Indian Journal of Cancer Medknow Publications on behalf of Indian Cancer Society ISSN: 0019-509X EISSN: 1998-4774 Vol. 48, Num. 3, 2011, pp. 328-334

Indian Journal of Cancer, Vol. 48, No. 3, July-September, 2011, pp. 328-334

Original Article

The role of magnetic resonance imaging in the diagnostic evaluation of malignant peripheral nerve sheath tumors

A Chhabra¹, T Soldatos¹, DJ Durand¹, JA Carrino¹, EF McCarthy², AJ Belzberg³

¹ Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins Hospital, 601 North Caroline Street, Baltimore, MD, USA
² Russell H. Morgan Department of Pathology The Johns Hopkins Hospital, 601 North Caroline Street, Baltimore, MD, USA
³ Russell H. Morgan Department of Neurosurgery, The Johns Hopkins Hospital, 601 North Caroline Street, Baltimore, MD, USA
Correspondence Address: A Chhabra, Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins Hospital, 601 North Caroline Street, Baltimore, MD, USA, achhabr6@jhmi.edu

Code Number: cn11087

PMID: 21921333

Abstract

Keywords: Magnetic resonance imaging, neurofibromatosis type-1, peripheral nerve sheath tumors

Introduction

Malignant peripheral nerve sheath tumors (MPNSTs) are sarcomas thought to arise from pre-existing benign peripheral nerve sheath tumors (BPNSTs), such as neurofibromas or schwannomas. ^[1] MPNSTs may arise in subjects with neurofibromatosis or in a sporadic fashion. Although the survival rate of patients with MPNSTs is poor, early diagnosis and treatment may improve the prognosis. ^[2],[3] In the general population, MPNSTs are exceedingly rare, with an incidence of 0.001%. Among patients with neurofibromatosis type-1 (NF1), however, MPNSTs are the leading cause of mortality, with an incidence of 4.6-13%. ^[4],[5] MPNSTs may manifest clinically with persistent pain, rapid enlargement and neurological deficit; however, the above symptoms may also be present in some BPNSTs. ^[2],[6] A non-invasive test to discriminate MPNSTs from BPNSTs could serve as a screening tool in individuals with NF1 and would assist in avoiding a large number of unnecessary biopsies. ^[4] Magnetic resonance imaging (MRI) has been considered the mainstay for the non-invasive evaluation of soft-tissue abnormalities. Imaging differences among MPNSTs of NF1 versus non-NF1 cases have not been described before. The objective of this study was to determine the potential differentiating MRI features between (a) MPNSTs of NF1 patients and non-NF1 patients and (b) MPNSTs and BPNSTs.

Materials and Methods

Patient population

Institutional review board approval was granted and informed consent was waived for this HIPAA-compliant study. The pathology database over a period of almost 14 years (November 1996 to July 2010) was searched for cases with a diagnosis of MPNST and BPNST. Twenty-one subjects with a diagnosis of MPNST and 35 subjects with a diagnosis of BPNST (24 cases of neurofibromas and 11 cases of schwannomas) were enrolled, creating a total study group of 56 patients (30 males and 26 females, mean age 50 ± 17 years, age range 15-92 years). The patients' medical records were evaluated for history of NF1 and presenting symptoms, such as regional pain, enlarging mass and skeletal muscle weakness. Of the 56 MRI examinations, seven were submitted by the patients in printed films and were digitized using dedicated scanning software (PacsSCAN, PacsGear, Pleasanton, CA, USA), five were performed at affiliated institutes and DICOM images were submitted via the web and 44 were performed in our institution on 1.5T or 3T scanners (Avanto, Verio, Trio, Siemens Medical Solutions, Erlangen, Germany). Although variable imaging protocols were employed, all studies included T1-weighted, T2-weighted and fat-suppressed T2-weighted or STIR images. Forty-two (75%) examinations additionally included post-contrast T1-weighted images. The fields of view ranged from 35 to 40 cm for neoplasms located in the axial body, whereas smaller fields of view were employed for lesions situated in the extremities.

