FormalPara Key Points

Malignant peripheral nerve sheath tumors are rare, aggressive cancers that tend to recur and metastasize, and have an extremely poor prognosis.

Although surgery is potentially curative in localized disease, and despite major advances in unravelling the molecular biology, there is no standard of targeted medical therapy.

1 Introduction

Malignant peripheral nerve sheath tumors (MPNSTs) are a rare group of aggressive cancers with a poor prognosis because of their high potential for local recurrence and metastasis. For localized disease, surgical resection with or without radiation is the only potentially curative option. Postsurgically, the 5-year median overall survival (OS) rate ranges from 30 to 60% [1]. Patients with metastatic disease have a poor OS [2]. Disease-specific survival (DSS) has not improved appreciably in recent decades. MPNST constitutes the main cause of morbidity and mortality in patients with the inherited condition of neurofibromatosis type-1 (NF1) [3].

MPNSTs often exhibit extensive nerve involvement and infiltration of surrounding structures, which can render them inoperable. The median OS and progression-free survival (PFS) in patients with unresectable MPNST are less than 5 and 2 months, respectively, in trials of investigational agents [4]. The generally poor outcomes for patients with MPNST after surgery, radiation therapy, and/or chemotherapy underscore an unmet medical need for novel disease-specific therapies.

2 Epidemiology

Accounting for approximately 2–10% of the 13,590 soft-tissue sarcomas (STS) diagnosed annually in the USA, MPNST is considered the sixth most common type of STS, which in turn comprises < 1% of all malignancies [5,6,7]. About 40–50% of cases occur in association with NF1 (von Recklinghausen’s disease), while 40–47% are sporadic [8]. A history of radiation exposure is present in approximately 10–13% of MPNST cases, with a typical latency period exceeding 16 (range 4–41 [9]) years [8, 10, 11].

In patients with NF1, about 95% have small, benign dermal neurofibromas, and 30% or more develop larger plexiform neurofibromas [12]. NF1-associated plexiform neurofibromas have an approximately 10% risk of malignant transformation [9], with a higher overall risk in more centrally located neoplasms of the torso and proximal limbs as well as large nerve trunks [13].

Sporadic MPNSTs are typically diagnosed in patients ages 30–50 (mean 41 [5]) years [13], while NF1-associated MPNST is more likely to occur in younger patients, typically between 20 and 40 years old [13]. However, MPNST can have a pediatric onset. One study following 52 NF1 patients with MPNST for 20 years included 8 patients ages 1–17 years (mean 12 years) [14], and one case report described MPNST in a 20-month-old toddler [15]. It is not clear if one sex is affected more than the other [7]. Patients with NF1 have a lifetime risk of 8–13% of develo** MPNST (versus 0.001% in the general population) [16].

A cohort study found that, of 666 (1607 total) patients with NF1 who developed additional neoplasms (other than neurofibromas), 295 (18.4% of the total) had gliomas and 243 (15.1%), MPNSTs [17]. Compared with the general population, patients with NF1 tended to develop a range of tumors, including MPNSTs, at a younger age (mean 33.88 versus 47.06 years; p < 0.0001). Compared with patients with NF1 who developed other neoplasms, those with MPNSTs had a significantly lower 5-year DSS (31.6 versus 92.4%; p < 0.001).

A large tumor registry at The University of Texas MD Anderson Cancer Center [8] documented respective 5-year DSS, local recurrence-free survival (LRFS), and distant recurrence-free survival (DRFS) rates of 53.6, 56.6, and 49.6%, respectively, in patients with high-grade MPNST, including a lower 5-year DSS in NF1- (52%) and radiation- (47%) associated MPNST compared with sporadic cases (67%).

A study of 95 Korean patients diagnosed with MPNST and treated over a 27-year interval further documented differences between NF1-associated and sporadic varieties [18]. Compared with their counterparts with sporadic MPNST, patients with a history of NF1 had a significantly greater tumor size (median 8.2 versus 5.0 cm; p < 0.001), lower median age at diagnosis (32 versus 45 years; p = 0.012), and a greater number of imaging studies and surgical procedures (each p ≤ 0.004). The 10-year OS rate for localized MPNST was also significantly reduced in those with NF1-associated tumors (45% ± 11% versus 60% ± 8%; p = 0.046). A European survival meta-analysis involving more than 1800 patients found that, although the odds ratios for OS and DSS were significantly higher in patients without (versus with) NF1, this gap in survival has been narrowing in recent years [19].

