| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Massachusetts General Hospital, 55 Fruit St., Boston,
MA 02150.
2 Present address: Department of Radiology, Brigham and Women's Hospital,
Boston, MA 02115.
3 Department of Radiology, Children's Hospital of Philadelphia, 34th Street and
Civic Center Blvd., Philadelphia, PA 19104.
4 Department of Radiology, Children's Hospital Boston, Boston, MA 02115.
5 Department of Medicine, Massachusetts General Hospital, Boston, MA
02114.
6 Department of Genetics, University of Alabama at Birmingham, Birmingham, AL
35294.
Received April 2, 2004;
accepted after revision July 20, 2004.
Supported by U.S. Army grant no. NF70002, "Natural History of
Plexiform Neurofibromas in NF1."
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Sixty-six patients (median age, 15 years) with superficial plexiform neurofibromas were compared with 56 patients with deep plexiform neurofibromas (median age, 12 years). All patients underwent axial STIR and coronal or sagittal STIR images.
RESULTS. Superficial neurofibromas were more likely to be asymmetric (p = 0.004) and extend to the skin surface (p < 0.001). Lesion borders were poorly defined with similar frequency in both superficial and deep groups (77% vs 68%, p = 0.31). The morphology of superficial neurofibromas was more likely diffuse (64% vs 11%, p < 0.001), whereas deep neurofibromas were more likely nodular or fascicular. Of neurofibromas that were nodular or fascicular in morphology, superficial lesions had a smaller maximal fascicle–nodule diameter (mean, 10.3 mm) than deep lesions (mean, 13.4 mm) (p = 0.013). Signal characteristics of deep neurofibromas were more likely to be targetlike (75%) compared with superficial neurofibromas (21%) (p < 0.001). Superficial neurofibromas had a smaller mean volume than deep neurofibromas (180 vs 444 cm3, p = 0.002).
CONCLUSION. Unlike the typical targetlike lesions along the course of major nerves seen in deep plexiform lesions, superficial plexiform neurofibromas in NF1 tend to be asymmetric, have nontargetlike signal intensity, lack nodular or fascicular morphology, and are likely to involve skin.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Plexiform neurofibromas often involve nerve plexuses, dorsal nerve roots, and other structures deep in relation to the muscle fascia and do not have any evident superficial extension. These deep lesions follow the course of major nerves. Past MRI studies have focused primarily on deep plexiform neurofibromas and show that they typically have a targetlike appearance on T2-weighted MR images, with central low signal intensity and peripheral high signal intensity [1–4].
Other plexiform neurofibromas occur superficially and can be cutaneous or subcutaneous [1]. They can arise from peripheral nerves with no deep involvement, or they can represent superficial extension of a deeper plexiform neurofibroma [1]. Cutaneous involvement and subcutaneous involvement in NF1 are common. McGaughran et al. [5] investigated the clinical diagnostic features of NF1, including the prevalence of cutaneous and subcutaneous neurofibromas. They reported that well-circumscribed cutaneous neurofibromas were seen in 217 (59.5%) of 365 patients with NF1 and that 150 (45.5%) of 330 patients with NF1 had discrete subcutaneous neurofibromas. In addition, they found that 80 (15%) of 523 patients with NF1 had diffuse subcutaneous plexiform neurofibromas. Iannicelli et al. [6] found that in a group of 46 patients with NF1 with soft-tissue lesions, 33 had deep plexiform neurofibromas and 18 had subcutaneous neurofibromas.
Imaging evaluation of superficial neurofibromas has been limited. In the series by Iannicelli et al. [6], the sonographic appearance of subcutaneous neurofibromas was described as either platelike or possessing deep digitations. The appearance of these superficial lesions on MRI has not been well characterized, with only a few cases presented in the literature [7–9]. On the basis of these limited cases, superficial plexiform neurofibromas do not typically possess the targetlike appearance seen in their deep counterparts and can be easily mistaken for other entities, such as venous malformations [8].
Our hypothesis is that superficial plexiform neurofibromas are common and have MRI characteristics that are different from their deep counterparts; these lesions need to be recognized and differentiated from other superficial abnormalities. It is the goal of our study to evaluate the MRI characteristics of superficial plexiform neurofibromas in NF1 and to compare them to the characteristics of deep plexiform neurofibromas. This study is part of a multiinstitutional trial that we are conducting to determine the natural history of NF1, specifically by collecting growth data on plexiform neurofibromas. It is our goal to correlate these growth data with the MRI characteristics of deep and superficial neurofibromas. Ultimately, we hope to determine whether a faster rate of growth is associated with certain imaging characteristics. Imaging-based prognostic indicators are potentially a very useful tool and could be used to triage which patients should pursue more aggressive therapy or monitoring.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
MR images from each patient were sent to a central location from each of these institutions. Human research committee approval was obtained at each institution, and informed consent was obtained from all participants.
