NF1 encodes neurofibromin, which keeps RAS proteins in their inactive state (32). Therefore, loss of function of NF1 leads to constitutive activation of RAS, promoting unchecked cell signaling through the RAS-MAPK pathway (33). In 2013, a study performed by Dodd et al. found that MEK inhibitors showed promise in treating NF1-deleted sarcomas in mice (34). Selumetinib, a MEK inhibitor, was subsequently trialed in human patients and showed effectiveness in treating inoperable plexiform neurofibromas and low-grade gliomas, as well as in improving progression-free survival in leiomyosarcoma patients (35, 36). A clinical trial with another MEK inhibitor, Cobimetinib, is currently ongoing for its utility in locally advanced and/or metastatic soft tissue sarcomas (37). As we enter the era of precision medicine, MEK inhibition represents a possible new target for NF1-deleted tumors and should be further examined for MFS specifically. Previous studies, however, have shown that cancer cells may quickly develop resistance to these agents through subsequent loss of inhibitory genes such as CDKN2 which may limit the effectiveness of these agents in MFS, since MFS has been shown to frequently exhibit CDKN2 driver mutations (15, 16, 35).

ATRX and the alternative telomere lengthening (ALT) pathway. Mutations to ATRX were another leading alteration, present in 14.68% of the collected samples. ATRX encodes proteins crucial for chromatin remodeling and is therefore essential for maintaining the stability of DNA (15). ATRX further functions to suppress the ALT pathway, which may be hijacked by cancer cells (38). When mutated or deleted, ATRX can no longer suppress this pathway, allowing cells to proliferate uncontrollably (38). Prior research has revealed that ATRX deletions are frequently seen in solid tumors, particularly in soft tissue sarcomas like MFS (39).

The current treatment guidelines for localized MFS recommend surgical resection with adjunctive radiation therapy (3, 12, 40). Radiation therapy reduces local recurrence rates but has not been shown to improve overall survival (3, 12, 40). However, it is possible that the role of radiation is more far-reaching than what is currently being utilized in clinical practice. Interestingly, a study conducted by Floyd et al. adopted a mouse model of ATRX-deleted soft tissue sarcomas, finding that these tumors were more sensitive to radiation therapy compared to tumors without the ATRX deletion (39). Together, these findings suggest that ATRX-deleted MFS tumors may be more responsive to radiation than ATRX-positive tumors, possibly requiring a lower dose of radiation or fewer treatment sessions than previously thought compared to tumors not demonstrating this mutational burden. Radiation therapy may also act as an increasingly important arm of therapy in the treatment of metastatic MFS when the tumors harbor ATRX deletions.

TERT pathway. Telomerase reverse transcriptase (TERT) mutations were present in 4.13% samples. TERT, which encodes telomerase, is reactivated in a multitude of cancers and is another method by which cancer cells may attempt to preserve their immortality (41). Interestingly, these mutations are not seen as frequently in MFS as they are in other cancers. Previous studies have investigated the presence of TERT promoter mutations in a variety of soft tissue sarcomas, reporting they are most frequently associated with myxoid liposarcomas compared to other soft tissue sarcomas (42, 43). Rather than being TERT-driven, MFS tumors may alternatively rely more heavily on ATRX mutations and the subsequent activation of the alternative telomere lengthening pathway (38, 44). This information may be helpful in better understanding the pathogenesis of MFS and delineating the targeted therapies to which it will best respond.

Early-stage clinical trials of TERT or telomerase inhibitors have unfortunately failed to demonstrate clinical efficacy thus far in a variety of solid tumors, with research ongoing to better understand the lack of response (45). A recent study from February 2025 explored the ALT pathway, finding that ATRX-deleted tumors are successful in activating this pathway due to concurrent suppression of the cGAS-STING pathway, allowing the cells to evade immune detection (46). The authors therefore concluded that a potential avenue for future research may involve reactivating STING in ATRX-deleted tumors, therefore suppressing the means by which the cancer cells maintain telomere length (46). These findings may be foundational to future treatments that can target ATRX-deleted tumors and should therefore be more closely examined in the context of MFS.

Co-occurring mutations. Several co-occurring mutations were identified, including TP53 and RB1, TP53 and ATRX, ATRX and RB1, ATRX and NOTCH3, and NF1 and TERT. The frequent co-occurrence of TP53, RB1, and ATRX has been previously established by several studies as these are consistently among the most common mutations seen in MFS (15, 16). While TP53 and RB1 deletions work synergistically to promote rapid, unchecked movement through the cell cycle, ATRX deletions allow for cancer cells to escape senescence (30, 31, 38). The collective presence of these three mutations lays a foundation upon which cells may replicate indefinitely, driving aggressive tumor formation.

TP53 and RB1 mutations may co-occur in up to 60% of MFS tumors (47). Tumors with these co-existing mutations positively express Skp2, a cell cycle proto-oncogene (47). Cancers with concurrent loss of both TP53 and RB1 become heavily dependent on Skp2 for degradation of p21 and p27 in order to drive progression through the cell cycle (47). Pevonedistat is a neddylation-activating enzyme inhibitor that blocks Skp2, and it has been shown to decrease growth of both TP53/RB1-negative MFS and undifferentiated pleomorphic sarcoma (UPS) in patient-derived cell lines and mouse xenografts (47). A phase I clinical trial demonstrated general efficacy for Pevonedistat in several different cancers, but neither this medication nor other Skp2 inhibitors have been specifically trialed in human patients with MFS (48). Hu et al. published a 2024 study investigating novel Skp2 inhibitors, finding a specific compound which shows promising utility in designing future cancer therapies (49). As cancer treatment continues to become more individualized, it may be worthwhile to continue exploring this pathway as a potential treatment target for MFS.

