Epidemiology and Risk Factors

It is estimated that from 5% to 10% of the 6000 soft tissue sarcomas diagnosed in the United States per year are malignant nerve sheath tumors, with an incidence of 0.001% in the general population.¹⁰ These tumors occur with equal frequency in males and females, although some series have shown a female preponderance.^3,4 There is no racial association. Most studies show that the peak incidence of MPNSTs is in the seventh decade of life in the general population but in the third or fourth decade in people with neurofibromatosis type 1 (NF1),^11,12 although these tumors may occur at a much younger age in either population.¹³

Most MPNSTs occur in patients with NF1, with a cumulative lifetime risk of up to 10%. Nonetheless, a surprising number occur as solitary MPNSTs, unassociated with neurofibromatosis or other predilections such as irradiation. Individuals with NF1 and internal plexiform neurofibromas are 18 times more likely to develop MPNSTs than patients without internal plexiform neurofibromas.¹⁴ In the general population, plexiform neurofibromas can undergo malignant transformation to an MPNST with an estimated lifetime risk of 3% to 5%, whereas in NF1 patients, it can be as high as 15% to 20%. The dermal neurofibromas seen in NF1, although more numerous and a more troubling cosmetic problem, do not undergo malignant transformation. Only rarely do MPNSTs arise from malignant degeneration of a schwannoma, ganglioneuroma, or pheochromocytoma. Ten percent of these tumors occur in patients who have undergone radiation treatments for other diseases, and they occur on average 15 years after the treatments.¹⁵

The effect of radiation on peripheral nerves was described initially in animal experimental work by Bergstrom and Cavanagh.^16,17 The incidence of radiation-induced MPNSTs reported in large series ranges from 5.5% to 11%.^6,13,18,19 There are also several published reports relating to patients with and without NF1 and radiation treatments.^13,20,21 Ducatman and associates¹³ described 12 patients with postirradiation MPNSTs, 7 of whom had NF1. Two of the 7 had received radiation therapy for optic gliomas 5 and 17 years previously. More recently, Loree and colleagues²¹ described two of four NF1 patients who developed MPNSTs after head and neck irradiation, whereas Baehring reported such after radiation for Wilms’ tumor and Hodgkin’s disease.⁶

Diagnosis

The diagnosis of these tumors remains problematic because it is based primarily on clinical suspicion. As with any patient, a history and physical examination are the place to begin the assessment for a peripheral nerve tumor. In the history, special note should be made of when the mass, if palpable, was noticed, and the onset of symptoms such as pain and motor or sensory deficit. Rapid increase in the size of a mass or rapid onset of symptoms should immediately alert the surgeon to the possibility of a malignancy (Fig. 245-1). A patient with a known history of NF1, neurofibromatosis type 2 (NF2), or schwannomatosis who presents with a tumor that shows recent rapid increase in size, a new or progressive neurologic deficit, or pain should alert the examiner to a suspected malignant degeneration. The examiner should question and record the location, quality, and radiation of pain. The location and extent of motor weakness, if present, and the location and extent of sensory deficit should be defined and recorded. A family history of peripheral nerve problems or any other genetic disorders should be closely questioned, and a history of previous radiation treatments should be discussed. Systemic diseases or any preexisting conditions that can contribute to peripheral nerve problems should also be questioned (i.e., diabetes mellitus, cancer). Any recent illnesses, even those as seemingly minor as flu, should be questioned and recorded. Because many prescription medications can cause peripheral neuropathies, a medication history should also be recorded.

FIGURE 245-1 A, A 71-year-old man had previously undergone complete excision of what was read histologically as a benign median nerve tumor in the axilla of the left arm. Within 1 year, he returned with a recurrent, fixed, painful mass in the same area. B, Tumor exposed. The tumor was once again resected and histologically was now identified as a malignant schwannoma that would require significant adjuvant therapy. This case underscores that a rapidly enlarging or recurring, painful, nonmobile mass should alert the surgeon to the possibility of malignancy.

During the physical examination, special attention should be given to examining for the presence of café au lait spots, axillary freckling, inguinal freckling, and Lisch nodules (pigmented iris hamartomas), which can indicate the presence of a genetic disease such as neurofibromatosis. Any spinal scoliosis that may indicate the possibility of intraforaminal tumors distorting the spinal column should also be noted. All four extremities should undergo a complete motor examination with standard motor strength grading as well as a sensory examination. Reflexes should be tested, and the patient should be examined for the presence of Tinel’s sign at the area of the mass or suspected tumor. The distribution and extent of Tinel’s sign should also be noted because it will offer clues as to which nerve is involved. If a mass is palpable, the size, quality, and mobility should be noted. Traditional teaching relates that a nerve tumor is mobile from side to side but not along the length of the nerve proximally and distally. The likelihood of malignancy is increased with increasing tumor size, a consistency that is hard to palpation, and a mass that is fixed to the surrounding soft tissue.

