Imaging of Soft Tissue Tumors
2nd Edition

9
Neurogenic Tumors
Neuromas are common, reactive, hyperplastic lesions that simulate neoplasms of nerve. True neoplasms of nerve sheath origin are not uncommon soft tissue tumors, making up approximately 12% of all benign and 7% to 8% of all malignant soft tissue neoplasms (1). Imaging characteristics often suggest a neurogenic origin for both reactive and neoplastic lesions. These distinctive features are typically related to location, such as those lesions in the region of a major nerve with an identifiable nerve entering and exiting the mass. Lesions discussed in this chapter include neuroma, schwannoma, neurofibroma, neurofibromatosis (NF), malignant peripheral nerve sheath tumor, neurothekeoma, nerve sheath myxoma, perineurioma, granular cell tumor, melanotic neuroectodermal tumor of infancy, clear cell sarcoma, paraganglioma, and primitive neural tumors.
Development and Histologic Characteristics of Normal Peripheral Nerves
We believe many of the imaging features of neurogenic tumors are a reflection of their similarity to the normal neurogenic tissue from which they are derived. Because of this relationship, an understanding of basic nerve anatomy and histologic characteristics is important.
Peripheral nerves are derived embryologically from neural crest tissue and migrating axons from the primitive neural tube (2). The two predominant supporting elements are the connective tissue stroma and Schwann cells that encase all peripheral nerve axons to varying degrees. A myelinated fiber results if only one axon is encased by one Schwann cell. Unmyelinated fibers result if a Schwann cell encases many axons.
Each peripheral nerve is surrounded by a thick connective tissue sheath called the epineurium. Within the nerve, groups of axons are surrounded and divided by a fibrous stroma called the perineurium, which creates multiple bundles of fibers or fascicles (Fig. 9.1) (2). This gross appearance of a normal nerve can be recognized at ultrasonography and MR imaging, particularly in large nerve trunks such as the sciatic nerve, and is described as having a fascicular appearance (Fig. 9.2) (3,4,5,6). On longitudinal ultrasonography, the normal nerve texture is composed of multiple, hypoechoic parallel but discontiguous, linear bundles separated by hyperechoic bands (7,8). On MR imaging, the normal nerve is seen as multiple small circular (seen en face) or longitudinal (seen in long axis) bundles with punctate areas of higher signal intensity on T1- and T2-weighting. Vascular supply to the peripheral nerves is relatively profuse, arises from adjacent vessels, and forms longitudinally oriented channels along the nerve.
Neuroma
Neuromas are not true soft tissue neoplasms, but represent lesions of nerve caused by reactive hyperplasia (9). Types of neuromas encountered in the soft tissue include traumatic, Morton, Pacinian, and palisaded encapsulated.
Traumatic Neuroma
Traumatic neuroma is not a true neoplasm but represents an attempted, insufficient, reparative proliferation of nerve tissue to regain axonal continuity usually related to the proximal end of a severed nerve (9,10,11,12,13,14,15,16). Typically, this is associated with trauma or amputation. The incidence of traumatic neuroma is markedly reduced, and the likelihood of successful regeneration markedly improved, if the nerve ends are closely approximated following injury (13).
P.329

Figure 9.1 Normal nerve structure. Photomicrograph (original magnification, approximately × 20; Bielschowsky silver stain) of an axial section of normal sural nerve. Nerve is surrounded by epineurium (straight arrows). Bundles of nerve fibers (asterisks) are surrounded by perineurium (curved arrows), creating a fascicular appearance. Adipose tissue (black arrowheads) and blood vessels (white arrowheads) are seen about the nerve.
Clinically traumatic neuroma may be asymptomatic or painful and presents as a firm nodule at a focal pressure site. Pain may be elicited with palpation or tapping on the lesion (Tinel sign). The most common location for traumatic neuromas is the lower extremity, followed by the head and neck (frequently in the oral cavity; more than 50% of these lesions are related to tooth extraction) (2,3,4,5,10,12,13,16,17,18,19,20,21,22,23,24,25,26). Other affected sites include the radial nerve and brachial plexus (27,28,29,30).
Traumatic neuromas are divided into two major categories based on the anatomic location of the tangled, multidirectional, regenerating axonal mass with respect to the proximal nerve end (9,10). Spindle neuromas are internal, focal, fusiform swellings secondary to chronic friction or irritation to a nondisrupted, injured, but intact nerve trunk. Lateral or terminal neuromas are the result of severe trauma with partial avulsion, disruption or total transection of a nerve (9,13,16). These lesions have a bulbous-end morphology in continuity with the normal nerve proximally. They arise 1 to 12 months after transection or injury, vary in size, although they are usually less than 5 cm, with no malignant potential (9,31).
Figure 9.2 Normal nerve appearance at MR imaging. Axial T1-weighted (TR/TE; 500/20) MR image of the upper thigh shows the normal sciatic nerve (arrows) with small circular low-signal-intensity areas, surrounded by a background of mildly higher signal-intensity areas representing the fascicular structure of the normal nerve.
At histologic analysis, traumatic neuromas are nonneoplastic, nonencapsulated tangled masses of axons, Schwann cells, endoneurial cells, and perineurial cells, in a dense collagenous matrix with surrounding fibroblasts (9,19,26). The disorganization of the neurogenic tissue (caused by multidirectional proliferation of cells in an abortive attempt to repair the injured nerve) allows traumatic neuromas to be distinguished from neurofibroma, although discrete bundles or fascicles can be recognized.
There are only limited reports of the ultrasonographic, CT, or MR imaging appearance of traumatic neuromas (14,15,30,31,32,33). Typically, a fusiform mass or focal enlargement with an entering and exiting nerve (spindle type), or only an entering nerve terminating in a bulbous shape (lateral or terminal type), is identified (Fig. 9.3) (14,15,26,31). Ultrasonography and MR imaging are the best modalities to identify the direct relationship of the nerve to the lesion (Fig. 9.3) (34). Peer et al. reported the ability to identify injured nerves accurately by high-resolution sonography in 18 patients (35). In addition, these authors also were able to distinguish among nerve swelling (thickening but intact fascicles), compromising surrounding scar tissue (hypoechoic mass without fascicles), neuroma, and insufficient repair on high-resolution sonography. However, these authors acknowledge that difficulty may occur in distinguishing spindle cell neuromas from nerve swelling (35). Lesions of small and/or superficial nerves may not be detected radiologically or may be seen as nonspecific soft tissue masses without an entering nerve (31). Lesion margins are often well-defined, although some irregularity (likely related to multidirectional cell proliferation) is seen with ultrasonography. These imaging characteristics allow traumatic neuromas to be distinguished from other causes of amputation stump pain, including recurrent malignant tumor, phantom limb pain, bone bruise, stress fracture, lymphadenopathy, osteomyelitis, abscess, bursitis, cellulitis, hematoma, heterotopic bone, foreign bodies, atrophied stump muscles, and cicatrization or scar formation (14,26,31,36). Distinguishing lymphadenopathy from traumatic neuroma after neck dissection can be very difficult (37). Yabuuchi et al. showed that statistically significant features favoring traumatic neuroma in this situation include small short-to-long axis ratio, small short axis diameter, central hypoechoic area, less likely contact with the carotid artery, and hypointense rim on T2-weighted MR images (38).
Figure 9.3 Traumatic terminal neuroma of tibial nerve in a man 33 years of age that developed after mid to distal tibial-level amputation. A,B: Coronal T1-weighted (TR/TE; 500/16) (A) and axial turbo T2-weighted (TR/TE; 4300/126) (B) spin-echo MR images show a mass (arrows) posterior to the tibia. Tubular structure (asterisks) represents the tibial nerve entering the mass. The axial T2-weighted MR image shows heterogeneous signal with a fascicular appearance. C: Axial T1-weighted (TR/TE; 600/16) spin-echo MR image following gadolinium administration shows mild enhancement (arrows).
P.330

P.331

Traumatic neuromas typically have intermediate signal intensity (similar to that of muscle) on T1-weighted MR images and intermediate-to-high signal intensity on T2-weighted images (26,31,32). Their signal intensity is frequently heterogeneous, with a ringlike pattern (“fascicular sign”), which we believe correlates with the histologic morphology of nerve fascicles and is optimally detected on T2-weighted images (Fig. 9.3) (26). Intrinsic ultrasound echogenicity (hypoechoic) and CT attenuation (similar to that of muscle) characteristics of these lesions are nonspecific (4,14,15).
Prevention of traumatic neuroma involves approximating the two severed nerve ends so that successful nerve repair and regeneration results (13,39,40). Nerve graft placement may also be performed if the nerve ends cannot be approximated. Multiple surgical techniques are available to remove the proximal nerve stump from the area of scar, which limits the potential for both traumatic neuroma development and recurrence (13,39,40). Initial conservative therapy, including acupuncture, cortisone injection (ultrasound guidance is useful), transcutaneous and direct nerve stimulation, and physical therapy is successful in up to 50% of patients (15,39,40,41,42). Surgical resection is reserved for patients in whom conservative treatment fails.
Pacinian Neuroma
Pacinian neuromas arise from hyperplasia or hypertrophy of the pacinian corpuscle. Pacinian corpuscles are mechanoreceptors and most numerous in the deep skin layers of the hands and feet, although they also occur in viscera walls, mesentery, and vessel wall adventitia. Reflecting the distribution of these normal structures, pacinian neuromas most frequently present as small superficial masses affecting the hands and feet. In the hand, lesions are most common in the index and long fingers, near the periosteum, along the lateral portion of the proximal and middle phalanges or beneath the flexor tendons at the level of the base of the proximal phalanges. The majority of pacinian neuromas are associated with a history of trauma. They are painful, and there is a female predilection (2:1 ratio). These lesions affect adults most commonly in the fifth and sixth decades of life.
At gross pathology these single or multiple lesions range from small nodules attached to interdigital nerves by a slender stalk, contiguous subepineural nodules, or simply enlargement of a pacinian corpuscle. Histologically, these lesions consist of enlarged or multiple pacinian corpuscles with fibrosis similar to that of a Morton neuroma.
Imaging of pacinian neuroma is not reported, to the best of our knowledge, although we would expect similar findings as seen in Morton neuroma (see subsequent section). Erosion of underlying bone is unusual but has been reported (9). Surgical excision with sparing of the associated nerve is curative, although symptoms occasionally persist until the pacinian corpuscle is resected.
Palisaded Encapsulated Neuroma
The palisaded encapsulated neuroma was initially described by Reed et al. in 1972 (43). Although similar to schwannoma pathologically, these lesions are clinically distinct, developing as small, asymptomatic, subcutaneous nodules affecting the face. Adults are primarily affected without a gender predilection, although some authors report a female predominance. Unlike other neuromas, these lesions are not associated with trauma and are typically encapsulated. Imaging features are not known, although we would expect a nonspecific intrinsic appearance on cross-sectional imaging. Treatment of palisaded encapsulated neuroma with surgical excision is curative.
Morton Neuroma
Morton neuroma is a term used to characterize a nonneoplastic lesion, originally described by Morton in 1876 as a “peculiar painful affection of the foot localized to the fourth metatarsophalangeal articulation” (44). Other terms for this lesion include localized interdigital neuritis, plantar neuroma, and Morton toe or node. It represents perineural fibrosis of the plantar digital nerve. The plantar digital nerve is usually affected at the level of the metatarsal head, and there is frequently an associated inflammatory response about the lesion. The nerve between the third and fourth metatarsals is most frequently involved, followed by that between the second and third metatarsals (9,44,45,46). Morton neuromas are uncommon between the first and second metatarsals and rare between the fourth and fifth (44,46). Similar lesions affecting the hand are described typically in males and are related to chronic occupational or recreational trauma.
Clinically, 90% of patients present with paroxysmal pain, usually elicited by exercise that may radiate into the toes or leg and is relieved by rest (44,45,46). Symptoms are usually
P.332