A retrospective review of 175 patients diagnosed with MPNST at the Mayo Clinic from 1985 to 2010 revealed 5- and 10-year DSS rates of 60 and 45%, respectively, as well as a local recurrence rate of 22% [22]. Significant adverse prognostic factors on multivariate analysis included tumor size ≥ 5 cm, location in the trunk, local recurrence, and high tumor grade [20]. Finally, a meta-analysis involving 28 studies from 1966 through 2020 found that the pooled 5-year OS rate was 49%, local recurrence rate 38%, and event-free survival rate 37% [21].

3 Histopathology, Anatomic Distribution, and Presentation

Previously termed “malignant Schwannoma,” MPNST is considered a spindle-cell sarcoma arising in cells of neural-crest origin in which fibroblasts are frequently present. A “herringbone” or “marbled” fascicular spindle-cell growth or a branching pattern of hemangiopericytoma-like vasculature may be observed, with light hypocellular areas alternating with dark hypercellular regions [22]. MPNSTs tend to be very cellular lesions, with pleomorphism, nuclear hyperchromasia, and coagulative necrosis and high mitotic activity [9, 22]. (A small proportion of MPNSTs have a solely epithelioid morphology [9].) Up to 15% of MPNSTs include heterologous elements, including glands with a benign appearance, angiosarcomatous foci, and bony or cartilaginous islands [23].

MPNSTs are frequently situated in the proximal lower and upper extremities and the pelvis, often at the confluence of large nerve roots and bundles. These include peripheral nerves, sciatic nerves, as well as the brachial and sacral plexi and spinal nerve roots. Other sites include the head and neck (including the brain), trunk/core/retroperitoneum, and, less frequently, skin, gastrointestinal tract (mesentery, bile duct, colon, and rectum), endocrine glands (thyroid), and prostate and breast [9]. Gross pathology may include fusiform nerve enlargement [9].

A Surveillance, Epidemiology, and End Results (SEER) program database analysis involving 3267 patients with MPNST from 1973 to 2013 [24] showed the following anatomic distribution: core, n = 1307 (40.0%); limb, n = 1022 (31.3%); head and neck, n = 449 (13.7%); cranium, n = 167 (5.1%); and spine, n = 119 (3.6%). Independent factors associated with decreased survival in some series [25] include tumor size, stage, older age, and partial resection/positive margins.

MPNST typically presents as a soft-tissue mass or swelling, often enlarging over a period of several months, with potential local/radicular pain, paresthesia, or paraparesis (on disease progression). Mixed sensory and motor symptoms are suggestive findings because neurologic deficits are infrequent in patients with benign masses. Incident pain in a patient with NF1 should trigger assessment for MPNST [23].

As mentioned, the prognosis is poor, with typically rapid growth and metastasis (mainly to the lungs more often than the bone or liver). In addition to frequent metastasis and disease recurrence, patients with MPNST are also approximately 30 times more likely than the general population to develop secondary malignancies, particularly in the breast, skin, lung, and soft tissues in females as well as skin and myeloid tissues in males [26].

4 Genetics and Pathogenesis: Potential Cellular and Molecular Signaling Pathways

A complex network of signaling pathways underlies the oncogenesis and progression of MPNST. These diverse pathways may open avenues for the development of targeted pharmacotherapies.

4.1 Genomic Alterations

4.1.1 Copy Number Alterations

MPNSTs are highly copy number aberrant and enriched in chromosome 8 [27]. Chromosome 8 gain, as well as aberrations in chromosomes 1, 12, 22, and 17 (on which the NF1 gene is located: 17q11.2), have been observed [11]. Regions with potentially relevant proto-oncogenes include chromosome bands 5p15, 7p11-p13, 8q22-q24, 12q21-q24, and 17q24-25 [28].