Imaging

The “gold standard” for imaging of peripheral nerve tumors has become magnetic resonance imaging (MRI). If a solitary palpable mass is encountered clinically in a patient and a peripheral nerve tumor is suspected, an MRI of the involved extremity, plain and with contrast, is indicated. If there are multiple tumors palpable, or if an indication of neurofibromatosis or schwannomatosis is noted on examination, the imaging should be more comprehensive, including a full spine series to define any spinal or foraminal masses. Contrast images should always be ordered to evaluate the enhancing quality of the mass. Information on the enhancing qualities of the mass, combined with its appearance on T1- and T2-weighted images, can give valuable clues to the histopathology that may be encountered.

On any contrast-enhanced image that is being performed on a peripheral nerve tumor, fat suppression sequences should be used to better define the nerve in question. Newer techniques using magnetic resonance neurography (MRN) have the potential to offer enhanced visualization and definition of peripheral nerve mass lesions and produce higher-resolution images of the nerves with greater separation from the surrounding soft tissue. Proper MRN requires special phased-array surface coils and radiologists familiar with the short tau inversion recovery (STIR) sequences necessary to obtain the desired images.

MRI can contribute significant information about the suspected pathology. This information is extremely useful preoperative information. Unfortunately, whether a tumor is benign or malignant cannot be discerned definitively from the image alone (Fig. 245-2). Areas of hemorrhage or necrosis, heterogeneous enhancement, and cystic areas may suggest a malignancy but are no means definitive and can even be seen in benign tumors.

FIGURE 245-2 A, A 34-year-old woman with a known history of NF1 and a painful right thigh mass associated with the sciatic nerve. Patient is in the prone position at operation. This ultimately proved to be a benign neurofibroma. Not all painful masses prove to be malignant; however, given the consequences of a missed malignancy, it is better to err on the side of caution with surgical excision of such lesions. B, T1-weighted, contrast-enhanced, coronal magnetic resonance imaging (MRI) of the patient seen in part A. Note that despite nonuniform enhancement of the mass, MRI cannot predict malignancy. This tumor was ultimately shown to be histologically benign.

Careful assessment of the images should include surrounding blood vessels and nearby vital structures and whether any infiltration of these surrounding structures is present.

Positron emission tomography (PET) with the glucose analog ¹⁸FDG is a dynamic imaging technique that permits the visualization and quantification of glucose metabolism in cells and reflects the increase in metabolism in malignant tumors.^22,23 A retrospective study of 18 NF1 patients demonstrated that ¹⁸FDG-PET is a potentially useful, noninvasive method for detecting malignant change in plexiform neurofibromas.²² However, the distinction between low-grade MPNSTs and benign plexiform neurofibromas was not clear in all the cases. The new tracer ¹⁸F-thymidine, which detects DNA turnover, might be helpful in distinguishing low-grade MPNSTs from active, benign plexiform neurofibromas in future PET-based studies.²⁴

Nerve conduction and electromyography (EMG) are generally not required in the preoperative evaluation of MPNSTs because they are not diagnostic, nor do they influence management. They may, however, be useful in documenting baseline neurological function.

Treatment

When the diagnosis of MPNST is suspected, surgery is the mainstay of treatment of these tumors.^24–26 Resectability depends largely on location and ranges from 20% in paraspinal MPNSTs to 95% in tumors of the extremity.^{2,6,13,18,19,27,28} The ultimate aim of surgery is complete removal of the lesion with tumor-free margins.^13,29

There are many different approaches to these tumors, some of which are controversial. Total excision, subtotal debulking, and biopsy may all be considered. Because the surgical resection to clean margins may require sacrifice of vital nerve and soft tissue structures, some surgeons advocate fine-needle biopsy first, before definitive surgery. Other surgeons believe that a fine-needle biopsy may miss malignant cell rests and lead to subsequent misdiagnosis or allow for spread of tumorigenic cells and therefore avoid this approach. A modification of this approach uses an open biopsy of multiple (at least four) quadrants of the tumor. Still others advocate a staged approach whereby first a gross total resection of the tumor is performed with care taken not to violate tumor capsule. A full pathologic examination of the entire tumor tissue ensues, and if malignancy is proved histologically on permanent sections, the patient is returned to surgery, in a timely fashion, for definitive oncologic surgery. The margins of the resection site and surrounding tissue are then explored, with sampled tissue sent for frozen section to be examined pathologically. As this surgery progresses, the definitive operation to clean margins to at least 2 cm in all directions is performed.