present for weeks to years, and focal tenderness to palpation is frequent. Compressing the intermetatarsal space may also elicit pain (44,45,46,47,48,49). A mass is not usually palpable related to the lesion, although associated synovial cysts may be clinically evident (50). The lesions are almost invariably unilateral. Marked female predilection (as high as 18:1) has led to the suggestion that the cause is related to irritation from the wearing of high-heeled shoes, which is thought to compress the nerve against the intermetatarsal ligament. Although this pathogenesis intuitively seems correct, it has not been substantiated biomechanically, raising the question of other causes, including ischemia (51).
Asymptomatic lesions may be relatively common, with a prevalence of 30% to 33% reported in two series of 70 and 85 MR examinations, respectively (51). These asymptomatic lesions, unlike symptomatic Morton neuromas, had no significant gender predilection and were also statistically smaller than symptomatic lesions (mean transverse diameter of 4.1 to 4.5 vs. 5.3 to 5.6 mm, respectively) in several studies (51,52,53).
A Morton neuroma appears pathologically as a fusiform enlargement of the plantar digital nerve at its bifurcation, with thickening of the epineural fascicles, perineural fibrosis with high collagen content (Renaut bodies), and loss of the myelinated fibers (51,54,55). These lesions represent a primarily degenerative, rather than proliferative process.
Radiography invariably shows normal findings and is most useful for excluding other causes of pain (56,57). Ultrasonography and MR imaging are superior to CT for identifying Morton neuroma (57,58,59,60,61,62). At ultrasonography, these neuromas commonly appear as round or ovoid, well-defined, hypoechoic masses located just proximal to the metatarsal heads in the intermetatarsal space (55,56,63). Quinn et al. reported lesions to be hypoechoic in 79% of cases, mixed echotexture in 12%, and anechoic in 8% (64). Interdigital nerve continuity could be seen in only 56% of cases (64). Small lesions (<5 mm in size) can be difficult to evaluate with ultrasonography (65). In a series of 45 surgically treated patients, Kaminsky et al. reported false-negative results in two cases related to small lesion size (56). Power Doppler ultrasonography may be a valuable adjunct for identifying these lesions and often shows increased vascularity (Fig. 9.4) (66). In the study by Quinn et al., ultrasonography revealed the prospective diagnosis of Morton neuroma in 85% of 27 cases with other reports of 95% to 98% sensitivity. MR imaging had an accuracy of 90%, positive predictive value of 100%, and negative predictive value of 60% in identification of Morton neuroma in the investigation by Zanetti et al. (54). In our experience and that of others, these lesions are most evident on coronal, small field-of-view, T1-weighted images (51,54,67,68,69,70). Zanetti et al. (51) suggested three MR imaging criteria for diagnosis of Morton neuroma: (a) lesion center in the neurovascular bundle, within the intermetatarsal space, and on the plantar side of the transverse metatarsal ligament; (b) the lesion is well-demarcated (excluding partial volume artifact from the adjacent joint capsule); and (c) the signal intensity of the lesion is similar to that of skeletal muscle on T1-weighted images and less than that of fat on T2-weighted images (likely reflecting high collagen content of fibrosis) (Figs. 9.4 and 9.5). Erikson et al. (67) reported intermetatarsal bursal fluid as an associated finding, proximal to Morton neuroma, in 67% of their cases. Zanetti et al. (51) reported a small amount of bursal fluid in the first three intermetatarsal spaces in 67% of asymptomatic patients. A large amount of intermetatarsal bursal fluid (>3 mm in transverse diameter) or fluid in the fourth intermetatarsal space should suggest an associated Morton neuroma (51).
Morton neuromas are markedly less conspicuous on T2-weighted MR images, making differentiation from surrounding muscle and fat difficult (Fig. 9.5). Use of fat-suppressed T2-weighted sequences may allow better delineation of these lesions. In our experience and that of others, the lesions often, but not invariably, enhance with intravenously administered contrast material (Fig. 9.4) (71,72). In a study of six patients with Morton neuromas, Terk et al. (69) demonstrated that fat-suppressed, contrast-enhanced MR imaging was superior for depicting lesions and detected neuromas in two patients in whom other MR imaging sequences (including fat-suppressed T2-weighted) failed to identity an abnormality. However, Williams et al. (70) reported that only 4 of 11 lesions were visible at enhanced MR imaging.
Weishaupt et al. described improved detection of Morton neuroma on MR imaging in the prone position compared to supine or upright weight-bearing positions (73,74). The supine and upright position may cause more dorsal location of the lesion, creating difficulty in detection. Quinn et al. reported 50% of lesions dorsally on sonography, which may also be a reflection of positional variation of location (73,74).
There are many cases in which metatarsal symptoms not related to Morton neuroma, had a clinical presentation suggesting that diagnosis. Zanetti et al. found that the initial clinical diagnosis of Morton neuroma was changed by MR evaluation in 28% of 54 feet (53). In 41% of cases of MR imaging, there was a significant change in location or number of neuromas, with altering of the treatment plan in 57% of cases (53). This study suggests significant implications and use for MR imaging in patients with a suspected clinical diagnosis of Morton neuroma (53).
Initial treatment of Morton neuroma is directed at modifying patient footwear. When conservative management fails, other modes of therapy are used, including neurolysis, steroid injection (ultrasound-guided), ultrasound therapy, and surgical release of the transverse metatarsal ligament for decompression (75,76,77). Surgical resection of the neuroma and involved nerve segment is the most successful treatment. However, the failure rate with resection is
P.333

P.334

approximately 10%, likely because of development of traumatic neuroma (78).
Figure 9.4 Morton neuroma in a woman 45 years of age. A: Short axis T1-weighted (TR/TE; 750/15) MR image shows a 6-mm mass (asterisk) in the interspace between the third and fourth metatarsals at the level of the metatarsal head. B,C: Short axis fat-suppressed contrast material–enhanced T1-weighted (TR/TE; 700/15) (B) MR and power Doppler ultrasound (C) images show marked enhancement and increased vascularity of the lesion (asterisk). D: Photograph of the resected specimen shows the entering plantar digital nerve (arrowheads) and the mass (arrows) distally representing perineural fibrosis.
Figure 9.5 Morton neuroma in a woman 39 years of age. A,B: Coronal T1-weighted (TR/TE; 700/15) (A) and T2-weighted (TR/TE; 2500/90) (B) spin-echo MR images show a mass (arrow) between the second and third metatarsal heads. The lesion has decreased signal intensity on all pulse sequences and is more difficult to detect on the long TR image. Intermetatarsal bursal fluid is also seen (arrowhead).
Benign Peripheral Nerve Sheath Tumors
Benign peripheral nerve sheath tumors (BPNSTs) are typically divided into two major benign groups: schwannoma (neurilemoma) and neurofibroma (79). Both the schwannoma and neurofibroma contain cells that are closely related to the normal Schwann cell, from which some would speculate they arise (9). However, multiple clinical and pathologic features usually allow distinction of these neoplasms.
Schwannoma (Neurilemoma)
Schwannoma is most commonly detected in patients between the ages of 20 and 50 years and occurs with equal frequency in men and women (79,80). Schwannomas are slightly less common than neurofibromas and comprise approximately 5% of all benign soft tissue tumors (81). Sites of involvement include the cutaneous nerves of the head and neck and the flexor surfaces of the extremities, particularly the peroneal and ulnar nerves. Schwannomas also show a predilection to affect sensory nerves. The posterior mediastinum and retroperitoneum are also commonly affected with deeply seated lesions. A schwannoma is almost invariably a slowly growing nonaggressive neoplasm that usually presents clinically as a painless mass, often smaller than 5 cm, without neurologic symptoms (9,82,83). Pain may be associated with large lesions, or in patients with schwannomatosis. Schwannoma is solitary in the vast majority of cases. Overall, approximately 5% of cases are plexiform or multiple and are rarely associated with von Recklinghausen disease (82,84). Schwannomas are sporadic in 90%, with 3% occurring in neurofibromatosis type 2 (NF2), 2% in patients with schwannomatosis, and 5% in association with multiple meningiomas with or without NF2 (85). Schwannomas are often mobile in the plane transverse to the involved nerve, but movement is restricted in the longitudinal axis by nerve attachment.
At gross pathology, schwannomas are fusiform-shaped lesions, representing the mass with the entering and exiting nerve. When large nerves are affected, the mass is eccentric in relationship to the involved nerve, with nerve fibers splayed about the neoplasm. Similar to neurofibromas, schwannomas of small nerves may obliterate the nerve of origin. Both the schwannoma and the affected nerve are within a true capsule composed of the epineurium.
The histologic hallmark of schwannoma is identification of Antoni A and Antoni B regions. Lesions are also S-100 protein positive at immunohistochemical analysis (1,82,83). Antoni A areas are more organized and are hypercellular composed of spindle cells arranged in short bundles or interlacing fascicles. Antoni B regions are hypocellular, less organized, and contain more myxoid, loosely arranged tissue, with high water content. These components are intermixed within schwannomas and occur in varying amounts. Schwannomas in which Antoni A areas predominate are often called cellular schwannomas (86). They are more frequently located in the posterior mediastinum and retroperitoneum and constitute 25% of extremity lesions (1). Large schwannomas commonly undergo degenerative changes, including cyst formation, calcification, hemorrhage, and fibrosis and are often referred to as ancient schwannomas (82,83,87). Additional histologic variants of schwannoma include epithelioid and melanotic subtypes. The melanotic schwannoma deserves special mention because of several unique features. Although maintaining the basic features of schwannoma, there is often heavy pigmentation. These lesions commonly involve the spinal nerves (96%), have a mild female predilection (1.4:1 ratio), and 55% occur in association with Carney complex (myxomas, spotty pigmentation, and endocrine overactivity) (79,80,88,89). In addition, unlike conventional schwannoma, the melanotic variant has a significant incidence of malignant behavior with metastases reported in 13% to 26% of cases (79,80,88,89).
Schwannomas are almost universally associated with genetic aberration in chromosome 22. The specific site is frequently referred to as the NF2 gene and this encodes the Merlin protein (also termed schwannomin) (90). The NF2 gene is a tumor suppressor, and loss of its function is likely the initial factor in schwannoma tumorigenesis (90).
Treatment of schwannoma is usually surgical excision. The affected nerve is typically separable from the neoplasm intraoperatively after incision of the epineurium, allowing the native nerve and its function to be preserved (26). Partial resection may be performed in cases that would otherwise require nerve resection for complete removal. Recurrence is unusual, even after incomplete resection, and malignant transformation is exceedingly rare.
P.335

Neurofibroma
Neurofibroma most commonly affects patients 20 to 30 years of age and demonstrates no sex predilection (24,79,80,83,91). These lesions constitute slightly more than 5% of all benign soft tissue tumors (1,81). Three types of neurofibromas are classically described: localized, diffuse, and plexiform (see discussion of neurofibromatosis type 1 [NF1]). The localized variety is the most common, representing approximately 90% of these lesions, and the vast majority are solitary and not associated with NF1 (1,26). Localized neurofibromas often affect superficial cutaneous nerves, although involvement of larger nerves also occurs, causing deep-seated lesions. Localized neurofibromas are slow-growing lesions, usually smaller than 5 cm at presentation and painless (1,26). The diffuse neurofibroma primarily affects children and young adults and most frequently involves the subcutaneous tissues of the head and neck. The majority of diffuse neurofibromas (90%) are isolated lesions not associated with NF1 (1,26). Diffuse neurofibromas demonstrate a plaquelike elevation of the skin with thickening of the entire subcutis.
At gross examination, localized neurofibromas show a fusiform shape, representing the mass with the entering and exiting nerve. Deep-seated lesions of large nerves (most frequently evaluated radiologically) often remain within the epineurium, and, similar to neurilemomas, they have a true capsule (3,64,65). Localized neurofibromas of small nerves commonly extend beyond the epineurium, although they often remain well-circumscribed. In contradistinction, diffuse neurofibroma is a poorly defined lesion in the subcutaneous fat that infiltrates along connective tissue septa. Unlike neurilemomas, neurofibromas are intimately intermixed and inseparable from normal nerve tissue.
At histologic analysis, a localized, solitary neurofibroma is composed of interlacing fascicles of wavy, elongated cells that often contain abundant amounts of collagen. Unlike neurilemoma, neurofibromas do not contain Antoni A and B regions (1,82,83). Myxoid areas and degenerative regions are also not as prominent as they are in neurilemoma. Diffuse neurofibroma contains very uniform, prominent fibrillary collagen. Both localized and diffuse neurofibroma are positive for S-100 protein at immunohistochemical analysis, although generally not as extensively as schwannoma (1). Histologic variants of neurofibroma include cellular, pigmented, and epithelioid types, similar to neurilemoma. Benign neurogenic lesions composed of a mixture of neural elements, skeletal muscle, and rhabdomyomatous elements (neuromuscular hamartoma/choristoma or benign triton tumor) are also described. Genetic aberration in sporadic neurofibromas awaits confirmation, although alterations in the NF1 gene may be involved.
Treatment of localized and diffuse neurofibromas (not associated with NF1) is often surgical resection. However, in contradistinction to neurilemomas, neurofibromas cannot be separated from normal nerve, and complete excision of the neoplasm requires sacrifice of the nerve (92). Although this treatment may be acceptable in cutaneous lesions, deep-seated lesions may only be debulked or managed conservatively with observation. Surgical resection of deep-seated lesions is associated with significant patient morbidity related to nerve damage and neurologic deficits from excision (93). Local recurrence after complete excision is unusual and is more frequently associated with diffuse neurofibroma because of its infiltrative growth. Malignant transformation of these lesions, not associated with NF1, is probably rare, although the exact risk is not known.
Neurofibromatosis
Several reports of neurofibromatosis predate von Recklinghausen’s description in 1882 (79). However, the first association of neural and fibrous components in this disorder is attributed to von Recklinghausen (94). At least eight variations of neurofibromatosis are described, although neurofibromatosis type 1 (NF1) and neurofibromatosis type 2 (NF2) account for 99% of cases (95,96). Musculoskeletal abnormalities predominate in the most
P.336