4.2 Mutations and Downstream Pathway Alterations

Inactivating mutations in the NF1 gene may contribute to MPNST oncogenesis through loss of the Ras-inhibiting enzyme neurofibromin. Attendant activation of Ras and its downstream pathways culminates in upregulation of phosphoinositide 3-kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR)/mitogen-activated protein kinase (MAPK) signal transduction mechanisms, which in turn regulate tumor cell proliferation, survival, angiogenesis, and metastasis. In this context, a preclinical study demonstrated synergistic effects of everolimus and an inhibitor of the MEK component of the MAPK pathway (PD-901) in increasing lifespan and decreasing tumor burden, but this has yet to translate into a clinical trial [29]. The SARC031 trial (NCT03433183) involving MEK inhibitor selumetinib combined with mTOR inhibitor sirolimus in patients with unresectable or metastatic MPNST demonstrated that partial metabolic responses (by PET PERCIST) did not translate into enhanced treatment outcomes [30].

Loss of NF1 gene expression is considered a necessary but not sufficient condition for MPNST development [31]; other cooperating genetic (and epigenetic) events are required. These may include inactivating mutations of the TP53 tumor suppressor gene, loss of tumor suppressor phosphatase and tension homolog (PTEN), amplification of epidermal growth factor receptor (EGFR), Trp53 loss, and Ink4a deletion or inactivation [32,33,34,35,36,37]. According to one research team, the finding that “many [human] MPNST are Ink4/Arf deficient but p53 competent provides a unique window of therapeutic opportunity for drugs such as MDM2 and/or CDK4/6 inhibitors as potential avenues of treatment.” [38] Genetic alterations that are prominent in MPNST include EGFR overexpression, TP53 alteration, and loss of PTEN and cyclin-dependent kinase inhibitor 2A (CDKN2A) [39,40,41,42,43].

The polycomb repressive complex 2 (PRC2) and its components/subunits are salient epigenetic regulators with potential molecular pathological roles in MPNST. These include enhancer of zeste homolog 2 (EZH2), embryonic ectoderm development (EED), and SUZ12. EED and SUZ12 genetic aberrations are considered to be among the most common somatic mutations in patients with MPNST. Loss-of-function somatic mutations in EED and/or SUZ12 are observed in 90% of radiation-associated, and 70% of NF1-associated, MPNSTs [44]. Preclinical research has demonstrated that EED or SUZ12 loss results in PRC2 inactivation, which in turn drives MPNST cell growth and metastasis (including upregulation of matrix-remodeling enzyme and collagen-dependent invasiveness), with amplification of Ras-driven transcription. In clinical samples, PRC2 loss correlated with elevated fibrosis, metastasis, and reduced patient survival [45].

SUZ12 mutation may be part of a “three-hit” sequence that converts neurofibromas to MPNSTs, with the earlier hits including germline NF1 mutation (first hit) and somatic (NF1 allele) inactivation from the nonaffected parent (second hit) [46]. Finally, preclinical research suggests that inhibition of the EZH2–miR-30d–KPNBI signaling pathway may be a viable targeted treatment strategy to inhibit MPNST cell growth and survival [47, 48].

Ras-driven transcriptome analysis of 2000 candidate genes (of which 339 showed significantly differential expression in human and murine NF1-associated tumors) demonstrated that the Aurora kinase A (AURKA) gene was amplified and overexpressed in MPNSTs (but not neurofibromas). An AURKA inhibitor and short-hairpin RNAs attenuated MPNST cell growth, while AURKA-selective inhibitor MLN8237 significantly augmented survival and stabilized tumor volume in murine MPNST xenograft models [49].

Activating mutations in TYK2 have also been identified in a subgroup of NF1-associated MPNST tumors [50]. TYK2 encodes a JAK-family protein that mediates cytokine signaling via the STAT pathway and hence may affect transcription of genes regulating cancer cell apoptosis, survival, proliferation, and invasiveness. The combination of TYK2 and MEK inhibition has also shown promising in vivo effects in a preclinical model. When combined, TYK2 inhibitors (e.g., deucravacitinib [BMS-986165]) and a MEK inhibitor (mirdametinib) exerted synergistic antiproliferative and apoptogenic effects in MPNST cell lines. Clinical trials may be warranted to determine the utility of combining TYK2 and MEK inhibitors to treat MPNST.

