Malignant Peripheral Nerve Sheath Tumors Associated With Neurofibromatosis Type 1: A Clinicopathologic and Molecular Study of 17 Patients

Objective  To identify potential prognostic factors and criteria for early detection of malignant peripheral nerve sheath tumors associated with neurofibromatosis type 1 (NF1).

Design  Retrospective study of malignant peripheral nerve sheath tumors in a cohort of 395 patients with NF1 followed up between October 1, 1988, and January 1, 1999; review of the clinical and histological characteristics of treatment and course; and analysis of p53 mutations and overexpression in tumors.

Setting  Teaching hospital referral neurofibromatosis center for adults.

Patients  Seventeen patients with NF1 (9 males and 8 females). Mean ± SD patient age at diagnosis was 32 ± 14 years.

Main Outcome Measures  (1) Clinical symptoms, (2) comparison of p53 mutations and overexpression in benign vs malignant tumors; and (3) median survival.

Results  Twelve patients had high-grade tumors. All tumors except 1 developed on preexisting nodular or plexiform neurofibromas. Pain and enlarging mass were the first and predominant signs. None of the benign tumors displayed significant p53 staining or p53 mutations. Six of 12 malignant tumors significantly overexpressed p53, and 4 of 6 harbored p53 missense mutations. Median survival was 18 months overall, 53 months in peripheral locations, and 21 months in axial locations.

Conclusions  Malignant peripheral nerve sheath tumors are highly aggressive in NF1. They mostly arise from plexiform or nodular neurofibromas. Investigations and deep biopsy of painful and enlarging nodular or plexiform neurofibromas should be considered in patients with NF1. Late appearance of p53 mutations and overexpression precludes their use as predictive markers of malignant transformation.

NEUROFIBROMATOSIS type 1 (NF1) is a common human genetic disease with an incidence of approximately 1 in 2500 to 3300 persons, an autosomal dominant mode of inheritance, and a high rate of new mutations.^1,2 Despite the presence of serious intractable complications, the priority of the patients is for cosmetic correction of the disease; therefore, the dermatologist and the plastic surgeon are frequently the first-line practitioners.³ Among these complications, malignant peripheral nerve sheath tumors (MPNSTs) (formerly neurofibrosarcoma) are the main cause of mortality in adult patients with NF1.^4-8 Diagnosis delays are frequent; symptoms are usually present for months or even years before the malignant neoplasm is recognized.

The molecular events involved in the malignant transformation of neurofibromas are poorly defined. The NF1 tumor suppressor gene is located on band 17q11.2 and encodes a protein called neurofibromin, which is involved in the control of the ras signal transduction pathway.^9-11 Inactivation of both copies of the NF1 gene has been demonstrated in benign neurofibromas,¹² hence before malignancy occurs. Chromosome arm 17p deletions, probably targeting the p53 tumor suppressor gene, have been associated with tumor progression.^13,14 p53 alterations have been further reported^15-17 in sporadic and NF1-associated MPNSTs. In these tumors, p53 overexpression has been linked to a poor prognosis.^18,19

Our aim was to identify potential prognostic factors of MPNST in NF1 and criteria allowing early detection of this complication. Thus, we retrospectively studied the clinical, histological, and molecular characteristics of MPNST cases in our multidisciplinary neurofibromatosis clinic, the Réseau NF-Mondor, Créteil, France, a referral center for adults.

Between October 1, 1988, and January 1, 1999, 395 patients with NF1 according to the National Institutes of Health consensus criteria²⁰ were seen at the Réseau NF-Mondor. Median follow-up was 4 years. Among these patients, we retrospectively studied those with histologically proven MPNSTs.

The medical chart was reviewed for each patient, and the following data were collected: demographic information (age and sex), date of diagnosis, signs at diagnosis (pain, enlargement of a mass, and neurological symptoms) and onset of these signs, location and size of the tumor, presence or absence of preexisting neurofibromas, treatment, and evolution.