MRI
For all patients, contiguous axial STIR (TR/TE, 6000/35; inversion time,
150 msec; echo-train length, 8) images and coronal or sagittal STIR images
were obtained. STIR is a fluid-sensitive, fat-suppressing sequence that was
chosen for the imaging protocol because of the very bright signal
characteristics of neurofibromas relative to surrounding muscle, fat, and
bone. Contrast-enhanced imaging was not performed for two main reasons: first,
Iannicelli et al. [6] showed
that plexiform neurofibromas are consistently bright in T2 signal intensity
and therefore are very conspicuous on STIR and other T2-weighted sequences. In
addition, they found that plexiform neurofibromas show variable enhancement
after IV gadolinium administration on T1-weighted sequences; therefore, their
conspicuity on contrast-enhanced images is inconsistent. Because IV contrast
material provides no gain in conspicuity of plexiform neurofibromas, it was
decided that there was no justification for the added risk or expense of
gadolinium-enhanced imaging in this group of voluntary subjects.
Because patients were referred as part of a multiinstitutional study, MRI was performed on scanners of different makes and field strengths, with 35 magnets at 1.5 T and one magnet at 1.0 T. Coil, slice thickness, and field of view varied according to the location and extent of the lesion. Slice thickness was 4 mm for the head and neck, 5 mm for the trunk, and 10 mm for the extremities. The extremities were imaged with a matrix of 512 x 160; other anatomic locations were imaged with a 256 x 256 matrix.
Image Analysis
One lesion in each patient was evaluated. If a patient had multiple
lesions, only the largest of the lesions was evaluated. We defined superficial
neurofibromas as lesions arising between the skin and muscle fascia. We
defined deep neurofibromas as lesions arising underneath the muscle fascia.
Some lesions contained both deep and superficial components. When a lesion had
a predominantly (> 50%) deep component with a smaller superficial
component, the lesion was considered a deep neurofibroma. Similarly, when the
lesion was predominantly superficial (> 50%), with a smaller deep
component, it was considered a superficial neurofibroma.
Each lesion was categorized by one radiologist as either being superficial or deep. The location within the body of each neurofibroma was then noted by the same radiologist. Location was categorized into one of the following three categories: trunk (chest, abdomen, and pelvis), extremity, or head and neck.
Each lesion was then graded with regard to numerous characteristics. Ten patients were reviewed jointly by two radiologists to establish a grading system of these characteristics. The remaining 112 patients were reviewed by one radiologist. Subsequently, 50 of these 112 patients were reviewed independently by the second radiologist, because this was the number of examinations required to establish the level of interobserver agreement (i.e., kappa statistic).
Note was made of whether the lesion was unilateral, bilateral and symmetric, or bilateral and asymmetric. The patient's midline was used as the axis of symmetry. Lesions involving a single extremity were categorized as unilateral. In addition, it was noted whether the lesion extended to skin surface (Figs. 1 and 2) and whether superficial lesions extended deep in relation to the muscle fascia.
|
|
The border definition of each lesion was evaluated and categorized as poorly defined or well defined (Figs. 3A, 4A, 5, and 6A). Poorly defined margins were those with a wide zone of transition, greater than or equal to 5 mm, between lesion and normal surrounding tissue or lesions having margins with numerous small digitations, smaller than 5 mm in diameter. Well-defined borders were those with a narrow zone of transition, less than 5 mm.
|
|
|
|
The morphology of each lesion was categorized as fascicular–nodular versus diffuse. Fascicular–nodular lesions were those consisting of a collection of smaller components that were tubular or spherical or both (Figs. 4B and 7). Diffuse lesions were those that lacked any definable geometry (Fig. 8). For each neurofibroma that was fascicular–nodular in morphology, the largest fascicle or nodule within the lesion was identified, and then its diameter was measured using Cheshire image analysis software (Parexel).
|
|
|
The signal characteristics of each lesion were categorized as homogeneous, targetlike, or heterogeneous without targets (Figs. 3B, 6B, and 9). Targetlike lesions are round, hypointense centrally, and hyperintense peripherally. Each lesion was evaluated for the presence of increased vascularity, as characterized by the presence of two or more flow voids (Figs.9 and 10) that were increased in size or number or both compared with the normal side or with similar uninvolved regions of the affected side in cases in which the contralateral side was not imaged. Finally, the volume of each lesion was calculated using the Cheshire image analysis software, which uses a maximum-intensity-projection algorithm to define the borders of the lesion. The borders of each neurofibroma, as calculated by the software, were then confirmed by a radiologist (one for body lesions and another for head and neck lesions) before volume calculation was performed. This additional review by the radiologist was necessary to exclude any T2 bright nonneurofibroma tissue inadvertently included by the software.
|
|
|
|
Statistical Analysis
We compared the MRI characteristics of deep and superficial neurofibromas.