AXL is a receptor tyrosine kinase associated with the survival, growth, and metastasis of MPNST cells. Expressed in more than 90% of MPNSTs and neurofibromas, AXL may be activated and abnormally expressed as an early event in MPNST development, conferring a tumor cell survival advantage by stimulating the PI3K/mTOR/MAPK prosurvival pathways as well as other signal transduction mechanisms related to cellular migration, growth, and invasiveness. A preclinical study demonstrated that pharmacologic inhibition of AXL and MEK1/2 decreased MPNST tumor volumes [51].

Another multifunctional tyrosine kinase receptor involved in tumor cell survival, proliferation, DNA repair, and differentiation is insulin-like growth factor-1 receptor (IGF1R). MPNST tissue samples show enrichment of IGF1R pathway copy number-altering events, including IGF1R gene amplification. A preclinical study showed that high IGF1R protein expression was related to inferior tumor-free survival and that pharmacologic inhibition of IGF1R resulted in reduced cellular invasiveness, proliferation, and migration, with decreased PI3K/AKT/MAPK signaling [52].

A further potential druggable cell membrane target is platelet-derived growth factor receptor A (PDGFRA). In an animal model, overexpression of wild-type PDGFRA (expressed by about 78% of human MPNSTs) in cells of neural-crest origin was associated with loss of NF1 and resulted in abnormal downstream Ras signaling and stimulated MPNST oncogenesis (tumorigenicity) [53].

4.3 Tumor Microenvironment (TME)

Cells and factors in the TME likely provide additional growth factors or chemokines (e.g., CXCL12) that accelerate the growth and progression of MPNSTs [54]. A receptor for the cytokine interleukin-13 [IL-13-receptor alpha 2 (IL-13Rα2)] is also overexpressed by MPNSTs. This receptor enables cancer cells to avoid death by binding to IL-13 without α1 receptor activation, and decreased α1 activation in turn augments tumor cell survival and metastasis; IL-13Rα2 also potentiates cancer metastasis and invasiveness by activating STAT3, PI3K, ERK, and AKT [55].

Expression of the hypoxia-inducible factor (HIF-1α) has been reported in approximately 75% of MPNST cases and is a poor prognostic factor [56]. Of potential interest, a preclinical study demonstrated the inter-relatedness of transcriptional factor STAT3 and HIF. STAT3 knockdown blocked HIF1α, HIF2α, and VEGF expression in four MPNST cell lines. Not only did targeted inhibition of STAT3 suppress tumorigenesis, cellular migration, and invasiveness, but these processes were dependent on HIF signaling, and HIF1α knockdown abolished these

STAT3-driven effects [57].

In addition, MPNSTs typically exhibit low or absent expression of PD-L1 and lack of PD-1 expression, with significant infiltration of CD8+ tumor-infiltrating lymphocytes and M2 macrophages [58, 59]. A large analysis of tumor-associated biomarkers in different forms of sarcoma found that macrophages comprised more than half of the total tumor immune infiltrates and that the M2:(M0 + M1) ratio in MPNST was 2.1:1 [59]. No signal for treatment has been seen for MPNSTs in immunotherapy-based clinical trials to date.

An immunohistochemical analysis found a significant increase in vascular density in MPNSTs compared with benign neoplasms. Although VEGF expression was significantly increased in MPNST compared with benign neoplasms (and basic fibroblast growth factor expression increased in neurofibromas versus normal Schwann cells), malignant transformation may require coordinated rises in VEGF receptors, EGFR expression, and TME mast cells, which can transmit inflammatory signals and stimulate angiogenesis in different kinds of tumors [60].

An extracellular matrix protein with properties of cytokines, osteopontin (OPN) was initially identified in osteosarcoma cells and mediates a range of potential cellular functions, including communication, immune-cell activation and migration, and adhesion. An analysis of copy number and gene expression in MPNSTs found an 85-fold increased expression in the secreted phosphoprotein 1 (SPP1) gene as compared with plexiform neurofibromas. Short-hairpin RNA knockdown of SPP1 in NF1-associated MPNST cells was associated with decreased tumor size and invasiveness across four MPNST cell lines [61].

Finally, recent preclinical work has identified a previously unrecognized mesenchymal stem-like subpopulation (nestin+ cells) underlying malignant transformation, with regulation by potentially druggable targets governing epithelial-mesenchymal transition and “stemness,” such as ALDH1A1 and ZEB1 [62]. Other dysregulated cellular and molecular pathways in MPNST are shown in Table 1 [23].