Study methods

All MRI examinations were reviewed in consensus by two radiologists with 14 (A.C.) and 6 (T.S.) years of radiology experience, respectively. Both readers were blinded to the original reports of the MRI studies. In all examinations, the lesions were assessed in terms of size (maximum diameter on T2-weighted images), margins (well- or ill-defined on T2-weighted images), T1 and T2 signal characteristics as compared with skeletal muscle and internal architecture (homo- or heterogeneous on T2-weighted images). If post-contrast images were available, the enhancement characteristics (central, peripheral or both), necrosis or cystic degeneration (non-enhancing area[s] of T2 hyperintensity) were registered for each neoplasm. In addition, all lesions were evaluated for the presence of invasion of adjacent tissues (vessels, bones and/or muscles) as well as for the presence of previously described characteristics of BPNSTs, such as tail sign, split-fat sign and target sign. ^[7],[8] A thorough search through the patients' records was also performed for conventional radiographs and computed tomography (CT) scans, on which lesions were assessed for the presence of calcifications and for ^[18] fluorodeoxyglucose positron emission tomography combined with CT (FDG PET/CT) studies, on which the standardized uptake values (SUV _max ) of the respective lesions were registered.

Statistical analysis

Summary data are expressed as mean ± standard deviation, unless otherwise indicated. Student's t-test for independent samples was employed to evaluate the differences in age and tumor size between patients with MPNST and patients with BPNST as well as between NF1 and non-NF1 subjects with MPNSTs. Fisher's exact test was used to compare the incidences of the various MRI features between MPNSTs and BPNSTs as well as between MPNSTs of NF1 subjects and MPNSTs of non-NF1 subjects. In addition, receiver operating characteristic (ROC) curves was obtained to specify the optimal cut-off value of tumor size for the prediction of malignancy. Optimal cut-off value was considered the threshold value that maximized the sum specificity and sensitivity. A probability value of less than 0.05 was considered as statistically significant. All data were stored on a spreadsheet (Excel 2007, Microsoft, Seattle, WA, USA) and analysis was performed using a commercially available statistical package (MedCalc 8.0, MedCalc Software, Mariakerke, Belgium).

Results

The demographics of the 21 subjects with MPNSTs along with the presenting symptoms and MRI characteristics of the respective neoplasms are summarized in [Table - 1]. The respective characteristics of all 56 patients and their neoplasms are presented in [Table - 2]. There was no difference in sex distribution between NF1 and non-NF1 subjects or between patients with BPNSTs and patients with MPNSTs. All 21 (100%) MPNST cases and 33 (94%) BPNST cases presented with multiple combinations of symptoms such as pain, enlarging mass and/or muscle weakness, whereas two (6%) BPNST cases presented as incidental findings on a routine chest radiograph and a prostate cancer follow-up MRI examination, respectively. There was no difference in the type of the presenting symptoms between NF1 and non-NF1 subjects or between patients with BPNSTs and patients with MPNSTs. The NF1-related MPNSTs presented at an earlier age compared with the non-NF1-related MPNSTs. Similarly, the patients with MPNSTs were significantly younger compared with the patients with BPNSTs, and additionally demonstrated a higher incidence of NF1.

With respect to the skeletal muscles, all neoplasms in the study exhibited intermediate T1 [Figure - 1]a and [Figure - 2]a and high T2 signal intensity [Figure - 1]b and [Figure - 2]b. In the contrast-enhanced studies, all tumors demonstrated enhancement [Figure - 1]c and [Figure - 2]c. Compared with the non-NF1-related MPNSTs, the NF1-related MPNSTs demonstrated a significantly higher incidence of necrosis/cystic degeneration. Compared with the BPNSTs, the MPNSTs were significantly larger at the time of the diagnosis. The optimal cut-off value of tumor size for predicting malignancy was 6.1 cm. The sensitivity and specificity of the above value were 62% and 86%, respectively. MPNSTs exhibited significantly higher frequency of ill-defined margins and invasion of adjacent structures [Figure - 3]a and b. The specificity, sensitivity and positive and negative predictive values of ill-defined margins in predicting malignancy were 91%, 52%, 79% and 76%, respectively. For invasion of adjacent structures, the corresponding values were 100%, 43%, 100% and 73%, respectively.