For categoric outcomes, chi-square tests were used to compare the statistical
difference between outcomes. For continuous outcomes, mean or median was used
to summarize the central tendency and Wilcoxon's rank sum tests were used to
compare the distributions. The level of interobserver agreement (kappa
statistic) was calculated on the basis of the 50 cases that had been reviewed
independently by two radiologists. According to Landis and Koch
[12], on the basis of kappa
values, agreement can be classified as almost perfect (
=
0.81–1.00), excellent ( = 0.61–0.80), good or moderate
( = 0.41–0.60), fair ( = 0.21–0.40), slight (
= 0.00–0.20), and poor ( = 0.00 or negative).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Border definition (Figs. 3A, 4A, 5, and 6A) is poor with similar frequency (77% vs 68%) in the superficial and deep groups, respectively. Morphology (Figs. 4B, 7, and 8) of superficial neurofibromas was more likely to be diffuse (64% vs 11%). Of lesions that were fascicular or nodular in morphology, the superficial lesions had a smaller maximal fascicle–nodule mean diameter of 10.3 mm than that of their deep counterparts (mean, 13.4 mm).
Signal characteristics (Figs. 3B, 6B, and 9) of superficial neurofibromas were infrequently targetlike (21%) and more likely to be homogeneous (35%) or heterogeneous without targets (44%); deep neurofibromas were more likely to be targetlike (75%) than homogeneous (4%) or heterogeneous without targets (21%). Superficial neurofibromas and deep neurofibromas exhibited increased vascularity (Figs. 9 and 10) with similar frequency (36% vs 48%).
The volume of the superficial neurofibromas was significantly smaller, with an average lesion volume of 180 cm3 for superficial neurofibromas and 444 cm3 for deep neurofibromas.
Subsequent MRI had been performed for 72 of our 122 patients, at follow-up intervals varying from 2 to 39 months. The rate of lesion growth was calculated for these patients in terms of both volume change and percentage of volume change over time. Although some of the lesions decreased in size, the overall pattern was one of lesion growth for both superficial and deep groups. The mean rate of volume growth for superficial neurofibromas was 18.7 cm3/year (median, 1.6 cm3; range, –185.0 to 399.1 cm3); for deep neurofibromas, it was 114.8 cm3/year (median, 25.8 cm3; range, –278.9 to 2,346.5 cm3) (p = 0.09). The mean rate of percentage growth for superficial neurofibromas was 40.0% per year (median, 16%; range, –100% to 556%); for deep neurofibromas, it was 19.3% per year (median, 10%; range, –69% to 275%) (p = 0.68).
Interobserver agreement ranged from very good to excellent for all the lesion characteristics evaluated, and simple kappa statistics ranged from 0.70 (excellent) to 1.0 (almost perfect). Levels of interobserver agreement are summarized in Table 2 [12].
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The targetlike MRI appearance of deep neurofibromas reflects their histologic composition. Woodruff [13] describes deep plexiform neurofibroma as microscopically consisting of a nerve or nerve fascicle distended by tumor cells, embedded in a rich myxoid matrix. Longitudinal bundles of residual nerve fibers are often centrally situated in the neurofibroma. This architecture could account for the centrally T2 dark (nerve fibers) and peripherally T2 bright (myxoid) appearance of these targetlike lesions [14, 15]. Woodruff [13] also described the histology of diffuse cutaneous and subcutaneous neurofibromas as spindle cells infiltrating around normal structures, such as skin adnexa, blood vessels, and adipose tissue. The ability of superficial neurofibromas to have poorly defined borders on MRI could be explained by this infiltrating microscopic architecture.