Table 1 Dysregulated molecular pathways in MPNST
Full size table

5 Diagnosis

In general, computed tomography (CT) and magnetic resonance imaging (MRI) help to establish tumor stage, size, and local invasiveness. MPNST is usually diagnosed as an enlarging soft-tissue mass infiltrating a peripheral nerve. Although one study showed that the mean tumor size of MPNSTs (7.4 ± 4.1 cm) was significantly greater than their benign counterparts (4.8 ± 2.7 cm; p = 0.008) [63], some degree of overlap in the sizes of these lesions may limit the clinical utility of using size as a discriminator. Similarly, frequent inhomogeneity of MPNSTs may challenge MRI to reliably discriminate between MPNSTs and benign lesions [23].

Diagnosis may be suspected before biopsy on the basis of symptoms, 2-[18F]-fluoro-2-deoxy-d-glucose (FDG) uptake on positron emission tomography (PET; in patients with plexiform neurofibromas), or relationship to a prior neurofibroma or a peripheral nerve [23]. On MRI, apparent diffusion coefficient (ADC) based on diffusion-weighted imaging (DWI) may help to discriminate MPNSTs from plexiform neurofibromas and identify MPNST in patients with NF1 [64, 65]. Concomitant FDG-PET/MRI enables comprehensive evaluation of both morphology (MRI) and metabolic data (PET) to correlate metabolic activity with lesion growth [66].

One potential diagnostic challenge is overlap** signs and symptoms of MPNST and benign lesions (Table 2) [23]. CT, contrast-enhanced MRI, and ultrasonography are important modalities to distinguish them. Core needle biopsy guided by MRI or ultrasound is used to confirm histopathology, which, along with clinical assessment, is considered the criterion standard of MPNST diagnosis [23].

Table 2 Diagnostic criteria for MPNST
Full size table

Pathologic diagnostic challenges include discriminating: (1) high-grade MPNSTs from other high-grade tumors and (2) low-grade MPNSTs from neurofibromas and/or atypical neurofibromas (Table 2). On hematoxylin and eosin staining, a range of high-grade sarcomas and other tumors may have similar spindle-cell morphology to MPNST [23]. A potential future diagnostic aid is cell-free DNA ultra-low-pass whole-genome sequencing [67].

Proactive monitoring of patients with NF1 may assist in surveillance for malignant transformation to MPNST. For instance, the total volume and number of neurofibromas are significantly correlated with the risk of such transformation. Potentially prodromic manifestations in patients with NF1 include hardening of soft neurofibromas, sleep disturbances, prolonged pain, and new neurologic deficits (including loss of sphincter control) [13]. The following conditions may warrant particularly vigorous monitoring: early MPNST diagnosis, a history of radiation therapy, as well as plexiform neurofibromas situated in the abdominal and pelvic regions, lumbosacral and shoulder plexi, and spinal nerve roots [13].

6 Potential Treatments

6.1 Surgery

Radical surgical treatment with wide local excisional margins is the standard of care for localized high-grade MPNST and is positively associated with life expectancy [23, 68]. Referral to a specialized center, together with multidisciplinary care involving surgical oncology experts, clinical geneticists, specialized neuropathologists, and radiation therapists, is a hallmark of effective surgical care [68]. Using a wide approach to the lesion and taking care not to rupture the tumoral capsule, an incision is made over the tumor, oriented along the nerve’s anatomical track, with palpation of the tumor and/or ultrasonography to assist in localizing tumors that are not palpable. Thoracic and abdominal CT screening may play a role in identifying any secondary lesions [68]. Care must be taken to assess the relationship of the neoplasm to its vascular supply, the possibility of tissue invasion, and the proximity of other peripheral nerves that might be damaged during retraction, transection, and suture [68].