The co-occurrence patterns of ATRX and NOTCH3 as well as NF1 and TERT in MFS are not well-characterized in the current literature. As previously discussed, TERT has not been frequently identified as a common mutation in MFS as the cancer seems to instead rely on ATRX deletions to achieve immortality (38, 42). Additionally, NOTCH3 has not been previously identified in other large-scale studies on the mutational landscape of MFS and therefore may be a novel mutation in our study as well (15, 16). As such, future research is warranted to better characterize the role of NOTCH3 and to understand the significance of its co-occurrence with ATRX. It is possible their co-existence is due to the general prevalence of ATRX, but additional studies are necessary to clarify the basis of their interaction.

Mutually exclusive alterations. Interestingly, total mutual exclusivity was demonstrated in two of the more prevalent genetic mutations, NF1 and RB1. The existing body of knowledge, while describing the high frequency of these mutations, does not specify whether these alterations co-exist within the same patient’s tumor or whether they are merely common to these tumors generally. To date, there are no studies detailing that these mutations have demonstrated total mutual exclusivity. While this observation seems contradictory to the current evidence, we hypothesize two potential explanations.

First, an early mutation in either NF1 or RB1 may make the other redundant. For example, loss of NF1 leads to constitutive activation of RAS. RAS, in turn, increases expression of cyclin D, which binds to CDK4/6 and phosphorylates RB1, thereby inactivating it (33, 50). Should the mutation in NF1 occur upstream of an RB1 mutation, an additional mutation in RB1 may not confer a selective advantage for the tumor. Conversely, an early loss of RB1 may make an upstream mutation in NF1 unnecessary as well, as the tumor already possesses the ability to bypass cell cycle checkpoints. The temporality of these mutations should be studied in subsequent research. Secondarily, it is possible that a mutation in either NF1 or RB1 is associated with a specific grade of MFS and is therefore exclusively seen in low- or high-grade tumors. Our present study did not distinguish which mutations were associated with which tumor grade, so it is entirely possible that one of these mutations may demonstrate preference for a certain tumor grade.

Primary vs. metastatic samples. As previously discussed, the majority of the common genetic mutations associated with MFS did not significantly differ in prevalence between primary and metastatic tumors. The only mutation exclusive to primary tumors was NCOA2. The existing body of knowledge does not describe NCOA2 mutations as being associated with MFS cases, though these alterations are present in other sarcomas as fusion genes, demonstrated in alveolar/embryonal rhabdomyosarcoma, spindle cell rhabdomyosarcoma, mesenchymal chondrosarcoma, uterine sarcoma, and small round cell sarcoma of infancy (24, 51). It is unlikely that this mutation has a significant role in the development of MFS primary tumors or metastatic lesions, but recognizing its incidence will help further delineate the molecular profile of MFS. It is most plausible that this mutation is exclusively identified in primary tumors due to its overall rarity in MFS, further supported by the fact that the mutation was only present in two samples from the cohort.

Several other mutations were exclusive to metastatic tumors, including HLA-DPA1, HLA-DRB1, KDM5D, PALLD, MLLT3, ERBB2, SOX2, SRSF2, and TBX3. None of these mutations are known as key drivers of primary or metastatic lesions of MFS. Given the low overall incidence of these mutations, combined with their failure to survive FDR-correction, it is reasonable to suggest that they do not have a major influence on the metastatic spread or overall aggressiveness of the tumors. The current dominant catalysts of metastatic transformation include genetic mutations of TP53, RB1, CDKN2, NF1, ATRX, and other common driver mutations (15). Heitzer et al. found that metastatic MFS tumors actually tended to demonstrate a less complex karyotype than the primary tumors (14). The significance of these more infrequent genetic alterations should be further investigated; as we progress toward being able to define MFS by a distinct molecular profile, it is important to note even small aberrations as they may further characterize tumor behavior or incidence in certain populations.

Study limitations. We acknowledge the limitations in our study. First, the relatively small cohort size may reduce the statistical power of our findings, underscoring the need for larger future studies. The GENIE database may introduce confounding factors, such as the inclusion of multiple tumor samples from the same patient; however, prior studies suggest this risk is minimal. Additionally, the platform lacks data on miRNA, DNA methylation, and transcriptional changes, limiting our ability to link mutations with pathway-level effects, tumor behavior, and treatment response. It also does not provide information on the timing of mutations, making it difficult to distinguish early drivers from later, potentially metastatic mutations. The absence of histological subtype or tumor grade stratification further prevented subtype-specific analysis. Moreover, the lack of treatment data precluded assessment of mutation-specific outcomes or therapeutic responses. Finally, variability in sequencing methods across contributing institutions may introduce reporting bias. Despite these limitations, our study contributes to the growing molecular understanding of MFS by highlighting key and potentially novel driver mutations relevant to future research and clinical applications.