This last approach, in our opinion, allows for a thorough pathologic examination of the initial tumor tissue so that nothing is overlooked, a situation that may not be possible if only a small biopsy sample is taken. Furthermore, it allows the surgeon to discuss with the patient, in the interval between the two operations, more accurately what can be expected and the greater likelihood of neurological deficit after surgery as diseased nerve and soft tissue are removed.

Unfortunately, for neurogenic sarcomas that involve brachial or pelvic plexus or the proximal portion of the arm, a wide resection to clean margins is not accomplished without paralysis or even limb loss necessitated by vascular supply sacrifice. Thus, wide local resection appears to work better for neurogenic sarcomas involving the more distal portions of the limb. For more proximal lesions, amputation of the limb may be required, but many patients, given the cosmetic deformity and overall poor prognosis of these tumors, often refuse such recommendations (Fig. 245-3).

FIGURE 245-3 A, This 16-year-old man with neurofibromatosis type I originally presented with a painful supraclavicular mass. This was partially resected because it extended deeply under the sternum, making total resection impossible. It was read as histologically benign. B, The patient was lost to follow-up for 1 year when he re-presented with this appearance. On open biopsy, the tumor now had undergone malignant degeneration to a malignant peripheral nerve sheath tumor. Initially, the family refused amputation, arguing that the arm remained functional. After further discussion, the family agreed to the necessary four-quadrant amputation of the limb, followed by adjuvant radiation therapy.

The reported local recurrence rate of MPNST following gross total resection is 32% to 65% after median intervals of 5 to 32.2 months.^{6,13,18,19,27,28,30} There have been statements that a microscopically non–tumor-free margin may be an indication of a highly aggressive, invasive tumor, rather than being a reflection of inadequate surgical technique, assuming that scrupulous attempts to attain microscopically tumor-free margins have been made.³¹ Local aggressive resection and control are thought to decrease the risk for systemic metastasis and to contribute to a better overall prognosis.

Reconstruction of the nerve after surgery for removal of malignant brachial and lumbosacral plexus lesions is not advocated. Because the needed adjuvant radiation and chemotherapy will compromise the ability of the axons to advance and grow down to the target organ, attempts at reconstruction are thought to be a futile exercise. Furthermore, the natural history of MPNST is often not long enough for effective reinnervation.²⁹ Amputation may be indicated for extensive tumors and for MPNSTs that recur after apparently adequate excision.²⁴

Radiotherapy and chemotherapy are unfortunately of only limited value. However, they are routinely applied because our armamentarium against these tumors is so limited. To date, only complete surgical excision before metastasis is likely to give a good prognosis.^2,6,13,21,32

Classification and Pathologic Diagnosis and Grading

The NF1 gene on chromosome 17q11.2 was identified by positional cloning, and its protein product, neurofibromin, functions as a tumor suppressor.^33–35 One of the functions of neurofibromin is to reduce cell proliferation by accelerating the inactivation of the proto-oncogene p21-ras, which has a pivotal role in mitogenic intracellular signaling pathways.³⁵ Patients with MPNST demonstrate aberrations in 17q11.2-22 and show loss of NF1 gene expression, with resultant increase in ras oncogene expression.

A soft tissue sarcoma is denoted to be of neurogenic origin if it fulfills one the following criteria: (1) macroscopic or microscopic association with a peripheral nerve, (2) malignant transformation of a preexisting neurofibroma, or (3) immunohistochemical or ultrastructural features consistent with peripheral nerve origin.¹¹

Classification of MPNSTs is based on that used for the much more common soft tissue sarcomas, from grade 1 to 3, depending on number of mitotic figures and degree of nuclear and cellular atypia.¹¹

Tumors are also classified based on gross size. Tumor size at surgery is stratified to be greater than or less than 5 cm. This has been based on previous reports in which soft tissue sarcomas larger than 5 cm correlated with a worse prognosis.^{11,13,36–39} The tumor size influences the treatments and outcomes in several ways. First, patients with neurogenic sarcomas larger than 5 cm present twice as often with neurological motor or sensory deficits compared with patients with neurogenic sarcomas 5 cm or smaller. Second, tumor size correlated with pathologic grade, an important predictor of survival and systemic spread. The most aggressive grade 3 tumors are found in patients with neurogenic sarcomas larger than 5 cm. Third, the size of the presenting neurogenic sarcoma influences the ability to obtain tumor-free margins in the first en bloc resection. Fourth, as with other sarcomas, the presenting size of neurogenic sarcomas appears to negatively influence the survival rate.¹¹