In a study of 123 patients with MPNST (90 sporadic and 33 NF1-associated), 28% of patients in whom surgery was attempted had negative surgical margins. These patients had a local recurrence rate of 6% as compared with 30% among the remainder with positive margins [69]. Consistent with a multivariate analysis showing that NF1 (versus sporadic) MPNST and tumor volume (≥ 200 mL) were independent predictors of inferior outcomes, the authors suggested that routine screening using MRI and/or FDG-PET in the setting of NF1 may be warranted. In an Italian case series involving 205 patients with localized MPNST who underwent surgery, positive margins were significantly associated with approximately 2-fold increased risks of cause-specific survival [hazard ratio (HR), 1.86; p = 0.021] and local recurrence (HR, 2.43; p = 0.001) [70]. A French study found that patients with R0 versus R1/2 resection had 5-year DFS rates of 45.2 and 34.3% (p = 0.0226) and 5-year OS rates of 54.6 and 41.0% (p = 0.1108), respectively [71]. Margin status was not statistically significantly associated with local recurrence or distant metastasis on multivariate analysis.

Anatomic location and tumor grade affect the surgical approach and rate of negative surgical margins. In one study [24], spinal masses were most frequently resected (83.0%) and intracranial lesions least frequently (58.1%). Although radical resection is plausible in most patients with limb MPNSTs, it may be necessary to remove the main nerve trunk (e.g., sciatic nerve).

6.2 Radiation and Chemotherapy

Multidisciplinary (multimodal) care based at an experienced sarcoma center is the cornerstone of effective management for localized disease [72]. In general, neoadjuvant chemo- and radiotherapy can be contemplated if clinical and imaging data suggest that a tumor may not be adequately resected, including preoperative planning for patients with tumors exceeding 5 cm or in the event that tumor mass needs to be rapidly decreased because the lesion is im**ing on a nerve or other structure [13]. In addition to downstaging neoplasms that are borderline unresectable, neoadjuvant chemotherapy can help to ascertain chemosensitivity in patients at risk for disease dissemination, for whom doxorubicin combined with ifosfamide may be associated with improved outcomes [23]. The role of adjuvant chemotherapy is equivocal at best [73]. One study confirmed that such treatment is not effective, with the exception of epirubicin with ifosfamide, which augmented median survival by up to 75 months after local treatments, such as radiation after wide resection, amputation, or surgery after preoperative radiation [74].

6.2.1 Radiation

Local control of MPNST may be improved by neoadjuvant or postoperative adjuvant radiation. A multivariate analysis from a French histopathologic study of 350 patients found a 4.5-fold higher risk of local recurrence (HR, 4.50; p = 0.003), but not distant metastasis, at 5 years in patients who did not receive adjuvant radiation therapy [71]. However, tumor grade was significantly associated with metastasis but not local recurrence. A study [75] of NF1-associated and sporadic tumors affecting chiefly the extremities (58%), trunk (36%), and head and neck (6%) found an overall median OS of 46.5 months with a 5-year OS rate of 43.7%. Radiation therapy, administered to 20 of the 33 patients in this study, included brachytherapy, external beam radiation, a combination of these modalities, and proton therapy, with median total doses of 59.4 Gy for NF1-associated, and 58.5 Gy for sporadic, MPNSTs. The median OS was 33.1 months among patients with NF1-associated MPNST who received radiation, as compared with 17.4 months among those who did not. In patients with sporadic MPNST, median OS values were 51.6 and 69.4 months in radiation treated versus untreated patients, and the authors concluded that radiation therapy overall was not a significant predictor of survival.

In a retrospective analysis of 24 patients with MPNST, Kar’s group reported a 5-year DFS rate of 42% for patients with mostly deep-seated, high-grade, large (mean 10.83 cm) MPNSTs who received postoperative radiation therapy (median dose 58 Gy) compared with 0% among those who did not; respective 5-year OS values were 65 and 38% [76]. An international consensus group [77] recommended radiation for local control, to defer recurrence, and for intermediate- to high-grade lesions (and lower-grade neoplasms after marginal excisions), although this intervention is not associated with long-term life expectancy.