No differences existed between NF1-related MPNSTs and non-NF1-related MPNSTs as well as between BPNSTs and MPNSTs regarding their location, type of internal architecture or contrast enhancement (central and/or peripheral), incidence of intralesional hemorrhage and presence of tail-, target- and split-fat signs.

From the patients with MPNST, three underwent both CT and radiography, five underwent CT alone and three underwent radiography alone, whereas from the patients with BPNST, one underwent both CT and radiography, six underwent CT only and one underwent radiography alone. In the above examinations, tumor calcification was observed in only one patient with MPNST. Furthermore, FDG PET/CT examinations were performed in five patients with MPNST, revealing an SUV _max ranging from 3.0 to 10.3, and in two patients with BPNST, revealing an SUV _max of 2.6 and 3.7, respectively.

Discussion

NF1 is an autosomal-dominant multisystem disorder characterized by orthopedic and neurocutaneous manifestations. ^[9] Nerve sheath tumors are an inherent characteristic of NF1 and may develop as cutaneous, subcutaneous and plexiform neurofibromas. ^[6] These neoplasms carry a potential to degenerate into MPNSTs, which tend to metastasize widely, heralding a poor outcome. ^[4] Because of the extensive morbidity associated with MPNSTs, in 2002, an international consensus group suggested that individuals with NF1, particularly those with many or very extensive plexiform lesions, warrant increased surveillance for the development of MPNST. ^[2] Because of its multiplanar imaging capabilities, high spatial resolution and superb soft-tissue contrast, MRI has been emerging as a strong candidate for the regular screening of NF1 patients and the definite non-invasive determination of the histological type of peripheral nerve sheath tumors. To the best of our knowledge, there are no reports in the literature describing the potential differences in MRI features between NF1-related and non-NF1-related MPNSTs that could provide additional assistance to the MRI readers when assessing studies of NF1 patients. Furthermore, divergence exists in the literature regarding the differentiating MRI features between MPNSTs and BPNSTs and therefore additional data could increase the accuracy of the modality in discriminating the two types and assist in avoiding a large number of unnecessary biopsies.

Because the evaluated MRIs were not screening studies, the patients presented with various combinations of pain, skeletal muscle weakness and enlarging mass. However, none of the above manifestations proved helpful in differentiating between the types of tumors studied. According to our data, the MPNSTs of NF1 patients presented at a significantly younger age compared with MPNSTs of non-NF1 patients, a finding that was probably expected based on the higher incidence of peripheral nerve sheath tumors in NF1.

All tumors of the study were isointense to skeletal muscle on T1-weighted images and hyperintense on T2-weighted images, a result that is in accordance with those of previous authors. ^[7],[10] However, central and peripheral contrast enhancement, which have been previously reported as specific for BPNSTs and MPNSTs, respectively, ^{[11],[12],[13]} were not observed in our study. NF1-related and non-NF1-related MPNSTs also demonstrated a similar type of contrast enhancement. On the other hand, the MPNSTs of NF1 subjects exhibited a higher incidence of necrosis/cystic degeneration compared with the respective lesions of non-NF1 subjects, a differentiating feature that has not been previously reported. No other significant differences were observed between the MPNSTs of these two groups. In agreement with previous authors, ^{[12],[14],[15],[16]} the presence of large size, ill-defined margins and/or invasion of adjacent structures proved highly specific, however not sensitive, for predicting malignancy in peripheral nerve sheath tumors. Other characteristics such as location, type of internal architecture, calcification, intralesional hemorrhage and presence of tail-, target- and split-fat signs did not prove helpful in discriminating NF1-related MPNSTs from non-NF1-related MPNSTs or BPNSTs from MPNSTs. The above findings suggest that biopsy remains the only definitive way to exclude malignancy in case the lesion is non-invasive.