Pathology

Gross inspection of MPNSTs reveals a fusiform, fleshy, tan-white mass with areas of degeneration and secondary hemorrhage. The nerve proximal and distal to the tumor may be thickened owing to spread of the tumor along the epineurium and perineurium. The cell of origin is the Schwann cell, although the mature Schwann cell marker S-100 may be absent in about 50% of cases because of dedifferentiation.²⁹

The minimum histologic examination should comprise sections stained with conventional tinctorial stains including hematoxylin and eosin stain and reticulin stain. At histologic analysis, MPNSTs are nonencapsulated infiltrating tumors composed of spindle cells arranged in a whorling pattern with irregular nuclei, cyst formation, and nuclear palisading.⁴⁰ Mitotic figures are readily visible, with more than one per high-power field, and 50% to 90% of cases are immunoreactive with S-100 protein staining.⁴¹ Necrosis, pseudocystic change, or hemorrhage may also be found. The pathologic criteria for malignancy include invasion of surrounding tissues by tumor cells, vascular invasion, marked nuclear pleomorphism, necrosis, and the presence of mitoses.⁴²

Using the hematoxylin and eosin–stained sections, the tumors are graded 1 to 3, with a scheme for soft tissue sarcomas based on cellularity, nuclear pleomorphism, anaplasia, mitotic rate (mitotic figures in 10 high-power fields), microvascular proliferation, and degree of necrosis and invasion.^3,11,43,44 In addition, immunohistochemical stains for S-100 protein, the skeletal muscle markers desmin and myogenin, and a proliferation marker (MIB-1) are required.

Other spindle cell tumors may be excluded with appropriate immunohistochemical markers. The diagnosis of a neurogenic sarcoma cannot be reliably made from examination of hematoxylin and eosin–stained sections alone because other soft tissue sarcomas, arising from fibroblasts or smooth muscle cells, may have similar appearances. Three immunohistochemical markers (S-100, Leu-7, and myelin basic protein), although not diagnostic by themselves because of significant false-positive and false-negative rates, are used to facilitate the diagnosis of neurogenic sarcomas.^13,45–49

Mitotic rates are evaluated as 0, 1, and 2 depending on the numbers as less than 5, 5 to 10, and more than 10 per high-power field. More than 5 mitoses per 10 high-power fields has been considered to denote a high-grade tumor because a single mitotic figure may be significant in a tumor with hypercellularity and nuclear atypia.²⁴ More than 5% cellular staining of an MIB-1 proliferation marker has been considered to denote a high-grade tumor.^50,51

On electron microscopic examination, ultrastructural features suggestive of a neurogenic origin for the tumors recapitulate the features of normal Schwann cells. These features include wavy, buckled, or comma-shaped nuclei arranged in sweeping fascicles with extensive perineural and intraneural spread of the tumor. Also often seen is a proliferation of the tumor in the subendothelial zones of vessels, so much so that neoplastic cells appear to herniate into the lumen.⁵² Also important is the lack or relative lack of ultrastructural features such as myofibrils, which suggests other origins for soft tissue sarcomas.

These tumors often create diagnostic problems because of their cellular origin and histopathologic similarities with other spindle cell sarcomas such as monophasic synovial sarcoma, leiomyosarcoma, and fibrosarcoma. Another interesting clinical feature of this tumor is multifocality and development of second primary tumors of the same histology.¹³ However, it is not always possible to demonstrate the origin from a nerve, especially when it arises from a small peripheral branch. This point was exemplified in a series by Nambisan and colleagues, in which nerves could not be identified in 61% of cases of MPNST,²⁶ and in the series by Bilge and coworkers, in which nerve origin could be identified only in 45% to 56% cases.⁵³ Still, there are several distinct features, such as proliferation of tumor in the subendothelial zones of vessels, with neoplastic cell herniation into vessel lumen, and proliferation of small vessels in the walls of the large vessels, which are characteristic features of MPNSTs.⁵⁴

Radiotherapy

Radiotherapy provides local control and may delay the onset of recurrence but has little effect on long-term survival. However, there are reports of routine postoperative radiotherapy and even radiotherapy as a single modality for MPNST in the literature.⁵⁵ Currently, postoperative radiotherapy is recommended by the Oncology Consensus Group²⁴ as part of a uniform treatment policy for MPNSTs, much like other high-grade soft tissue sarcomas,^2,8 even if clear surgical margins are obtained. Adjuvant radiotherapy should be given whenever possible for intermediate- to high-grade lesions and for low-grade tumors after a marginal excision. Although radiation therapy does provide local control and delay recurrence, it has little effect on long-term survival.