common form, NF1, as opposed to the central nervous system manifestations (bilateral acoustic neuromas, gliomas, meningiomas) of NF2 (97); therefore, we largely limit our discussion to NF1 (98,99,100,101,102,103).
NF1 is one of the most common genetic diseases, with an estimated frequency of 1 case in every 2,500 to 3,000 births (98,99,100,104,105). Higher frequencies are reported in Arab and Israeli subpopulations. This disease is a mesodermal dysplasia, affecting multiple organ systems, and it is inherited as an autosomal dominant trait. The penetrance of NF1 is 100%, although the expression is variable, with slightly more than 50% of patients only mildly affected (101). However, at least 50% of cases are believed to arise from new mutations, and its mutation rate (1 per 10,000 gametes per generation) is greater than that of many other common genetic disorders (98,99,100,104,105,106). Advanced paternal age (>35 years) is a predisposing factor in an estimated 80% of cases, producing a twofold increase in new mutations, although other factors are also important (9,100). The genetic abnormality is localized to the pericentromeric region of chromosome 17, which is the site of a tumor suppressor gene (9,100). This genetic focus encodes the production of the protein neurofibromin, which likely has some control in cell growth regulation.
Table 9.1 lists the clinical criteria for the diagnosis of NF1 (26), and the classic triad consists of cutaneous lesions, skeletal deformity, and mental deficiency. Café-au-lait spots are identified in approximately 90% of patients, usually within the first several years of life (9,98,99,100,104,105). Although café-au-lait spots are not pathognomonic of NF1 (because they are also seen in tuberous sclerosis and fibrous dysplasia), their size, distribution, and shape in NF1 aid in differentiating this condition from other diagnoses (100,104,105). Café-au-lait spots are caused by increased melanin pigment in the basal epidermal layer and represent another manifestation of the underlying neural crest abnormality. The axilla is a frequent location of café-au-lait spots (9,26). Their extent often parallels disease severity. Another pigmentation abnormality seen in more than 90% of patients with NF1 (and not associated with NF2 or seen in the normal population) is the Lisch nodule, which is an asymptomatic pigmented hamartoma of the iris (98,99,100).
TABLE 9.1 CRITERIA FOR DIAGNOSIS OF NF1a
  • Six or more café-au-lait spots
  • Over 5 mm in greatest diameter in prepubertal patients
  • Over 15 mm in greatest diameter in postpubertal patients
  • Two or more neurofibromas (any type) or one plexiform neurofibroma
  • Axillary or inguinal freckling
  • Optic glioma
  • Two or more Lisch nodules (i.e., iris hamartomas)
  • Distinctive osseous lesions (e.g., sphenoid dysplasia, pseudarthrosis)b
  • First-degree relative with NF1, as diagnosed by preceding criteria
aTwo or more of these criteria are required for diagnosis.
bSee Table 9.2.
From Neurofibromatosis. Conference statement. In: National Institutes of Health Consensus Development Conference. Arch Neurol.
1988; 575–578. Modified, with permission.
TABLE 9.2 OSSEOUS ABNORMALITIES ASSOCIATED WITH NF1
  • Scoliosis (short or long segment)
  • Kyphosis (often predominant deformity)
  • Facial or orbital dysplasia
  • Lambdoid suture defects (left-sided)
  • Pseudarthrosis (particularly of the tibia and congenital)
  • Periosteal abnormalities (reaction, cysts)
  • Multiple nonossifying fibromas or fibroxanthomas
  • Rib deformity (particularly ribbon ribs)
  • Posterior vertebral body scalloping (dural ectasia)
Skeletal abnormalities are common in NF1, occurring in approximately 25% to 40% of cases, again reflecting the multiorgan effects of the mesodermal dysplasia (26,100). Table 9.2 lists these osseous manifestations (Fig. 9.6) (26).
The most frequent skeletal abnormality is kyphoscoliosis. Additional features associated with NF1 include epilepsy, neuropathy, hydrocephalus (aqueductal stenosis), pseudogynecomastia in men, and fibromuscular dysplasia (renal and large cervical vessels) in women. Numerous neoplasms may also be associated with NF1 and include optic glioma, astrocytoma, glioblastoma multiforme, rhabdomyosarcoma, Triton tumor, pheochromocytoma, carcinoid tumor, nephroblastoma, gastrointestinal stromal tumor, and juvenile chronic myeloid leukemia.
The hallmark of NF1 is the neurofibroma. Neurofibromas usually occur initially in childhood or adolescence, subsequent to detection of café-au-lait spots. These lesions can occur in any location of the body, including soft tissues (superficial or deep) and viscera, and, in some reports, more commonly affect males (9). Growth of neurofibromas is usually slow; however, more rapid episodes of growth can be associated with pregnancy, puberty, or malignant transformation.
All three types of neurofibromas (localized, diffuse, and plexiform) can be associated with NF1. Localized neurofibroma is the most common type seen with NF1. However, histologically, both localized and diffuse neurofibromas are not characteristic of NF1 because most of these lesions occur in an isolated pattern which is not associated with this underlying condition. In contradistinction to localized neurofibromas in patients without any underlying disease, those associated with NF1 more frequently involve large deep nerves (particularly the sciatic nerve and brachial plexus). These neurofibromas are large in size and invariably multiple in number. Localized neurofibromas in NF1 also often affect
P.337

the dermis and subcutaneous tissue and are referred to as fibroma molluscum when pedunculated (9,26,100).
Figure 9.6 Neurofibromatosis 1 with ribbon ribs resulting from multiple plexiform neurofibromas in a 20-year-old man. A: Chest radiograph shows multilobulated extrapleural masses (arrowheads) and scalloping of all ribs with a ribbonlike appearance. B: Photograph of the autopsy gross specimen reveals that the rib abnormalities resulted from multiple neurofibromas of intercostal nerves (asterisks) arising from the spinal cord (arrows).
Plexiform neurofibromas are essentially pathognomonic of NF1; development of these lesions usually occurs in early childhood and precedes cutaneous neurofibromas (98,99,100,104,105). Pathologically, a plexiform neurofibroma represents diffuse involvement of a long nerve segment and its branches with tortuous expansion, and its gross appearance is ropelike and is described as a bag of worms. Because of their large size, these lesions commonly extend beyond the epineurium into the surrounding tissue. Plexiform neurofibromas may be associated with massive and disfiguring enlargement of an extremity called elephantiasis neuromatosa, although the segmental variant form (107,108,109) of NF can have a similar appearance without the other stigmata of NF1 (110,111,112,113). This condition may be accompanied by osseous hypertrophy related to chronic hyperemia.
Treatment of patients with NF1 is complicated by the multiplicity of lesions and is often nonsurgical (114). Attempts at surgical resection are usually reserved for markedly symptomatic lesions that substantially compromise function. Because of the large size of many of these lesions, surgical resection is often incomplete, leading to frequent recurrences (92). Malignant transformation to malignant peripheral nerve sheath tumor (MPNST) is the most-feared complication of NF1. The estimated prevalence of malignant transformation varies from 2% to 29%, with an average of approximately 5% (9,26,115). The highest incidence of malignant transformation appears to be in plexiform lesions. Overall, we believe the lower figure of 2% more accurately reflects the true incidence of malignant transformation in NF1, similar to the results of the longitudinal (39-year follow-up) Danish study of 212 patients by Sorenson et al. (116).
Malignant Peripheral Nerve Sheath Tumor (MPNST)
P.338

The World Health Organization (WHO) Committee for the Classification of Soft Tissue Tumors standardized malignant peripheral nerve sheath tumor (MPNST) as the accepted nomenclature for a spindle cell sarcoma arising from nerve, or neurofibroma, or demonstrating nerve tissue differentiation (17,80,117,118). MPNST replaces the former terminology of malignant schwannoma, malignant neurilemoma or neurofibrosarcoma. MPNST accounts for 5% to 10% of all soft tissue sarcomas and usually affects adult patients 20 to 50 years of age (118). Overall, there is a slight female predilection (1,17). It is estimated that these lesions are associated with NF1 in 25% to 70% of cases and that in these patients, MPNSTs occur approximately a decade earlier (28 to 40 years of age) and have a male predilection (80% of patients with MPNSTs and NF1 are male) (17). MPNSTs show a distinct propensity to affect large major or medium-sized nerves, including the sciatic nerve, brachial plexus, and sacral plexus (119). Patients present with pain and neurologic symptoms of motor weakness and sensory deficits more frequently than do patients with benign peripheral nerve sheath tumors (BPNST). In patients with NF1, sudden increase in size of a previously stable neurofibroma should be viewed with great suspicion of malignant transformation, and lead to immediate biopsy. MPNST can also be a secondary neoplasm related to previous radiation therapy. These tumors develop after a long latent period (10 to 20 years) following irradiation and account for 10% to 20% of MPNSTs, with the higher figure associated with paraspinal lesions (17,120).
MPNSTs are fusiform, a shape caused by the entering and exiting nerve, which is clearly evident at gross pathologic examination (17). The tumor frequently spreads along the entering and exiting nerve, with the epineurium and perineurium becoming thickened proximally and distally to the mass, a feature not seen in BPNSTs. The tumor cells are arranged in fascicles, resembling those seen in fibrosarcoma, and areas of hemorrhage and necrosis are frequent. Additional heterotopic regions are seen histologically in 10% to 15% of tumors, and include foci of mature cartilage and bone, rhabdomyosarcoma elements (malignant Triton tumor), angiosarcoma regions, and glandular or epithelioid components (17). MPNST with rhabdomyosarcoma, angiosarcoma, or glandular elements tend to occur in association with NF1. Epithelioid MPNST is less commonly associated with NF1 and accounts for 5% of these malignancies. Epithelioid MPNST arises in relation to a large deep nerve in 50% to 80% of cases, with the remainder in the superficial soft tissues. The majority of MPNSTs are considered to be high-grade sarcomas.
The cytogenetics of MPNST now suggest a multistep mechanism of tumorigenesis. The initial step in the process in patients with NF1 is inactivation of one or both alleles of the NF1 gene allowing neurofibroma formation (121). Progression to MPNST is likely associated with several additional chromosomal aberrations, with particular involvement of the tumor suppression gene on chromosome 17p (122). The p53 focus is suggested as the specific culprit in allowing increased proliferation and angioinvasive capacity (122). Scientific implication of this pathway is supported by the fact that patients with MPNST show a high percentage of abnormalities at this locus, which are not seen in patients with BPNST.
Treatment of MPNST is complete surgical excision with wide resection margins. Adjuvant chemotherapy and radiation are also often employed. Radiation therapy reduces the incidence of local recurrence; however, despite this aggressive treatment, local recurrence and distant metastases are common, seen in 40% to 65% and 40% to 68% of patients, respectively (17,119). Wanebo et al. (120) reported a 5-year survival rate of only 44%. Two additional studies show 5- and 10-year survival rates ranging from 34% to 52% and 23% to 34%, respectively (120). Lesions in sites that are difficult to achieve adequate surgical margins (paraspinal, thoracic, retroperitoneal) are associated with even lower survival rates (15%). Worsened prognosis is associated with older patient age, larger tumor size (>5 cm), more central location of the tumor, high mitotic rate (>20/10 high power field [HPF]), and positive margins after resection. Although some researchers believe patients with NF1 have a significantly poorer prognosis, this contention is now somewhat controversial. Several newer reports show that all patients with MPNSTs have a similar prognosis regardless of the presence of underlying NF1 (120,123). Metastases most frequently affect the lung, bone, pleura, and retroperitoneum. Regional lymph nodes are involved in 9% of cases (17,119,120,123).
Imaging of Schwannoma, Neurofibroma, Neurofibromatosis Type 1, and Malignant Peripheral Nerve Sheath Tumor
The most common abnormality of peripheral nerve sheath tumors (PNSTs) (including BPNST, NF1, and MPNST) at radiography is a nonspecific soft tissue mass. Radiographs are also frequently normal. In rare cases, a fusiform soft tissue mass with surrounding fat may be seen. Occasionally, soft tissue and osseous overgrowth associated with elephantiasis neuromatosa and other skeletal manifestations of NF1 or the segmental variant may be recognized on radiographs (100,124). A primary osseous location for PNSTs is exceedingly rare, and bone involvement by either extrinsic erosion or invasion is unusual. Calcification (osteoid, chondroid, or amorphous) is uncommon and mild in extent when present (Fig. 9.7) (100,124). Both
P.339