Tumor specimens were reviewed by 2 of us (M.-C.V. and J.W.) for histopathological grading. We used the French grading system,²¹ which is based on a score generated by evaluating 3 variables: tumor differentiation, mitotic rate, and amount of tumor necrosis. Grade I corresponds to low grade, grade II to intermediate grade, and grade III to high grade.

Paraffin-embedded biopsy specimens were assayed using the following method: (1) deparaffinization of 5-µm tissue sections in xylene; (2) rehydration; (3) immersion in 0.01M EDTA buffer followed by 3 microwave treatments (750 W, 5 minutes each); (4) blocking step in 5% nonfat milk in Tris/sodium chloride buffer, pH 7.6; and (5) 1-hour incubation in a 1:50 dilution of the anti-p53 monoclonal antibody Do-7 (Dako SA, Trappes, France). Monoclonal antibody Do-7 was then revealed with the standard alkaline phosphatase and anti–alkaline phosphatase technique. The study was performed in MPNSTs and in neurofibromas if previously or simultaneously obtained from the same area. In all tumors, the intensity and extent of p53 reactivity (expressed as percentage of nuclear staining) were evaluated. Only nuclear staining was regarded as specific. Intensity of staining was graded as weak, moderate, or strong.

p53 gene mutations were studied in available frozen material from MPNSTs and neurofibromas (if previously or simultaneously obtained from the same area). DNA was prepared from frozen tissue sections by proteinase K digestion and phenol-chloroform extraction after hematein-eosin-safran coloration of a 5-µM control section. Mutations in exons 5 to 8 of the p53 gene were detected by analysis of the electrophoretic migration profile of amplified DNA samples on denaturing gradient gel electrophoresis, as previously described.²² Mutated DNA samples were amplified with p53-specific oligonucleotides containing M13 primer sequences at their 5′ end and were sequenced on both strands with ready reaction dye-primer sequencing kits (Applied Biosystems, PE France SA, Courtaboeuf, France).

Survival rates were determined using the Kaplan-Meier method. Effects of p53 and MPNST location on prognosis were assessed using the χ² test.

Seventeen patients (9 males and 8 females) with NF1 and MPNSTs were included in the study. Six patients had familial NF1 and 11 were sporadic cases. Table 1 summarizes the demographic and clinical characteristics of these patients. Mean ± SD age at MPNST diagnosis was 32 ± 14 years. The prevalence of MPNSTs in the 395 patients was 4.3% (95% confidence interval, 2.5%-6.8%). Because median follow-up was 4 years, the incidence of MPNST was evaluated to 0.01 per patient per year of follow-up. Mean ± SD delay to diagnosis after the appearance of the first symptom was 13 ± 11 months.

Table 2 summarizes the initial clinical characteristics of the tumors and treatment. Tumors were located on the lower limbs (n = 4), upper limbs (n = 3), trunk (n = 2), head (n = 5), pelvis (n = 2), and abdomen (n = 1). At the time of diagnosis, the mean ± SD size of the tumors was 14 ± 6 cm. Histoprognostic grading was as follow: grade I (n = 2), grade II (n = 3), and grade III (n = 12). The clinical aspect of the tumor in patient 4 is shown Figure 1. A biopsy specimen of the grade III MPNST from patient 17 is shown Figure 2. All MPNSTs except 1 developed on preexisting nodular or plexiform neurofibromas.

At the time of diagnosis, no metastasis was detected. Surgical treatment of the primary tumor was performed in 15 patients: 6 underwent large excision of the tumor (safety margin, 3-5 cm), 2 underwent amputation, 3 underwent marginal excision (no safety margin), and 4 underwent incomplete excision. Postsurgical radiotherapy was performed in 5 of 6 patients who underwent large excision, 2 of 3 who underwent marginal excision, and 2 of 4 who underwent incomplete excision. Adjuvant chemotherapy was performed in 2 of 6 patients who underwent large excision of the tumor. Palliative chemotherapy was performed in the 4 patients who underwent incomplete surgical excision, 1 who underwent large excision, and 1 for whom no surgery was possible.