The timing of irradiation (either before or after surgery) and the associated pros and cons are actively being studied by the group at the University of Toronto. Postoperative radiotherapy involves irradiation of the entire operative field, with a 5-cm field margin. Preoperative radiotherapy involves irradiation of the overt tumor alone, again with a 5-cm margin. The usual dose administered is 6000 to 7000 cGy Irradiation before surgery has been recommended if the location, size, and distribution of the tumor make it more technically difficult to provide optimal radiotherapy after excision, dissection is anticipated along a major neurovascular bundle (with the possibility of leaving microscopic disease in critical structures), or remote tissue flaps or skin grafts are required for wound management after resection.^11,56,57

Chemotherapy

Like most soft tissue sarcomas, MPNSTs are traditionally chemotherapy insensitive. Chemotherapy for adult soft tissue sarcomas is usually confined to the treatment of metastatic disease. Systemic spread, especially pulmonary metastasis, is the terminal event and, despite limited efficacy, chemotherapy is indicated in this situation.¹¹ Because of the rare incidence of MPNSTs, large trials for effectiveness of chemotherapy in MPNSTs are impossible, and most current data are based on case reports or small case series or on regimens proved successful for other soft tissue sarcomas.

Few drugs have been shown to be effective, and treatment comprises single-agent doxorubicin or a combination of doxorubicin and ifosfamide,⁵⁸ with a partial response rate of 20% to 25%. With doxorubicin, there was little impact on survival. Dacarbazine was later noted to have activity against these tumors and was combined with doxorubicin in the CYVADIC regimen.⁵⁹ In the late 1980s, a number of phase II trials demonstrated the superiority of ifosfamide to cyclophosphamide in soft tissue sarcomas. The European Organization for Research and Treatment of Cancer (EORTC) reported a phase III study comparing doxorubicin to doxorubicin plus ifosfamide to CYVADIC and demonstrated an insignificantly increased response rate with combination therapy, at a cost of increased toxicity with doxorubicin plus ifosfamide.⁵⁸

Chemotherapy is not curative, and therefore its use is controversial. There is a benefit at 10 years for progression-free survival for both local and distant relapse as shown by a meta-analysis.⁶⁰ However, the magnitude of any overall survival benefit is small (4%, not statistically significant).

Prognosis

Patients (and parents of pediatric patients) with NF1 should be made aware of the relatively high risk for MPNST associated with the disease. They should be instructed to contact their physician sooner rather than later should rapid enlargement, pain, or neurologic deficit occur.¹² MPNSTs have a poor prognosis, and in most studies, the prognosis appears to be worse in patients with NF1.^13,30 MPNSTs in broad terms can be considered fatal, with the worst outcomes related consistently to incomplete tumor resection.^13,19 Sarcomatous cells spread extensively within the fascial planes, resulting in high recurrence rates and ultimately systemic spread.^11,61

MPNSTs have a poor prognosis because metastases to the lung, liver, brain, soft tissue, bone, regional lymph nodes, skin, and retroperitoneum are common.¹³ Adverse prognostic factors include large size (>5 cm), higher tumor grade, advanced histology, non–tumor-free surgical margin, and associated NF1.¹¹

Basso-Ricci demonstrated a 56% disease-free survival rate using combined surgery and radiation therapy for MPNST.⁶² MPNSTs have the highest recurrence rate of any sarcomas,⁶³ and adequate initial treatment gives the best chance for survival.⁶⁴ Even with aggressive therapy, local recurrence of tumor is seen in 50% of patients.⁶⁵ Angelov and coworkers recommend that sacrifice of major neural structures such as the sciatic nerve should be performed, without exception, to achieve resection with tumor-free margins.¹¹ Other studies using preoperative radiotherapy or interstitial radiotherapy demonstrated that only a minority of patients with non–tumor-free margins ultimately develop locally recurrent or distant metastatic disease.^66–68 Hematogenous metastatic spread occurs most commonly to the lungs. The 5-year survival rates in large series have been reported to range between 16% and 52%.^13,18,19,69 The reported 5-year survival rate for patients with MPNSTs without NF1 is as high as 50%. It drops to as low as 10% for MPNSTs in patients with NF1.³⁰