bone involvement and mineralization are more common in larger lesions and MPNST. This reflects the heterogeneity of these lesions seen pathologically.
Angiography of deep PNSTs demonstrates displacement of major vascular structures related to the site of origin of the lesion within the neurovascular bundle. The degree of increased vascularity is variable, but typically more prominent in MPNST (125,126,127,128,129). A characteristic at angiography that suggests a neurogenic neoplasm is the identification of corkscrew-type vessels at the upper or lower poles of the tumor. This appearance represents hypertrophy of nutrient nerve vasculature (125,127) (Fig. 9.8).
Bone scintigraphy findings of PNSTs are nonspecific and reflect the vascularity, bone involvement, or mineralization associated with the tumor. Typically, only mild uptake of radionuclide is seen on all phases of imaging, unless calcification or bone involvement is extensive. However, several reports describe gallium-67 citrate imaging as very helpful for differentiating BPNST from MPNST (126,130). Although only small numbers of patients were involved, significant uptake of gallium-67 citrate was seen in MPNST compared to minimal or no accumulation in BPNST (Figs. 9.9,9.10 and 9.11) (126,130,131,132).
Early results with FDG-PET imaging in evaluation of neurogenic neoplasm have now reported (133). Several reports by Ahmed et al., Beaulieu et al., and Shah et al. identify marked radionuclide uptake in a significant number of patients with schwannomas (5% to 50%) with standard uptake values (SUVs) numbers greater than 2 (134,135,136). This limits use of FDG-PET to distinguish schwannoma from MPNST. However, Ferner et al. and Solomon et al. reported success in distinguishing plexiform neurofibromas that have undergone malignant degeneration to MPNST from benign lesions in patients with NF1 using FDG-PET (137,138).
As with other soft tissue neoplasms, these lesions are more easily characterized with cross-sectional imaging techniques (ultrasound, CT, MR) (139,140,141,142,143,144,145). In our opinion and experience, the most important imaging feature that should always suggest the diagnosis of neurogenic neoplasm is recognition of a fusiform mass (Figs. 9.12
P.340

and 9.13) (146). This appearance is a direct reflection of the underlying gross morphology of schwannomas, localized neurofibroma, and MPNST, representing the tubular entering and exiting nerve in a typical nerve distribution. This relationship is usually easy to detect in lesions affecting large, deep nerves that are frequently imaged because the clinical presentation is that of a nonspecific soft tissue mass (Figs. 9.12 and 9.13). In contradistinction, in superficial PNSTs, it is often difficult or impossible to identify this appearance, and imaging findings are nonspecific. However, these superficial lesions, as with other cutaneous or subcutaneous lesions, often are not imaged because of the so-called ease of clinical assessment. In our experience, MR imaging and ultrasonography are superior to CT for demonstrating the virtually pathognomonic fusiform appearance of deep-seated neurilemoma, localized neurofibroma, and MPNST because of their large field of view and multiplanar capabilities (145,147,148). Cerofolini et al. (149) reported this finding in 94% of their 17 cases. PNSTs of the paraspinal region often reveal a dumbbell shape with extension into an enlarged neural foramen (150). This
P.341

neural foraminal component is analogous to the entering nerve seen in peripheral lesions (Fig. 9.14) (151,152,153,154,155,156,157). Extrinsic erosion of the posterior vertebral body and, rarely, invasion simulating a more aggressive process may also be seen (153,158). Spinal PNSTs must be differentiated from meningoceles (70% to 80% of the latter lesions occur in patients with NF1 and are often multiple) because of differences in treatment (100). Meningoceles characteristically are cystic, fill with contrast material after myelography (with or without CT), and have a posterior mediastinal location without calcification. Spinal neurofibromas in patients with NF1 are often bilateral, and a prominent disparity in size should be viewed with suspicion that the larger lesion harbors MPNST (Fig. 9.15) (151,152,153,154,155,156).
Figure 9.7 Mineralization in nerve sheath tumors in two different patients. A: Foot radiograph shows faint calcification between the first and second metatarsals in a patient with a malignant peripheral nerve sheath tumor. B: Foot radiograph shows densely mineralized neurofibroma in the fifth toe. Histology (not shown) revealed areas of chondroid and osteoid metaplasia.
Figure 9.8 Schwannoma of the thigh in a man 29 years of age. A: Radiograph shows a noncalcified fusiform soft tissue mass. B: Early arterial phase film from conventional arteriogram reveals displaced femoral artery with multiple feeding vessels. Nutrient arteries of the sciatic nerve are tortuous at both the proximal (arrow) and distal ends of the tumor. C: The late-phase angiogram image shows an intense tumor stain and draining veins.
Figure 9.9 Schwannoma of the thigh in a man 61 years of age. Bone scintigraphy demonstrates mild accumulation of radionuclide (arrowheads).
Theoretically, differentiation of deep-seated localized neurofibroma from neurilemoma should be possible because of their differences in location relative to the affected nerve. In neurilemoma, the mass is eccentric and separable from normal nerve, but in neurofibroma the two structures are intimately related, intermixed and indistinguishable. Indeed, Cerofolini et al. (149) believed they could discern this relationship in 65% of their 17 cases. In our experience, however, this distinction can be difficult because both lesions are often in deep locations within the epineurium and have similar intrinsic imaging characteristics relative to affected nerve. This similarity often precludes radiologic discrimination between these lesions in many patients, and this distinction is best accomplished in neoplasms affecting large nerves such as the sciatic nerve. We believe significant eccentricity in any imaging plane of the nerve relative to the mass strongly suggests neurilemoma versus localized neurofibroma (Figs. 9.12 and 9.13) (145).
Figure 9.10 Malignant peripheral nerve sheath tumor in the left thigh of a man 22 years of age. Left thigh demonstrates intense uptake of gallium-67 5 days after injection.
Recently, Jee et al. reported the following characteristics as statistically significant in distinguishing extra-axial localized neurofibroma (n = 12) from neurilemoma (n = 40): target sign on T2-weighted MR images (58% of neurofibromas, 65% neurilemomas); central enhancement (75% neurofibromas, 8% neurilemomas); and target sign and central enhancement (63% neurofibromas, 3% neurilemomas) (159). Features favoring extra-axial
P.342

neurilemoma versus localized neurofibroma were the fascicular appearance on T2-weighted MR images (25% neurofibromas, 63% neurilemomas); a thin hyperintense rim on T2-weighted images (8% neurofibromas, 58% neurilemomas); a combination of the above mentioned features (8% neurofibromas, 48% neurilemomas); and diffuse contrast enhancement (13% neurofibromas, 67% neurilemomas) (159). No statistically significant difference was seen between the lesions for the following features: central entering/exiting nerve (42% neurofibromas, 23% neurilemomas); peripherally entering/exiting nerve (58% neurofibromas, 77% neurilemomas); a cystic area (38% neurofibromas, 64% neurilemomas); low-signal intensity margin (100% for both lesions); peripheral contrast enhancement (13% neurofibromas, 26% neurilemomas); or target sign on contrast-enhanced MR images (11% neurofibromas, 31% neurilemomas) (159).
Figure 9.11 Schwannoma of the sciatic nerve in a woman 57 years of age shows relative photopenia with no significant radionuclide uptake of gallium by the tumor (asterisk) 2 days after injection.
Plexiform neurofibromas invariably show a pathognomic imaging appearance identical to that of their gross pathologic features of diffuse nerve thickening (Figs. 9.16 and 9.17) (160,161,162,163,164). There is often nodularity and involvement of nerve branches that create the appearance of a serpentine “bag of worms.” Patients with multiple schwannomas may reveal multiple masses (beadlike appearance) with an intrinsic appearance identical to that of other neurilemomas, usually along a single nerve distribution best seen on long axis MR images (Fig. 9.18). Unlike plexiform neurofibroma, the intervening nerve is not typically thickened, and the nerve branches are not usually affected (Fig. 9.18) (165,166). Diffuse neurofibromas show a reticulated linear branching pattern within the subcutaneous tissue replacing the fat and creating a honeycomb appearance (Fig. 9.19) (26,167).
TABLE 9.3 IMAGING SIGNS OF NEUROGENIC NEOPLASMS
Sign Modality Depicting the Sign
Fusiform MR, ultrasonography (less well-seen with CT)
Entering and exiting nerve MR, ultrasonography (less well-seen with CT)
Low attenuation Unenhanced CT
Target sign T2-weighted MR (less well-seen with CT, ultrasonography)
Fascicular sign T2- and proton density–weighted MR
Split-fat sign T1-weighted MR (less well-seen with CT, ultrasonography)
Associated muscle atrophy T1-weighted MR
Table 9.3 lists the intrinsic imaging characteristics of neurogenic neoplasms. On unenhanced CT scans, PNSTs frequently have low attenuation (often as low as 5 to 25 HU) (Figs. 9.15 and 9.16) (168,169). This appearance is attributed to several factors, including high lipid content of myelin from Schwann cells, presence or entrapment of fat, endoneurial myxoid tissue with high water content (Antoni B areas in neurilemomas or myxoid areas in neurofibromas), and cystic areas (hemorrhage, necrosis or degenerated areas in ancient schwannomas) (170,171,172,173,174,175). Heterogeneity and higher attenuation may be seen in neurogenic neoplasms and is a more common feature of MPNST (128,129).
On MR images, the signal intensity of neurogenic neoplasms is relatively nonspecific and is similar to or lower than that of muscle on T1-weighted images and higher than that of fat on T2-weighted MR images. Ultrasonography typically reveals a well-defined hypoechoic mass with variable posterior acoustic enhancement (176,177). Color Doppler studies are useful to identify lesions that may simulate a ganglion on other imaging modalities with the identification of intrinsic blood flow (176). Diffuse neurofibromas often show predominant low signal intensity on T2-weighted MR images, a finding we believe is related to the high collagen content of these lesions. Heterogeneity in PNSTs is variable, although it is more prominent in MPNST (including, in uncommon cases, fluid levels and hemorrhage). Mann et al. (178) attempted to quantitate this heterogeneity in distinguishing benign PNST from MPNST by using fuzzy cluster analysis and found separation to be difficult by these characteristics alone.
Figure 9.12 Schwannoma of the peroneal nerve in a woman 49 years of age. A,B: Coronal T1-weighted (TR/TE; 500/20) (A) and axial proton density (TR/TE; 2000/30) (B) spin-echo MR images reveal a fusiform intermuscular soft tissue mass with entering and exiting nerve (asterisks), surrounding fat is well seen in A, while fascicular sign is well demonstrated in B. The axial MR image shows the peroneal nerve in the periphery of the mass (arrow); the coronal image reveals mild eccentricity of the entering/exiting nerve in relationship to the mass. C: Intraoperative photograph shows the same relationship with the eccentric nerve (asterisk) easily separable from the tumor (T), although this was only apparent after incising of the epineurium.
P.343