6.2.2 Chemotherapy

MPNSTs exhibit intermediate (to poor) chemosensitivity [77]. The chief role of chemotherapy (particularly anthracyclines + ifosfamide) is in the management of advanced, unresectable, or metastatic MPNST, including within palliative settings [78]. The multicenter phase 2 trial, SARC006 (NCT00304083) [79] of 48 patients with NF1-associated (n = 34) or sporadic (n = 14) MPNSTs supported the role of neoadjuvant chemotherapy in patients with high-grade or metastatic MPNST. The investigators reported nine partial responses (PRs; 5/28 [17.9%] NF1-associated and 4/9 [44.4%] sporadic) and 26 stable disease (SD; n = 22 NF1-associated and n = 4 sporadic) after two 21-day treatment cycles each of ifosfamide with doxorubicin and ifosfamide with etoposide. Neoadjuvant chemotherapy regimens supplanting doxorubicin with epirubicin may have a lower risk of cardiotoxicity. A retrospective analysis at Washington University in St. Louis found a 60% objective response rate (ORR) and 100% clinical benefit rate (CBR) in five patients with biopsy-documented MPNST who received neoadjuvant epirubicin (60 mg/m2 on days 1 and 2) and ifosfamide (1800 mg/m2 on days 1–5), with similar responses in those with either NF1-associated or sporadic MPNST [80]. Although anthracycline plus ifosfamide has clinical activity, it is also associated with a higher toxicity burden. A retrospective study [81] of 115 patients treated because of metastatic or unresectable MPNST from 2000 to 2019 found that, although doxorubicin was the most frequently used first-line treatment, there were no significant differences between the most commonly used regimens in terms of PFS, including 3.6 months for doxorubicin + dacarbazine, 2.7 months for high-dose ifosfamide, 3.1 months for etoposide + ifosfamide, 1.9 months for doxorubicin + ifosfamide, and 5.9 months for doxorubicin + cisplatin. The median OS in individuals treated with palliative intent was 15.0 months. Compared with regimens based on doxorubicin, use of gemcitabine-based regimens was associated with a significantly poorer prognosis [HR, 1.83; 95% confidence interval (CI), 1.12–3.00; p = 0.016] [81]. The median PFS was 3.9 months for anthracycline-based regimens [without local treatments (surgery or radiation)] in first line, as compared with 2.4 months in later lines (p = 0.143).

6.3 Targeted Therapies

Despite multiple cellular and molecular “leads” (biomarkers), no single driving pathogenetic event has been identified, and no particular targeted therapy has emerged as a clearly superior regimen against MPNST. Completed, small trials involving targeted therapies for mainly unresectable and/or metastatic MPNST have yielded a cumulative [response evaluation criteria in solid tumors (RECIST)] disease control rate of 22.9–34.8% (Tables 3, 4). On the basis of their analysis of data from targeted therapies in patients with MPNST (including studies summarized in Table 4), Akshintala and colleagues have suggested a threshold 4-month PFS rate of 30% for efficacy in single-arm trials [82].

Table 3 Summary of clinical data on targeted therapies in MPNST
Full size table
Table 4 Summary of clinical data on targeted therapies in MPNST (including patients with MPNST in “basket” studies)
Full size table

Encouraging but preliminary results were reported in a Japanese phase 2 investigator-initiated trial of the small-molecule multikinase inhibitor pazopanib, an antiangiogenic compound inhibiting VEGFR and PDGFR [83]. This study of 12 patients (median age 49 years) with tumors that were mostly metastatic (91.7%), deep-seated (83.3%), and/or extremity based (58.3%); exceeded 5 cm (58.3%); and/or had Eastern Cooperative Oncology Group Scale of Performance Status (ECOG PS) of 0 or 1 who received oral pazopanib 800 mg once daily found a RECIST or Choi CBR at 12 weeks of 50% [one PR and five SDs by RECIST (and three PRs and three SDs by Choi; Table 3]. Grade 3 or 4 adverse events included neutropenia, lipase or transaminase increase, leukocyte decrease, and left-ventricular (systolic) dysfunction (leading to treatment discontinuation).

Further potentially promising outcome activity of a multikinase (VEGFR) inhibitor was revealed in a phase 2 trial of tivozanib. Among five patients ages ≥ 18 years with metastatic, locally advanced, or recurrent MPNST; 1–4 prior therapies; and ECOG PS < 2, one PR was reported [84]. (Numbers of MPNST patients with SDs were not specified.) Grade 3 adverse events included hypertension.