It is important to recognize the particular MRI characteristics of superficial neurofibromas for two main reasons. First, not all plexiform neurofibromas have the typical targetlike appearance. The absence of the targetlike appearance should not speak against a lesion being a neurofibroma, particularly if the lesion has a superficial location. In addition, it is important to consider neurofibromatosis as a differential diagnostic possibility when encountering an infiltrating, branching, highsignal-intensity lesion. These lesions may closely resemble a lymphatic or venous malformation, a hemangioma (when there are multiple associated flow voids), or, less commonly, a traumatic or inflammatory lesion of the subcutaneous tissues (Fig. 11).
|
One limitation of this study is that for reasons described in Subjects and Methods, we did not evaluate whether there is a difference in the contrast-enhanced appearance of deep versus superficial plexiform neurofibromas. Although Iannicelli et al. [6] reported inconsistent enhancement in their series of 46 patients, it may be of future interest to gather additional data regarding contrast enhancement of these lesions to determine whether the pattern of enhancement has any relationship to their biologic behavior. The existence of such a relationship is suggested by Mautner et al. [16], who described three asymptomatic NF1 patients with inhomogeneously enhancing plexiform neurofibromas on MRI. All three of these lesions were in fact shown to be malignant peripheral nerve sheath tumors.
The therapeutic options for superficial neurofibromas are limited. Surgical management offers limited benefit, particularly for superficial small, asymptomatic neurofibromas [17]. It is currently impossible to predict which small neurofibromas will eventually grow and invade surrounding tissues and which will remain stable in size.
Growth of plexiform neurofibromas can occur at any time in life. Periods of more rapid growth and development of new lesions have been observed in early childhood, during puberty, and during pregnancy. However, the overall growth pattern of plexiform neurofibromas is unpredictable, and even when untreated, spontaneous periods of stability or remission can occur. It is not known whether the location, histology, or radiologic appearance of a plexiform neurofibroma has any relation to its growth pattern. We hope that the study of the natural history of these lesions will help in understanding these questions. Our preliminary results based on 72 patients with follow-up imaging do not show a statistically significant different growth rate between deep and superficial neurofibromas, although the p value approaches statistical significance (p = 0.09) for the change in lesion volume over time. It is our hope that as we acquire more data points, this relationship will become better understood.
This was a multiinstitutional study; therefore, there is variation in the technical factors and the quality of the images. STIR imaging results in high signal intensity of slow-flowing vessels, which are difficult to differentiate from the neurofibromas. Plexiform neurofibromas are very complex lesions, and the categorization into superficial and deep lesions is somewhat arbitrary. Nonetheless, the signal characteristics of the superficial lesions are very different from those of their deep counterparts, and we believe that this classification identifies two groups of lesions that have different histology and possibly biologic behavior. Histologic correlation is not available because, as described under Subjects and Methods, these lesions are not usually biopsied in patients with a clearly established clinical diagnosis of NF1. The data to show whether superficial and deep neurofibromas have different growth rates will not be available for several years. We hope to use these observations to further investigate the natural history of NF1, in the hope that imaging-based prognostic factors can be identified.
In summary, it is important to recognize that superficial plexiform neurofibromas have signal characteristics that differ from the typical targetlike lesions along the course of major nerves seen in deep plexiform lesions. Superficial plexiform neurofibromas in patients with NF1 are infiltrating lesions with diffuse morphology, asymmetrical distribution, and nontargetlike signal intensity; they have smaller fascicles and smaller volumes than deep plexiform neurofibromas.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  D. F. Papp, A. J. Khanna, E. F. McCarthy, J. A. Carrino, A. J. Farber, and F. J. Frassica Magnetic Resonance Imaging of Soft-Tissue Tumors: Determinate and Indeterminate Lesions J. Bone Joint Surg. Am., October 1, 2007; 89(suppl_3): 103 - 115. [Full Text] [PDF]
|
|
|
|
 |
|
 Y-D Han, Y Huan, J-L Deng, Y-G Zhang, and C H-S Zhang MRI appearance of multiple eccrine spiradenoma Br. J. Radiol., January 1, 2007; 80(949): e27 - e29. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. F. Blacksin, D.-H. Ha, M. Hameed, and S. Aisner Superficial Soft-Tissue Masses of the Extremities RadioGraphics, September 1, 2006; 26(5): 1289 - 1304. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Ostlere and R Graham Imaging of soft tissue masses Imaging, December 1, 2005; 17(3): 268 - 284. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Handra-Luca, D Vidaud, M-P Vullierme, N Colnot, D Henin, P Ruszniewski, P Bedossa, and A Couvelard Plexiform neurofibroma mimicking a pancreatic cystic tumour Gut, November 1, 2005; 54(11): 1658 - 1659. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