The target sign is described as being nearly pathognomonic of neurofibroma on T2-weighted MR images and consists of low-to-intermediate signal intensity centrally, with a ring of high signal peripherally (Figs. 9.16 and 9.20) (151,152,155,156,179). These central areas enhance with contrast and, in the study by Lin et al., strongly suggested a lesion is a PNST as opposed to MPNST (141). This MR imaging finding corresponds pathologically to fibrous tissue (with high collagen content) centrally and more myxoid tissue peripherally. It is most frequent, in our experience, in plexiform neurofibroma. Suh et al. (81,156) described this finding in 70%
P.344

of 10 cases of neurofibromas. In our experience, however, this prevalence is an overestimate of the frequency of this finding, which is supported by the study of Bhargava et al. who reported this sign in 52% of 23 neurofibromas (180). We strongly agree that the MR target sign should always suggest a neurogenic neoplasm, although it can be seen less frequently in neurilemoma (Fig. 9.20) and MPNST, as well as neurofibroma (both deep-seated and superficially located lesions). The target sign was also reported in 8% of 12 MPNSTs by Bhargava et al. (180). The central areas of fibrous tissue enhance more prominently than the peripheral myxoid tissue following intravenous contrast, as reported by Ogose et al., which could also be considered a target sign (181). CT scanning can also demonstrate the target sign (peripherally low attenuation with central higher attenuation), although not as well as MR imaging, reflecting the superior contrast resolution of the latter modality (173,174,175). The target sign can also be identified on ultrasonography (Figs. 9.13 and 9.17) with higher echogenicity centrally, similar to CT scanning, but not as well as with MR imaging in our opinion (182).
Figure 9.13 Schwannoma of the median nerve in a man 52 years of age. A,B: Short axis (A) and longitudinal sonograms (B) show a fusiform mass with an exiting nerve (arrow in B). The nerve is eccentric in relationship to the mass, typical of a schwannoma. Hyperechoic rim (arrowheads) represents the split-fat sign of an intermuscular mass, and target sign is seen with increased echogenicity centrally (black asterisk) and decreased echogenicity peripherally (white asterisk).
In our experience, another intrinsic MR imaging characteristic that should suggest neurogenic neoplasm is the fascicular sign (Figs. 9.12 and 9.21) (26). This feature may be seen in both superficial and deep-seated lesions. The fascicular sign manifests as multiple, small, ringlike structures (with peripherally higher signal intensity) on either T2-weighted or proton density–weighted MR images. We believe this sign corresponds to the fascicular bundles seen pathologically in neurogenic neoplasms, particularly in more differentiated PNSTs. This appearance recapitulates that seen in normal nerves, as described both on MR imaging and ultrasonography (4,5). The fascicular sign, as expected, is more frequent in BPNST than in MPSNT. The latter are more anaplastic neoplasms, in which this sign may be present only in small foci of the lesions.
The margins of BPNSTs are usually well-defined at ultrasonography, CT, and MR imaging (Figs. 9.12,9.13 and 9.14, 9.18, 9.20, and 9.21) (89,183,184,185,186,187,188,189,190,191,192). In fact, a capsule representing the epineurium (higher attenuation rim by CT, echogenic rim on ultrasonography, low signal intensity rim on all MR images) may be apparent. Unlike Beggs (32), we do not find this particularly helpful in distinguishing deep-seated localized neurofibroma from neurilemoma. The defined margin of these lesions on imaging reflects the underlying pathologic feature of both of these lesions frequently being contained within the epineurium (particularly schwannomas). Neurofibromas, although more common to extend beyond the epineurium, remain well-circumscribed in most cases, also causing a defined margin on imaging. However, indistinct margins are far more frequent in MPNSTs as a result of more infiltrative growth (Figs. 9.22 and 9.23). We believe any ill-defined margin, particularly on MR imaging (T2-weighted, short-tau inversion recovery [STIR], or postcontrast), of a deep-seated nonplexiform PNST strongly suggests MPNST. Plexiform neurofibromas may also show ill-defined margins, and diffuse neurofibromas always appear indistinct and infiltrative because of their subcutaneous spread along connective tissue septae (26,160).
Figure 9.14 Multiple spinal neurofibromas in a woman 38 years of age with neurofibromatosis 1. A,B: Coronal T1-weighted (TR/TE; 500/16) spin-echo MR images of the lumbar spine preceding (A) and following (B) gadolinium administration show multiple paraspinal masses. Nonenhancing areas represent central hemorrhage or necrosis. C: Axial T1-weighted (TR/TE; 717/16) spin-echo MR image following gadolinium shows the dumbbell shape to better advantage, representing the entering nerve extending into the neural foramina. Note small contralateral lesion.
Figure 9.15 Multiple spinal neurofibromas in a man 40 years of age with neurofibromatosis 1. Axial CT shows low attenuation paraspinal lesions (asterisks) with extension into the neural foramina (arrows). The much larger lesion on the left proved to be a malignant peripheral nerve sheath tumor.
Figure 9.16 Plexiform neurofibroma of the forearm and lower leg in a boy 8 years of age with neurofibromatosis 1. A,B: Coronal T1-weighted (TR/TE; 500/20) (A) and axial turbo spin-echo T2-weighted (TR/TE; 6500/115) (B) MR images show serpentine plexiform neurofibroma with a “bag of worms” appearance. The target sign with high signal intensity peripherally and low signal centrally (arrows) is seen in several lesions on the axial T2-weighted image. C: CT of lower leg reveals multiple low attenuation neurofibromas (arrows). D: Intraoperative photograph also shows serpentine “bag of worms” appearance.
Figure 9.17 Plexiform neurofibroma in a woman 33 years of age with neurofibromatosis 1. A,B: Long (A) and short (B) axis sonograms show plexiform morphology with diffusely thickened nerve (arrows), with higher echogenicity centrally (smaller asterisks), and lower echogenicity peripherally (larger asterisks). C,D: Coronal T1-weighted (TR/TE; 500/15) (C) and fat-suppressed T2-weighted (TR/TE; 3000/50) (D) MR images reveal the diffusely thickened and nodular sciatic nerve (black asterisk) and involvement to the nerve branches (arrowheads).
Figure 9.18 Schwannomatosis in a 35-year-old man without neurofibromatosis. Coronal proton density MR image (TR/TE; 2500/30) shows two schwannomas (arrows) with normal thickness intervening peroneal nerve (arrowhead).
P.345

P.346

P.347

P.348

A rim of fat (split-fat sign) is often present about deep-seated neurogenic neoplasms and has been described previously on CT scans, although is much easier to appreciate on T1-weighted MR imaging (Figs. 9.12 and 9.17) (170). The split-fat sign can also be seen on ultrasonography as a hyperechoic rim about the mass, as well as the entering and exiting nerve (Figs. 9.13 and 9.21). Because the neurovascular bundle is surrounded by fat, masses arising in this site maintain a rim of fat about them as they slowly enlarge. Although not a specific sign for PNST, this finding suggests tumor origin in the intermuscular space about the neurovascular bundle of which neurogenic neoplasms are the most frequent cause. Intramuscular masses do not demonstrate this feature unless they involve the entire muscle compartment extending to the fat in the intermuscular septae. The split-fat sign is more common in BPNST and lesions of large nerves. MPNST less frequently demonstrates a complete fat rim, reflecting its more infiltrative growth pattern.
Muscle atrophy with striated increased fat content and/or decreased size is associated with PNST and reported by Stull et al. (191) to occur in 23% of cases. Muscle atrophy is not commonly seen with other soft tissue masses. This finding can be quite subtle in muscle supplied by the affected nerve, may require comparison to the normal side, and is best seen on T1-weighted MR imaging (Fig. 9.24).
TABLE 9.4 DIFFERENTIATING BENIGN AND MALIGNANT PERIPHERAL NEUROECTODERMAL TUMORS
  BPNSTa MPNSTb
Fusiform shape Common Common
Entering/exiting nerve Common Common
Target sign Common Uncommon
Fasicular sign Common, diffuse Uncommon, focal
Split-fat sign Common, complete Common (may be incomplete)
Size <5 cm >5 cm
Margins Defined Ill-defined
Central necrosis Uncommon Common
Gallium uptake No Yes
FDG-PET uptake Variable Yes
Rapid growth No Yes
Vascularity Variable Prominent
aBenign peripheral nerve sheath tumor
bMalignant peripheral nerve sheath tumor
Contrast enhancement on CT or MR imaging, similar to angiography appearances, is variable in both BPNST and MPNST. Generally, more contrast enhancement is apparent in MPNST (26,193). The pattern of enhancement is also variable, commonly either heterogeneous and diffuse or peripheral (Fig. 9.21) (194). However, as described previously, lesions demonstrating the target sign typically enhance more prominently centrally (Fig. 9.25) (195). In addition, Lee and Boles reported a case of schwannoma with no enhancement (196,197). Irregular nodular peripheral enhancement with central necrosis is typical of MPNST. However, central necrosis can also be seen in ancient schwannomas (Figs. 9.14 and 9.26) (198,199,200). Ancient schwannoma is one of few benign lesions that can demonstrate this feature, and it represents an important exception to the general rule that central necrosis implies malignancy (Figs. 9.14 and 9.26) (201,202). Contrast-enhanced MR images (and T2-weighted sequences) may demonstrate growth of neurogenic neoplasm into the entering and exiting nerve and surrounding soft tissues. This imaging appearance is an ominous sign of MPNST as opposed to BPNST in nonplexiform lesions, and reflects the growth seen pathologically (Figs. 9.22 and 9.23) (132).
Differentiation of BPNST from MPNST is also often very difficult. Imaging features suggestive of malignancy include large size (>5 cm), prominent vascularity or enhancement, infiltrative margins, marked heterogeneity with central necrosis, rapid growth, and increased uptake of gallium-67 citrate (Figs. 9.10, 9.13, 9.20,9.21,9.22 and 9.23) (Table 9.4). Recognition of these imaging features is important for prospective diagnosis and to help guide therapy in the clinical management of these patients.
Figure 9.19 Diffuse neurofibroma of the buttock in a woman 31 years of age without neurofibromatosis. A,B: Coronal T1-weighted spin-echo MR images (A) before and corresponding fat-suppressed postcontrast T1-weighted image (B) show an infiltrative mass in the subcutaneous fat, with linear branching components. There is significant enhancement after contrast injection. C: Coronal proton density spin-echo MR image reveals heterogeneous high signal intensity.
P.349

Neurothekeoma and Nerve Sheath Myxoma
Neurothekeoma and nerve sheath myxoma were previously thought to represent variations of the same lesion, but they are now considered to be distinct lesions based on histologic differences. Nerve sheath myxomas reveal schwannian features, whereas neurothekeoma shows fibroblastic or myofibroblastic derivation. However, both lesions typically involve the superficial tissues of the head, face, neck, or shoulders and show a female predilection (2:1 ratio). Deep-seated lesions are very rare (Fig. 9.27). Nerve sheath myxoma affects adults in the third to fifth decade, whereas neurothekeoma involves primarily children and young adults.
Similar to other subcutaneous lesions, these tumors are rarely subject to radiologic evaluation. We would expect that both lesions would demonstrate nonspecific features of a subcutaneous mass, but that imaging would reflect the high water content of nerve sheath myxoma (Fig. 9.27). These benign lesions have no malignant potential, and complete excision is curative.
Perineurioma
P.350

Perineurial cells form the outer lining of peripheral nerve fascicles, analogous to the meningeal cells in the pia-arachnoid membrane. Perineurioma, described in 1978 by Lazarus and Trombetta, represents a rare soft tissue tumor composed of cells resembling normal perineurium without Schwann cell components (203). These lesions are rare and represent less than 1% of soft tissue neoplasms. There are two distinct types of these lesions: the intraneural perineurioma and the extraneural, or soft tissue, perineurioma. More than 30 cases of intraneural perineurioma have been reported to date (204). These lesions were formerly diagnosed as hypertrophic neuropathy and considered to be reactive lesions. However, further pathologic evaluation with clonal abnormalities (monosomy) related to chromosome 22 now provides proof of their neoplastic nature. Adolescents or young adults are usually affected with an equal sex distribution. The most common symptom is muscle weakness, and atrophy may be apparent. Nerves of the upper extremity are most commonly involved, followed by those of the lower extremity. Only one nerve is typically affected, with a single case of adjacent spinal nerve involvement reported (204). At gross pathologic examination there is symmetric tubular enlargement of the affected nerve (from 2 cm to 10 cm), without involvement of its branches. Histologically, ropelike bundles of perineurial cells surround indistinct nerve fibers.
Figure 9.20 Neurilemoma in a man 35 years of age with a palpable soft tissue mass in the calf. A: Coronal T1-weighted (TR/TE; 500/20) MR image shows an elongated, low-signal-intensity mass (arrowheads). B: On the axial T2-weighted (TR/TE; 2000/90) MR image, the mass has peripheral high signal intensity (white arrow) with low signal intensity centrally (black arrow), representing the target sign. No entering or exiting nerve is seen because the affected nerve is a small gastrocnemius intramuscular branch.
Soft tissue perineuriomas are most common in the subcutaneous tissues of the hand, affect adults of all ages, and have a female predominance (4:1) except for the sclerosing subtype (male predilection). These lesions are solitary, well-circumscribed, but nonencapsulated masses of perineurial cells, but unlike their intraneural counterpart, they do not surround a nerve. Soft tissue perineuriomas demonstrate the same cytogenetic abnormality as intraneural lesions.
Radiologic evaluation of these lesions is only rarely reported. In our experience, intraneural perineuriomas have a distinctive, pathognomic appearance, particularly on MR imaging. The affected nerve is diffusely thickened for a long extent with high signal intensity on T2-weighting and enhancement following intravenous contrast (Fig. 9.28). Similar morphologic features would be expected on CT scanning or sonography and reflect the underlying pathology. In contradistinction to plexiform neurofibroma or an infectious neuritis (such as that seen in leprosy), the nerve branches are not involved (no “bag of worms” appearance), and no surrounding inflammation is identified. The imaging appearance of soft tissue perineurioma is not described, to the best of our knowledge, but we would expect nonspecific characteristics with detection of a soft tissue mass.
These lesions are benign, with only rare and controversial description of a malignant variant. Surgical excision of soft tissue perineurioma is curative and there have been no reports of recurrence. Intraneural perineurioma demonstrates neither risk of recurrence nor metastases on long-term follow-up, and thus biopsy alone is considered sufficient for diagnosis. Biopsy should be directed to fascicles that are nonfunctional by direct nerve stimulation. We would suggest that imaging alone is diagnostic and may supplant the need for biopsy. Surgical resection of intraneural perineurioma is to be avoided to retain nerve function even if only partial. Resection, even though curative, of nonfunctional, localized,
P.351