First-in-class oral inhibitor of exportin 1 (XPO1) selinexor also showed promise in an international retrospective study of nine patients with unresectable or metastatic MPNST. Of these patients, seven received selinexor monotherapy at 60 mg orally twice weekly and two, a combination of the XPO1 inhibitor 80 mg once weekly and doxorubicin 75 mg/m2 [85]. Three patients experienced PR (two with selinexor monotherapy), four had SD, and the remainder had progressive disease. Tumor size reductions ranged from 10% to 56% at the time of best response in seven patients, and the mean (range) duration of disease control was 4.5 (2.5–13.5) months [85]. Adverse events included fatigue, anorexia, and biochemical or hematologic abnormalities, with four of nine patients requiring dose reductions because of intolerable AEs .

No objective responses were observed after treatment of MPNST via a wide range of targeted therapeutic mechanisms, including the BCR-ABL1 TKIs imatinib (in SARC001 [86]; one patient had SD at 2 and 4 months) and dasatinib [87] (number of SDs in MPNST not reported); the EGFR TKI erlotinib [88] (number of SDs not reported); an mTOR inhibitor combined with Hsp90 inhibitor ganetespib (in SARC023 [89]) anti-VEGF monoclonal antibody bevacizumab plus everolimus [in SARC016 [90]; according to World Health Organization (WHO) criteria, the best response was SD in 3/25 (12%) patients] or multikinase (CSF-1R) inhibitor pexidartinib with sirolimus [91] (6 SDs); multikinase (RAF/MEK/ERK/VEGFR/PDGFR) inhibitor sorafenib [92] [3/12 (25%) MPNST patients had SD; the median PFS was 1.7 months, and the median OS 4.9 months], and AURKA inhibitor alisertib (no SDs reported; the median PFS was 13 weeks) [93]. Given the high incidence of MTAP deletion in MPNST, PRMT5 inhibitors [e.g., TNG908 (NCT05275478), TNG462 (NCT05732831)] are emerging as potential targeted therapies and have demonstrated antiproliferative and antitumor effects in preclinical models of MTAP-deleted MPNST [94].

Ongoing clinical trials in MPNST are summarized in Table 5. Many are pursuing combinatorial approaches directed toward multiple signaling pathways. Signaling pathways of interest that are not targeted in the aforementioned trials (summarized in Tables 3, 4) include EZH2 with tazemetostat (NCT04917042), immune checkpoints with PD-1 inhibitor nivolumab and CTLA4 inhibitor ipilimumab (NCT04465643), and c-Met with AL2846, as well as MDM2-p53 and PD-1 with alrizomadlin and pembrolizumab (NCT03611868). Oncolytic viral approaches are also being investigated (NCT02700230).

7 Limitations

Given the overall low incidence of MPNST, the limitations of this review (and the field in general) center on the inclusion of retrospective evidence (e.g., single-center case series). Additional high-quality, prospective trials (and/or observational registries) involving larger numbers of patients with MPNST are warranted to evaluate the relative merits and pitfalls of distinct therapeutic strategies and better understand the roles of radiation and chemotherapy in MPNST management. Finally, the range of benchmark DCR values for targeted therapies in MPNST (22.9–34.8%) that we report on the basis of our literature review is not representative of pooled datasets but rather a cumulative value of several small sample size studies, recognizing the bias of the variability of patient populations (Table 3, 4).

Table 5 Summary of ongoing clinical trials of targeted therapies for MPNST
Full size table

8 Summary

Malignant peripheral nerve sheath tumors are rare, aggressive soft-tissue sarcomas with a tendency for local recurrence and metastasis and are recognized to have among the poorest prognoses of all sarcomas. Although surgical resection is potentially curative in localized disease, overall outcomes with this therapeutic modality and others are suboptimal, establishing an urgent unmet medical need. Advances in genetics and molecular biology have unveiled a range of potentially pathologic signaling pathways, yet further research is warranted to unravel the complex pathogenesis of this condition, no single druggable target has emerged, and combinatorial approaches addressing multiple pathways are hallmarks of ongoing clinical trials. To date, the prognosis of this disease, which affects children, young adults, and mid-life adults in their prime is very poor, with a maximum (after surgery with curative intent) median OS of 95.8 months and median PFS of 26.3 months (< 5 and < 2 months, respectively, in unresectable MPNST) [95], which have not improved over the last 20 years, and a reliance on only toxic chemotherapy without significant progress in targeted therapy therapeutic options, which are dearly needed.