Expression of p53 protein was studied in 12 MPNSTs and 6 preexisting neurofibromas either previously or simultaneously removed from the same area (Table 3). p53 exons 5-8 gene mutations were assessed in 6 MPNSTs and 5 related neurofibromas. None of the benign tumors displayed significant p53 staining or p53 mutations. p53 nuclear staining was observed in 8 of 12 MPNSTs studied and was considered significant in 6 (50%). p53 staining of the MPNST in patient 17 is shown in Figure 3. Four of 6 patients studied exhibited p53 missense mutations in exon 5 (patients 10 and 14) and exon 8 (patients 16 and 17). These mutations were mostly transitional events (3 of 4 patients), occurring outside CpG dinucleotides. The corresponding wild-type nucleotide was weak or absent on direct sequencing of the amplified region in patients 10, 14, and 16, reflecting loss or reduction to homozygosity of the wild-type allele (Figure 4). Missense mutations were associated with significant p53 overexpression in 3 patients and scattered positive cells (<5%) in 1.

Median survival of the 17 patients with MPNSTs was 21 months. Median survival was 53 months in peripheral locations vs 21 months in axial locations (Figure 5). This difference was not significant (P = .14). Median survival was 18 months in patients with p53-positive MPNSTs vs 24 months in those with p53-negative MPNSTs. This difference was not significant (P = .65).

The incidence of MPNSTs in patients with NF1 has been estimated to be 2% to 5% compared with 0.0001% in the general population. More than 50% of patients with MPNSTs also have NF1.^23,24 In our series of 395 patients with NF1, the prevalence of MPNSTs was approximately 4%. This prevalence is in accordance with other hospitalized NF1 cohorts,²¹ leading to an evaluated incidence of 0.01 per patient with NF1 per year of follow-up. This high rate should be interpreted with caution and certainly stems from collection bias of referred patients.

These neoplasms are noticed around the fourth decade of life in the general population and seem to appear earlier, around the third decade, in patients with NF1⁸ (32 years in our series). Nevertheless, of the 17 patients in our series, 6 were teenagers and a seventh was only 20 years old. Indeed, MPNST is a critically important complication in the second decade of life in patients with NF1.

Malignant peripheral nerve sheath tumors are highly malignant, and metastases are common. Patients with MPNST and without NF1 have reported 5-year survival of 50%.⁵ Concurrence of NF1 affects the outcome of patients with MPNSTs adversely,⁵ with reported 5-year survival of 15%. The main prognostic factors seem to be the size of the neoplasm, probably linked to the possibility of curative surgery, and the axial location.⁵

In patients with NF1, from a clinical perspective, 4 types of neurofibromas can be distinguished: (1) discrete cutaneous neurofibromas of the dermis or epidermis, (2) subcutaneous neurofibromas that lie deeper in the skin, (3) deep nodular neurofibromas, and (4) diffuse plexiform neurofibromas. Malignant peripheral nerve sheath tumors arise mostly from nodular and plexiform neurofibromas. About half of all MPNSTs develop in the axial location. An MPNST usually presents as a progressively enlarging mass with pain and, later, neurological symptoms. Symptoms are present for months before the malignant neoplasm is identified correctly.⁶ Partial biopsies often do not help make the diagnosis. Thus, to shorten the delay to diagnosis of MPNST, persons with NF1 should be educated to be alert to unexplained pain or changes in growth patterns of their tumors. Such signs or symptoms are usually investigated by magnetic resonance imaging, and suspicious areas might undergo biopsy, but there is a possibility of false-negative results because of the large size of some plexiform tumors. If a biopsy does not reveal malignancy but pain or tumor growth continue, close clinical and radiographic monitoring should ensue, and another biopsy should be considered.