affected nerve with graft placement and reconstruction, may not be associated with recovery of function.
Figure 9.21 Schwannoma of the forearm (ulnar nerve) in a 15-year-old girl. A–C: Axial T1-weighted (TR/TE; 616/25) (A), proton density (TR/TE; 2000/25) (B) and T2-weighted (TR/TE; 2000/90) (C) spin-echo MR images show a well-defined mass with fascicular sign (best seen on long TR images). The relationship of the ulnar nerve to the mass (schwannoma versus neurofibroma) is difficult to distinguish. D,E: Sagittal T1-weighted (TR/TE; 578/16) (D) spin-echo MR image preceding intravenous gadolinium, and corresponding T1-weighted (TR/TE; 600/15) (E) image following contrast reveal a fusiform mass. Nerve is seen proximally (arrow in D). Note surrounding rim of fat (split-fat sign) and mild peripheral enhancement.
Granular Cell Tumor
Initially the granular cell tumor was believed to be of muscle origin, and the term granular myoblastoma was used. Evidence now strongly suggests that this lesion is of nerve origin. Other former terms for this neoplasm included granular cell neurogenic tumor, granular cell neuroma, granular cell neurofibroma, and granular cell schwannoma (9). Granular cell tumors are relatively common lesions. They occur most frequently in the fourth to sixth decade, are twice as common in women, and have a predilection for blacks. Lesions are often in the skin or subcutaneous tissue, although a deep intramuscular location (205) can also occur and usually develops in close association to small-to-medium-sized nerves. The most common location is the head and neck (particularly the tongue) (Fig. 9.29), followed by the chest wall, breast, and arm (9,206,207,208,209). These lesions most frequently present as small painless masses, and approximately 10% to 15% occur at multiple
P.352

locations, with as many as 50 sites reported in some cases. Multiple lesions may be metachronous or synchronous in occurrence.
Figure 9.22 Malignant peripheral nerve sheath tumor of the radial nerve in a man 24 years of age with neurofibromatosis 1 and clinical rapid enlargement of a forearm mass. A: Spliced long axis sonogram shows the mass (asterisk) with entering and exiting nerve (arrows) and split-fat sign (arrowheads). B–D: Multiple MR images including axial T1-weighting (TR/TE; 500/20) (B), sagittal enhanced, fat-suppressed T1-weighted (TR/TE; 500/20) (C) and sagittal T2-weighted (TR/TE; 3000/115) (D), reveal a large (>5 cm) necrotic forearm mass (asterisk) with irregular margins and extension along the entering and exiting nerves (arrowheads).
Pathologically, the vast majority of lesions are benign, with only 1% to 2% being malignant and showing metastatic potential (9,210). Lesions are often closely associated with or replace adjacent peripheral nerves (9,206,207). Benign granular cell tumor is usually smaller than 3 cm as opposed to the malignant variety, which is typically larger than 4 cm. Malignant granular cell tumor should demonstrate three or more of the following histologic features: necrosis, spindling, vesicular nuclei with prominent nucleoli, increased mitotic activity (>2 mitoses/HPF), high nucleocytoplasmic ratio, or pleomorphism. Histologically, the lesions are nonencapsulated and are composed of cells with prominent granular eosinophilic cytoplasm.
Imaging of these lesions has only been recently reported, again related to the fact that superficial lesions are often not evaluated radiologically (3,208,209,211,212,213). The largest series are by Blacksin et al. and Elkousy et al. of 5 and 10 patients, respectively (214,215). These neoplasms are seen as subcutaneous nodular masses. Deep-seated lesions may show an intimate relationship to a nerve
P.353

P.354

(212). Low signal intensity on T2-weighted MR images, with a peripheral rim of high signal intensity that may reveal infiltrative margins has also been described as a suggestive feature of this diagnosis (Fig. 9.30) (216). Lesions larger than 4 cm with invasion of adjacent structures, including bone, should raise suspicion of a malignant granular cell tumor.
Figure 9.23 Malignant peripheral nerve sheath tumor (MPNST) in the thigh of a man 42 years of age with neurofibromatosis 1 (NF1). A–C: Coronal T1-weighted (TR/TE; 634/18) (A) spin-echo, coronal short-tau inversion recovery (STIR) (TR/TE/TI; 3000/60/160) (B) and axial T2-weighted (TR/TE; 2500/100) (C) MR images show a large fusiform mass in the mid-right thigh. Thickened sciatic nerve (asterisks) is seen entering and exiting the mass, and mild irregular margins are demonstrated on the axial image. D: Coronal T1-weighted (TR/TE; 634/18) spin-echo MR image following intravenous gadolinium shows extensive peripheral enhancement and tumor extending along thickened entering and exiting sciatic nerve. Nonenhancing central necrosis (asterisk) has low signal on T1-weighting and high signal on T2-weighting. Numerous superficial neurofibromas are also seen (several are marked with arrowheads), suggesting NF1. Large tumor size (>5 cm), irregular margins, and extension along the nerve all suggest MPNST as opposed to benign neurofibroma. E: Photograph of gross specimen shows multinodular thickening of a plexiform neurofibroma of both sciatic nerve and its branches (white arrows) and the necrotic MPNST (asterisk).
Figure 9.24 Malignant peripheral nerve sheath tumor in a girl 12 years of age without neurofibromatosis. A: Noncontrast axial CT shows a mass (arrow) related to the sciatic nerve, with marked atrophy of gluteal musculature (asterisk). B,C: Axial T1-weighted (TR/TE; 600/24) (B) and T2-weighted (TR/TE; 2000/90) (C) spin-echo MR images reveal a relatively well-defined mass (arrow). The fascicular pattern is best seen on axial T2-weighted image. Left gluteal muscle atrophy is again noted. D: Coronal turbo T2-weighted (TR/TE; 4000/108) spin-echo MR image shows the sciatic nerve entering the mass (asterisk).
Surgical resection is usually curative for benign granular cell tumors. Malignant granular cell tumors demonstrate significant metastatic potential with a 40% risk of mortality. Metastases in malignant granular cell tumors (50% of patients) usually require several years to occur following excision of the original lesion, and usually follow local recurrence. Metastases most frequently affect the lymph nodes, lung, liver, and bone. Reports of small numbers of patients suggest that adjuvant radiation therapy and chemotherapy are not effective in treatment of malignant lesions.
Melanotic Neuroectodermal Tumor of Infancy
The original description of the melanotic neuroectodermal tumor of infancy was by Krompecher in 1918, and approximately 150 to 200 cases have been reported to date (217). The vast majority of evidence suggests a neural crest origin. Additional terms for this lesion include congenital
P.355

melanocarcinoma, melanotic adamantinoma, retinal anlage tumor, pigmented epulis of infancy
, and melanotic progonoma (217,218,219,220,221). This tumor usually develops in patients younger than 6 months of age, but is only occasionally identified at birth. Rapid enlargement causing facial disfigurement and a protuberant blue-black mass (related to melanin) arising from the mandible, maxilla, or calvarium, is usually seen at presentation. Other rare reported locations include the epididymis, mediastinum, uterus, and shoulder. Multicentric melanotic neuroectodermal tumors of infancy are also rarely reported.
Figure 9.25 Neurofibroma of the peroneal nerve in a woman 77 years of age. A: Axial T2-weighted (TR/TE; 2200/80) spin-echo MR image shows a characteristic target appearance, with a central zone of low signal intensity and a peripheral zone of higher signal intensity. B,C: Axial T1-weighted (TR/TE; 577/16) spin-echo MR images before (B) and after (C) gadolinium administration show enhancement of the central area that corresponds to a solid central, tightly packed, cellular portion with fibrous tissue and xanthomatous areas. The peripheral nonenhancing region corresponds to loosely arranged myxoid stroma.
A neural crest origin of this tumor is supported pathologically by the ultrastructure identification of neurofilaments and neurosecretory granules. Variable amounts of melanin pigment are also present. Elevated levels of vanillylmandelic acid (VMA) are also rarely associated with this tumor.
Treatment primarily involves surgical resection; however, recurrence is seen in nearly 50% of patients (9,217,218,219,220,221). In addition, although melanotic neuroectodermal tumor of infancy is generally considered to be a benign lesion, metastases are reported in 2% to 10% of cases (9). Sporadic reports of the use of adjuvant chemotherapy are also available.
Imaging of this tumor is not extensively reported, reflecting its rare occurrence, with the largest series consisting of five patients reported by Mirich et al. in 1991
P.356

(217,218,219,220,221). The lesion typically involves both bone and soft tissue, with lytic destruction and expansile remodeling of the mandible, maxilla, or skull. Reactive hyperostosis and osteogenesis may also be present. CT generally reveals relatively homogeneous replacement of bone with an associated soft tissue mass. CT after contrast typically reveals uniform enhancement (220,221). MR imaging with its multiplanar imaging capability is optimal to delineate lesion extent and involvement of surrounding structures prior to surgical resection (Fig. 9.31) (218,221). Mirich et al. reported intermediate signal intensity on T1-weighted and slightly hyperintense signal intensity with heterogenicity on T2-weighted MR images (221). In the case report by Atkinson et al., low intensity was present, presumably reflecting the paramagnetic properties of melanin (218). Our experience with this lesion on MR imaging is that the T2-weighted images show heterogeneity with prominent low-to-intermediate signal intensity, and these tumors mildly enhance following intravenous contrast administration (Fig. 9.31). Homogeneous stain during the venous phase of angiography has also been reported (220). We believe that patient age, lesion location, and MR features are distinctive of this lesion.
Figure 9.26 Ancient schwannoma in a man 51 years of age without neurofibromatosis. A: Axial CT shows low attenuation mass (asterisk) in the superior mediastinum. B,C: Fat-suppressed coronal T1-weighted (TR/TE; 500/20) MR images preceding (B) and following (C) contrast show a mass in the left upper chest with a large focus of nonenhancing hemorrhage inferiorly (asterisk). More solid tumor superiorly shows diffuse enhancement (arrow in C). D: Axial T2-weighted (TR/TE; 3500/110) MR image shows high signal intensity from both hemorrhagic and nonhemorrhagic areas.
Clear Cell Sarcoma
Clear cell sarcoma is a rare neoplasm accounting for 1% of all soft tissue sarcomas. It was originally described by Enzinger in 1965 and is also referred to as malignant melanoma of soft parts (222). Young adults between 20 and
P.357