In the relatively limited number of studies that have evaluated the MRI features of MPNSTs, previous authors have focused on several different imaging criteria and have concluded on varying results. Bhargava et al.^[17] compared MPNSTs with BPNSTs regarding the presence of the target sign and concluded that the latter should be used to establish benignity in peripheral nerve sheath tumors. Ogose et al.^[11],[12] suggested that absence of motor weakness and central enhancement are indicative of BPNST, whereas severe rest pain and ill-defined margins are suggestive of MPNST. Van Herendael et al.^[13] reported heterogeneity on T1-, T2- and contrast-enhanced T1-weighted images as malignant MRI features of peripheral nerve sheath tumors. According to Li et al., ^[14] well-defined margins and presence of the split-fat sign are MRI features of benignity, whereas large size and infiltrative margins are features of malignancy. Matsumine et al.^[16] provided several imaging criteria that can discriminate MPNSTs from BPNST, including an irregular tumor shape, unclear margins, intratumoral lobulation, high signal intensity on T1-weighted images, absence of target sign and heterogeneous enhancement. In the most recent and largest study available, Wasa et al.^[15] advocated that large size, peripheral enhancement, perilesional edema and cystic degeneration are suspicious of malignancy, and that if a lesion has two or more of the above four features, it should be subjected to a biopsy. In NF1 patients, the authors suggested that heterogeneity on T1-weighted images can differentiate MPNSTs from BPNSTs. In the aforementioned studies however most of the suggested MRI features failed to combine a high sensitivity with a high specificity, reflecting the relatively limited accuracy of MRI in characterizing peripheral nerve sheath lesions.

It has recently been proposed that post-operative MRI follow-up may be useful for the early detection of MPNST recurrence. ^[18] Our data and the results of previous authors suggest that MRI should be employed for pre-surgical planning and post-surgical surveillance of MPNSTs. Unless the lesion is invasive or metastatic, attempts to differentiate BPNSTs versus MPNSTs by means of MRI may be futile. In addition, if surveillance for malignancy is to be performed in NF1 patients to assess rapid interval growth or invasion, MRI should be instituted at a younger age due to the earlier onset of malignancy in these subjects.

Current trends in the management of NF1 have also focused on the use of FDG PET/CT as a tool for the early detection of sarcomatous transformation. The initial results have been promising as the method may detect malignant change in plexiform neurofibromas; however, it remains difficult to differentiate between low-grade MPNSTs and benign lesions. ^[2] In this study, FDG uptake was noted in 5/5 cases of MPNSTs, but also in 2/2 cases of BPNSTs [Figure - 2]d. In addition, one of the above five patients with MPNST had a history of NF1, and the respective FDG PET/CT study also revealed intense uptake by multiple other lesions elsewhere in the body, presumably BPNSTs, and with SUV _max similar to the index MPNST. Although evaluating the accuracy of FDG PET/CT was not the primary endpoint of our study, the results of this series indicate a potential inability of FDG PET/CT to accurately differentiate MPNSTs from BPNSTs.

In summary, our study showed that NF1-related MPNSTs have an increased incidence of necrosis/cystic degeneration compared with non-NF1-related MPNSTs, and the above feature should alert MRI readers for the presence of malignancy when reporting studies of NF1 patients. Our results also confirmed the considerable overlap that exists between BPNSTs and MPNSTs regarding their intralesional MRI characteristics and presenting symptoms. However, peripheral nerve sheath tumors that demonstrate a large size, ill-defined margins and/or invasion of adjacent structures on MRI studies should be considered as highly suspicious for malignancy.

References

Copyright 2011 - Indian Journal of Cancer

The following images related to this document are available:

Photo images