Preferred treatment for these neoplasms is surgical excision,⁶ and long survival times in our series were associated with radical surgery. Patients treated with more than adequate surgery, ie, wide or deep resection and amputation, fared better than those whose neoplasms were removed with margins that were too narrow. One can expect that shortening the delay to diagnosis would lead to a decrease in the size of the tumor and would help the performance of radical surgery (complete and large excision). Radiotherapy and chemotherapy have been used, but, as in other studies, the results were discouraging.^5,6

We did not detect p53 overexpression or mutation in neurofibromas. On the other hand, 50% of MPNSTs assayed overexpressed p53. Indeed, overexpression of p53 is common in MPNSTs associated with NF1.^15-17 Four p53 mutations were detected in the 6 MPNSTs examined. p53 staining was weak and scarce in patient 10 despite a missense mutation of codon 151. It is possible that this particular mutation does not interfere with normal protein turnover, in contrast to other common missense mutations. Three of 4 mutations were associated with loss of the p53 wild-type allele, in agreement with previous studies showing frequent 17p deletions in these tumors.^13,14 To date and including this study, to our knowledge, there have been only 12 reported p53 mutations, including one silent nucleotide substitution, in NF1 solid tumors.^14,25-27 Most of these mutations differ from common p53 hot spot mutational events.²⁸ Six mutations seem to be rare (representing <1% of the 11 106 reported mutations of the p53 database²⁸), and 5 seem to be uncommon (<5% of reported mutations). These few genetic studies are nevertheless strengthened by several immunohistochemical studies showing frequent p53 overexpression in MPNSTs and rare p53 overexpression in neurofibroma areas, suggesting that p53 alterations are probably key elements in progression to malignancy. Indeed, recent identification of a tumor surveillance pathway involving p53 in response to oncogenic, hyperproliferative signals²⁹ might constitute a pathophysiological basis for the recurrent inactivation of p53 in MPNSTs. Oncogenic ras activation causes a cell cycle arrest associated with p16^INK4a and p53 accumulation in normal primary cells.³⁰ This p53 accumulation results from the up-regulation of p19ARF, which binds to and inactivates Mdm2.³¹ It is tempting to speculate that the constitutive ras activation due to NF1 disruption leads to similar antitumorigenic responses and that the cellular proliferation associated with malignant transformation requires inactivation of this p19ARF-p53 tumor surveillance pathway. Because p53 seems to be inactivated in more than 50% of MPNSTs, it would be of interest to determine whether wild-type p53 tumors exhibit p19ARF inactivation, as lymphoblastic leukemias³² or Mdm2 gene amplification, as reported in soft tissue sarcomas.³³ Furthermore, p53 overexpression has been linked to a poor prognosis in soft tissue sarcomas.^33,34 This finding was not confirmed in our series. This might be because of the low number of cases studied or might reflect a bias linked to the location of the tumors and the possibility of curative surgery. From a practical point of view, p53 overexpression seems to be a late event in the malignant transformation of neurofibromas and, therefore, is not useful in the early detection of MPNSTs.

In conclusion, MPNST is a critically important complication in the second and third decade of life in patients with NF1. Malignant peripheral nerve sheath tumor is a highly aggressive neoplastic disease that often arises from plexiform or nodular neurofibromas. To shorten the delay to diagnosis, repeated deep biopsies of painful and enlarging nodular or plexiform neurofibromas should be considered in patients with NF1. The presence of p53 overexpression or mutations lends support to the notion that p53 alterations play a role in the development of MPNSTs from NF1-associated neurofibromas, but their late appearance precludes their use as a predictive marker of malignant transformation.

Accepted for publication March 12, 2001.

This study was supported in part by funds from Paris XII University, Créteil, France.

Presented in part at the Journées Dermatologiques de Paris, Paris, France, November 22, 1996, and at the Congrès Annuel de Recherche Dermatologique, Paris, December 2, 1998.

Corresponding author and reprints: Pierre Wolkenstein, MD, PhD, Department of Dermatology, Henri-Mondor Hospital, F-94010 Créteil CEDEX, France (e-mail: pierre.wolkenstein@hmn.ap-hop-paris.fr).