40 years of age are most commonly involved. Women are affected slightly more frequently than men (3:2 ratio) (223,224,225). Clear cell sarcoma is a lesion intimately associated with or in a tendon, ligament, or aponeurosis. The lower extremity is involved in 75% of cases, with a particular predilection for the foot/ankle (43% of cases) followed by the knee and thigh (222,223,224,225,226,227,228,229,230,231). Other less commonly affected sites include the upper extremity (22% to 25%), trunk (2%), and head/neck (1%) (17,232). In the upper extremity, involvement of the hand and wrist is most common, paralleling the distribution in the lower extremity.
Figure 9.27 Neurothekeoma of the popliteal fossa in a 35 year old woman. A,B: Sagittal T1-weighted (A) and T2-weighted (B) spin-echo MR images show a lobular mass in the popliteal fossa. The mass reveals low-to-intermediate signal intensity on the short TR image and areas of marked high intensity on the long TR image. (Case courtesy of Martha C. Nelson, MD.)
Pathologically, clear cell sarcoma encompasses nearby tendon, ligament, or aponeurosis and is composed of nests or fascicles of cells with clear abundant cytoplasm (222,223,227,229). Intracellular melanin and premel-anosome granules can be identified in 72% of cases (222,223). Small amounts of iron are also seen in many patients. Immunohistochemical stains suggest neural crest origin and associated melanin production (S-100 protein, neuron-specific enolase homatropine methylbromide [HMB45]). Unlike malignant melanomas, cytogenetic abnormalities are reported in patients with this lesion, particularly translocation of the long arms of chromosomes 12 and 22 (17,227,233,234).
Treatment usually involves aggressive resection or amputation and adjuvant chemotherapy and radiation therapy (228,229). Unfortunately, prognosis in these patients is poor, with frequent local recurrence and metastases. Mortality rates range from 37% to 59% in the largest series (17,228,229). Survival at 5, 10, and 20 years was 67%, 33%, and 10%, respectively, in series at the Mayo Clinic (224,225). Multiple local recurrences are not uncommon and often precede development of distant metastases. Regional lymph node involvement develops in up to 50% of patients. In fact, regional lymph node dissection is often part of the initial treatment because of this common route of extension. Unfavorable prognostic factors include tumor size greater than 5 cm, necrosis, and local recurrence.
Radiographs in cases of clear cell sarcoma are often nonspecific with evidence of a soft tissue mass, although small lesions may not be detected. Associated bone involvement, by direct extension, has been considered rare in previous reports. However, we reviewed our cases at the Armed Forces Institute of Pathology (AFIP) and osseous invasion was seen in 5 of 14 patients (Fig. 9.32) (36%). The osseous involvement had an aggressive lytic appearance on radiographs, except in one case, where periosteal reaction alone was present. This frequency of osseous involvement should perhaps not be surprising, considering the primary tumor focus is within tendon or aponeurosis immediately adjacent to bone. Not unexpectedly, bone scans in patients with osseous extension show increased uptake of radionuclide. Calcification is only rarely reported in clear cell sarcoma (17).
Cross-sectional imaging of clear cell sarcoma by sonography, CT, or MR imaging more clearly suggests primary
P.358

involvement of the tendon, ligament, or aponeurosis. It is this location that should strongly suggest the diagnosis, and this anatomic relationship is best evaluated by MR imaging. CT scanning generally shows a soft tissue mass with attenuation equal to that of muscle. The lesion margins are often indistinct without evidence of a pseudocapsule. Because masses are often large at presentation and engulf multiple surrounding structures, it can be difficult to determine mass origin to a single tendon, ligament, or aponeurosis on axial CT images. Reports of sonographic evaluation of clear cell sarcoma are limited, but we would expect the lesion to appear as an aggressive infiltrating solid mass.
Figure 9.28 Perineurioma in a girl 13 years of age with right lower extremity neurologic symptoms. A–C: Coronal T1-weighted (TR/TE; 420/22) (A), enhanced fat-suppressed T1-weighted (TR/TE; 500/16) (B), and axial T2-weighted (TR/TE; 5000/106) (C) MR images show diffuse thickening of the sciatic nerve (arrows) with intermediate-to-high signal intensity on the long TR image and enhancement. Only the sciatic nerve is involved and not its branches, and no surrounding inflammation is seen.
The largest imaging series of 21 patients, reported by De Beuckeleer et al., noted a slightly higher signal intensity than that of muscle on T1-weighted MR images in 52% of patients as a helpful finding suggesting clear cell sarcoma (235). Although we have also observed this imaging appearance, in our experience, this is not uncommon in other neoplasms, and it is so mild that we do not find it useful as a distinguishing feature. MR imaging typically reveals relatively homogeneous signal intensity on T1-weighting with more heterogeneity on T2-weighting. In our 10 cases with MR imaging, 70% had overall higher signal intensity than that of fat, and only 30% showed low-to-intermediate signal intensity (lower than that of fat) on T2-weighted images (Figs. 9.32 and 9.33). Low signal intensity on long TR MR images and high signal on T1-weighting in clear cell sarcoma is attributed to the effects of melanin (paramagnetic relaxation enhancement of surrounding tissues) (Fig. 9.32). Low signal intensity on long TR images was reported in 50% of lesions in the study by Wetzel et al.
P.359

(231). They suggested this finding may be related to lesions with a large amount of melanin (231). Others have questioned this effect of melanin, and additional causes of low MR signal intensity include hemosiderin and fibrous content (236). However, the majority of cases of clear cell sarcomas, in our experience, do not show low signal intensity on T2-weighted MR images, but reveal nonspecific intermediate-to-high signal intensity. Prominent enhancement after intravenous contrast administration is often apparent on CT or MR imaging (237).
Figure 9.29 Granular cell tumor of subglottic region in a girl 8 years of age. A: Lateral radiograph after barium ingestion shows subglottic mass (asterisk). B: Axial noncontrast CT reveals a posterolateral mass (asterisk).
We do not believe the minute amounts of melanin in these lesions are likely to cause significant or characteristic signal intensity in agreement with studies that suggest up to 23% melanin is required to increase the T1-weighted signal. The most distinctive feature of clear cell sarcoma on MR imaging is its primary involvement of a tendon, ligament, or aponeurosis with fusiform growth proximally and distally along and within this affected fibrous structure (Figs. 9.32 and 9.33). This manifestation is particularly well-shown on long axis MR images. This intimate relationship of tumor to ligament, tendon, or aponeurosis is best appreciated in lesions that involve large tendons or are imaged early in their course, prior to extensive invasion of surrounding structures. Clear cell sarcomas that do not demonstrate this relationship to tendons have a nonspecific cross-sectional imaging appearance (Fig. 9.34). Involvement of subcutaneous tissues and dermis with ulceration is not infrequent in cases of clear cell sarcoma.
Paraganglioma
Paragangliomas are rare neuroendocrine neoplasms accounting for 0.03% of all neoplasms and 0.6% of all head and neck neoplasms (238,239,240). These tumors arise from the paraganglia, which represent neural crest cells associated with autonomic ganglia distributed from the skull base to the pelvic floor (241). Paragangliomas may be functional (producing catecholamines) or nonfunctional (not producing catecholamines-nonchromaffin). The World Health Organization subdivides these lesions into
P.360

four types, including: (a) those arising from the adrenal medulla (pheochromocytoma; not further discussed); (b) aorticosympathetic, related to the sympathetic chain and retroperitoneal ganglia; (c) parasympathetic, associated with the carotid, vagal, and visceral autonomic ganglia; and (d) other paragangliomas. Common sites of extra-adrenal paragangliomas include those of the carotid body (Fig. 9.35), jugulotympanic region, vagal body (Fig. 9.36), mediastinum (aortic body), and retroperitoneum, although numerous unusual foci are also described (242,243). Paragangliomas of either the carotid body or glomus jugulare account for 80% of all lesions.
Figure 9.30 Granular cell tumor in a woman 48 years of age with painless subcutaneous mass. A,B: Coronal T1-weighted (TR/TE; 420/14) MR image preceding (A) and fat-suppressed image following (B) contrast show a subcutaneous mass (asterisk) with prominent heterogeneous enhancement. Note mild surrounding edema (arrow). C: Sagittal T2-weighted (TR/TE; 4737/60) MR image shows the mass (asterisk), with low signal intensity, and mild surrounding edema (arrow).
Figure 9.31 Melanotic neuroectodermal tumor of infancy about the mandible in a newborn girl. A: Noncontrast CT reveals a destructive lesion of the mandible with associated protuberant soft tissue mass (asterisk). B,C: Axial T1-weighted (TR/TE; 450/13) (B) spin-echo and coronal short-tau inversion recovery (STIR) (TR/TE/TI; 4300/56/150) (C) MR images also show the extensive protuberant mass involving the nasopharynx (asterisk). There is predominantly intermediate signal intensity on T1-weighting and a heterogeneous pattern on the long TR image with areas of intermediate signal intensity (black asterisk) and other regions with low signal intensity (white asterisk).
P.361

Carotid body paragangliomas are commonly referred to as chemodectomas, a term proposed by Mulligan in 1950. Reports conflict on whether carotid body or glomus jugulare lesion is the most frequent extra-adrenal paraganglioma. However, many reports suggest carotid body lesions are most common and account for 60% of all head/neck paragangliomas. These paragangliomas are generally nonfunctional and arise from the carotid body (chemoreceptor to monitor arterial oxygen levels and blood pH) along the posterior aspect of the common carotid artery bifurcation (239). Chemodectomas are more frequent in high-altitude regions and in patients with chronic obstructive pulmonary disease, perhaps related to chronic hypoxia-induced hyperplasia. Patients are typically affected in the fifth to seventh decades (peak prevalence: 45 to 50 years of age), and the sex distribution is equal, except for lesions associated with high altitude where women predominate (241). Patients usually present with a slowly growing (5 mm per year) painless mass. Bruit may be apparent, and pressure applied to the mass may initiate symptoms of increased heart rate (carotid sinus syndrome). Other symptoms include hoarseness, stridor, vertigo, dysphasia, and tongue paresis. Treatment is usually surgical, which is curative in
P.362

most cases if the resection is complete (244). The metastases of malignant chemodectoma (6% to 9% of patients) most frequently affect regional lymph nodes (50% of metastases), lung, and bone (241,245,246,247). All large benign unresectable and malignant paragangliomas are radiosensitive.
Figure 9.32 Clear cell sarcoma of the foot in a woman 28 years of age. A: Lateral radiograph shows a large plantar mass with associated destruction of the calcaneus (asterisk). B: Noncontrast oblique coronal CT demonstrates the mass (asterisk) to be heterogeneous with an attenuation similar to that of muscle, extending into the calcaneus. C: Sagittal T2-weighted spin-echo MR image (TR/TE; 2000/90) reveals a large mass (large asterisk) involving the calcaneus, and extending into the plantar aponeurosis (small asterisk). Note low signal intensity of mass.
Multiple paragangliomas are present in up to 20% of patients, either synchronously or metachronously (248). This is particularly common in familial cases (10% of chemodectomas), 90% of which arise in the carotid body and show an autosomal dominant inheritance pattern (249). Familial paragangliomas occur at a younger age and reveal cytogenetic aberrations with evidence of linkage to 11q22.3-q23.2. Chemodectomas are bilateral in 2% to 7% of patients, and in patients with a familial lesion, the incidence of bilaterality increases to 31% (250,251,252). Paragangliomas may also be associated with multiple endocrine neoplasia (MEN), neuroectodermal syndromes (neurofibromatosis), and rarely, von Hippel-Lindau disease.
The second most common extra-adrenal paraganglioma in most series, is the jugulotympanic lesion, referred to as glomus jugulare tumor. These paragangliomas arise from the auricular branch of the vagus nerve, the tympanic branch of the glossopharyngeal nerve, or the jugular vein bulb (253,254,255,256). They are the most frequent neoplasms of the middle ear, and the vast majority are nonfunctional (241,256). Women are affected far more frequently than men (4–6:1 ratio), with the highest incidence in the fifth decade of life. Clinical symptoms are common, including pulsatile tinnitus, conductive hearing loss, and other cranial nerve palsies (40% of patients). Multiple staging systems accurately depict the clinical course of these lesions. Lesions that are amenable to complete surgical resection rarely recur. The incidence of malignant lesions with metastases (similar sites as seen with chemodectoma) is estimated at 1% to 4% (239).
Vagal paraganglioma was first described by Stout in 1935 and is the third most common extra-adrenal
P.363

paraganglioma, accounting for 5% of lesions (239). These lesions most commonly arise behind the angle of the mandible (83% of cases) (257). Vagal paragangliomas usually extend along the perineurium, as opposed to a focal mass, and they are slowly growing and painless. There is a female-to-male ratio of 2.7:1, and 10% to 15% of cases have multiple lesions (257). Surgical resection is difficult without sacrificing the vagus nerve. Malignancy with metastases occurs in 10% to 20% of lesions (239).
Figure 9.33 Clear cell sarcoma of the quadriceps tendon in a woman 69 years of age. A,B: Axial T1-weighted (TR/TE; 357/10) (A) and turbo T2-weighted (TR/TE; 4000/90) (B) spin-echo MR images reveal a mass centered in the quadriceps tendon (asterisk) with extension into the anterior soft tissues. The mass shows intermediate signal intensity on T1-weighting and heterogeneous intermediate-to-high signal on T2-weighting. C: Sagittal T2-weighted (TR/TE; 2200/80) spin-echo MR image shows the relationship of the mass (arrows) to the tendon to better advantage, as well as the high (black asterisk) and intermediate (white asterisk) signal intensity areas.
Other extra-adrenal paragangliomas arise related to the aortic body, within the posterior mediastinum or in the retroperitoneum along the periaortic sympathetic chain (including the organs of Zuckerkandl). The vast majority of these lesions are nonfunctional, except retroperitoneal lesions that produce norepinephrine in 25% to 60% of cases (258,259,260). Clinical symptoms are usually related to the mass and involvement of adjacent structures, although functional lesions may produce systemic signs such as hypertension, headaches, and palpitations with norepinephrine production as opposed to hypotension and cardiac arrhythmias in epinephrine-producing lesions. The incidence of malignant paragangliomas with metastases from these primary sites ranges from 2% to 16% (241). Metastatic foci usually involve regional lymph nodes, bone, liver, and lung.
Figure 9.34 Clear cell sarcoma of the buttocks in a girl 15 years of age with a nonspecific appearance. A,B: Axial T1-weighted (TR/TE; 700/15) (A) and T2-weighted (TR/TE; 2100/80) (B) spin-echo MR images reveal a mass in the gluteus maximus. The mass shows an intermediate signal intensity on T2-weighted images. There is marked associated edema. C: Axial T1-weighted image following gadolinium administration shows mild enhancement of the mass, with diffuse enhancement of the surrounding edema. D: The lesion is poorly delineated on corresponding noncontrast CT.
P.364

Pathologically, extra-adrenal paragangliomas are most commonly solid, defined masses. When bisected, a fibrous pseudocapsule and multiple blood vessels are well seen on the cut surface. Histologically, these lesions are composed of chief cells and sustentacular cells, with a surrounding fibrovascular stroma. The chief cells are arranged in compact nests (termed zellballen ball of cells).
Paragangliomas are richly vascular tumors, and radiologic imaging reflects this characteristic. Most masses are not apparent on radiographs, although larger lesions, particularly in a mediastinal location, may be apparent on chest radiographs. Ultrasonography of carotid body tumors is reported by Derchi et al. to reveal solid heterogeneous (1.2 cm to 5.0 cm) masses within the carotid bifurcation in 22 of 23 cases (261). Lesions are often heterogeneously hypoechoic, and Doppler analysis typically detects low-resistance arterial blood flow (Fig. 9.35) (262,263). CT and MR imaging also detect these masses and involvement of surrounding structures. Splaying of the common carotid bifurcation is a common manifestation of chemodectoma on ultrasound, CT, or MR imaging (264). CT reveals a soft tissue mass with marked postcontrast enhancement (Fig. 9.35) (265). Areas of osseous destruction and expansion are not uncommon with paragangliomas of the skull base and middle ear; these manifestations are best depicted by CT (256). MR imaging appearance has been described by Olsen et al. with demonstration of serpentine areas of signal void representing high vascular flow (12 of 15 cases), similar to arteriovenous hemangioma (Figs. 9.35 and 9.36) (246). Lesions are generally low-to-intermediate intensity on T1-weighted MR images and intermediate-to-very-high signal on T2-weighting (Figs. 9.35 and 9.36). Lesions were heterogeneous on all pulse sequences, creating a “salt and pepper” appearance on long TR sequences (Figs. 9.35 and 9.36) (246). The “salt” represents high signal intensity in slow-flow vascular components or hemorrhage, and the “pepper” represents areas of signal void from high-flow regions. The salt and pepper appearance occurs in lesions larger than 1 cm, but is not considered pathognomonic because it may occur in other hypervascular neoplasms (particularly renal or thyroid metastases). Similar to that seen on CT, an
P.365

intense pattern of enhancement occurs following intravenous contrast administration. Paragangliomas may reveal the so-called dropout phenomenon with MR angiography (266). Standard doses of gadolinium show a progressive increase in signal intensity caused by T1-shortening. However, with higher doses of contrast, magnetic susceptibility leads to a transient decreased signal intensity (T2-shortening) 24 to 42 seconds after injection, followed by progressive increased signal intensity (267,268).
Figure 9.35 Paraganglioma of the carotid body in several different patients. A: Angiogram shows intense staining in the chemodectoma (asterisks) with splaying of carotid vessels (arrows). B: Sonogram shows heterogeneous solid mass. Internal (large asterisk) and external (small asterisk) carotid arteries are well-seen. C,D: Contrast-enhanced axial CT (C) and sagittal reconstruction (D) show marked tumor enhancement (asterisk), splaying of carotid vessels (arrow), and high-flow serpentine internal vascularity (arrowheads). E,F: Axial T1-weighted (TR/TE; 366/22) (E) and T2-weighted (TR/TE; 4066/122) (F) spin-echo MR images show the mass (white arrow) to have well-defined margins, with heterogeneous, but predominantly intermediate signal intensity on long TR images. Low signal intensity flow voids (arrowheads) results from rapid blood flow, and the carotid vessels are displaced (black arrows). G: MR angiography study also shows splaying of carotid vessels by tumor. H: Intraoperative photograph demonstrate mass with splaying of carotid vessels (arrows).
Angiography of paragangliomas usually reveals profuse vascularity with large serpentine nutrient vessels, heterogeneous dense capillary stain, and early draining veins (Fig. 9.35) (250,252,269). In evaluation of chemodectomas, the contralateral carotid artery should also be studied to evaluate for a second tumor as well as to assess adequacy of cross filling, in case major vessels must be sacrificed at surgical resection (250,252). Preoperative embolization may be performed to lessen blood loss during surgery.
Scintigraphy with I-131 (or I-123) metaiodobenzylguanidine (MIBG), a structural analogue of norepinephrine, is also used to detect functional paragangliomas with a high degree of accuracy (242,270,271,272,273). In the study by von Gils et al., MIBG demonstrated 88% of extra-adrenal functioning paragangliomas as focal areas of marked radionuclide uptake (242). MIBG scanning is
P.366

reported to detect tumors as small as 0.2 grams and is also very useful to identify local recurrence and metastatic disease (242,271). However, false positives can occur associated with other neural crest tumors (neuroblastoma, schwannoma, medullary carcinoma of thyroid) and false negatives with certain medications (insulin, antidepressants, cocaine, amphetamines). In addition, spatial resolution limitations of scintigraphy usually require evaluation by CT or MR imaging as well, prior to surgical resection. Currently indium-111 octreotide is the agent of choice for nuclear medicine detection of paragangliomas (274,275,276). Octreotide imaging shows high sensitivity in detecting paragangliomas larger than 1.5 cm, but it is insensitive for lesions smaller than 1 cm (274,275,276). Octreotide is more sensitive than MIBG for identifying lesions, primarily because of its ability to detect both functional and nonfunctional paragangliomas (274,275,276). Octreotide is particularly useful in demonstrating multiple lesions, metastases, local recurrence, and in differentiating neurogenic tumors.
Figure 9.36 Paragangliomas (glomus jugulare and glomus tympanicum) in a woman 18 years of age complaining of pulsatile tinnitus. A,B: Axial T1-weighted (TR/TE; 600/15) (A) and fat-suppressed enhanced T1-weighted (TR/TE; 750/20) (B) spin-echo MR images show paragangliomas with prominent enhancement (arrow). Signal voids are indicative of the high degree of tumor vascularity. The patient had an additional paraganglioma with a small lesion about the ipsilateral carotid (not shown).
P.367

Primitive Neural Tumors
Neuroblastoma/Ganglioneuroblastoma/Ganglioneuroma
Primitive neural tumors arise from neural crest origin and thus, they occur along the sympathetic ganglia and within the adrenal medulla (not discussed further). In order of least to most differentiated, neuroblastoma is composed of primitive neuroblasts (resembling the fetal adrenal and representing a small, round, blue cell tumor), ganglioneuroblastoma contains both primitive neuroblast and maturing ganglion cells, and finally, the ganglioneuroma is made up of mature Schwann cells and ganglion cells (Figs. 9.37,9.38 and 9.39) (277,278,279). As expected, the malignant potential and behavior is directly related to the degree of differentiation. Clinically, patients usually present with evidence of a soft tissue mass or the secondary effects of the lesion on surrounding structures. Neuroblastoma patients frequently (80% to 90%) demonstrate elevated amounts of catecholamines and their byproducts, with resulting clinical symptoms (277). Ganglioneuroblastoma and ganglioneuroma less commonly manifest elevated catecholamine levels. Ganglioneuromas, even if large, are often asymptomatic. Numerous cytogenetic abnormalities are detected, with the most common related to the short arm of chromosome 1 (described in up to 70% of patients with neuroblastoma) (277).
Neuroblastoma is the third most common malignant tumor in childhood (1 per 10,000 live births), following leukemia and brain tumors, and it causes 15% of cancer deaths in children (277,278,280,281,282,283,284). The vast majority of these lesions occur sporadically, with a small number following an autosomal dominant pattern of inheritance. Neuroblastoma represents the most frequent solid neoplasm arising outside the central nervous system in infants and children. This tumor is significantly less common in blacks. Neuroblastoma and ganglioneuroblastoma are generally seen in young patients. In fact, 50% of neuroblastomas are found before the age of 2 years (25% congenital
P.368

and most of these are adrenal), 90% by 5 years of age, 96% in the first decade of life, and 3.5% in the second decade (277,278,281,285). The peak age of presentation of neuroblastoma is 18 months. Ganglioneuromas, in contrast, are usually diagnosed after the age of 10 years (277,285).
Figure 9.37 Ganglioneuroma of the posterior mediastinum in a girl 5 years of age. A: Chest radiograph shows a left paraspinal mass longer in its vertical extent with secondary osseous erosion (arrowheads). B: Contrast-enhanced axial CT reveals low attenuation, heterogeneous, posterior mediastinal mass (asterisk) with extension into neural foramina (arrowhead). C,D: Axial T1-weighted (TR/TE; 533/20) (C) spin-echo MR image preceding and coronal T1-weighted (TR/TE; 600/20) (D) image following gadolinium shows a large paraspinal mass (asterisk) with intermediate signal intensity and heterogeneous contrast enhancement. E: Axial T2-weighted (2000/100) spin-echo MR image reveals the mass (asterisk) to have high signal with extension into the neural foramina (arrowhead).
Neuroblastomas and ganglioneuroblastoma are located in a paramidline position from the skull base to the lower pelvis, with the adrenal (not discussed further) being the single most frequent site. Extra-adrenal neuroblastoma locations include the abdomen (63%), most of which are retroperitoneal (Fig. 9.39), posterior mediastinum (14%), neck (5%), pelvis (5%), and brain (2%) (281,286). Ganglioneuroblastomas most frequently affect the abdomen, followed by the mediastinum, neck, and lower extremity (Fig. 9.38). Ganglioneuromas are most common in the posterior mediastinum (39% to 43%) and the retroperitoneum (32% to 52%) (277,285) (Fig. 9.37). Neural immunohistochemical markers are usually positive but are nonspecific.
Treatment of these lesions is usually complete surgical removal, except in neuroblastoma patients with stage 4 disease. The most important prognostic factor for neuroblastoma is the clinical stage according to the International Neuroblastoma Staging System (Table 9.5) (284).
Figure 9.38 Ganglioneuroblastoma of the posterior mediastinum in a girl 2 years of age. A: Chest radiograph reveals a right posterior mediastinal mass (arrows). B: Noncontrast axial CT shows the large soft tissue mass (asterisk) with extension into a neural foramen (arrowhead). C,D: Axial T1-weighted spin-echo (TR/TE; 500/20) MR image preceding (C) and following (D) gadolinium administration show diffuse heterogeneous enhancement and extension into neural canal. (arrowhead by the mass).
Figure 9.39 Neuroblastoma of the retroperitoneum in a boy 18 months of age. Laboratory tests demonstrated elevated catecholamines. A: Abdominal sonogram shows an echogenic mass (asterisk) anterior to the kidney. Small echogenic focus with shadowing (arrowhead) resulting from calcification is seen. B: Contrast-enhanced axial CT reveals the mass (asterisk) containing small calcifications (arrowheads). Calcifications were not well-seen on radiographs (not shown).
TABLE 9.5 INTERNATIONAL NEUROBLASTOMA STAGING SYSTEM
Stage Definition
1 Localized tumor confined to area of origin; complete gross excision, with or without microscopic residual disease; representative ipsilateral lymph nodes negative for tumor.
2A Localized tumor with incomplete gross excision; representative ipsilateral nonadherent lymph nodes negative for tumor microscopically.
2B Localized tumor with or without complete gross excision, with ipsilateral nonadherent lymph nodes positive for tumor.
  Enlarged contralateral lymph nodes must be negative for tumor microscopically.
3 Unresectable unilateral tumor infiltrating across the midline, with or without regional lymph node involvement; or localized unilateral tumor with contralateral regional lymph node involvement; or midline tumor with bilateral extension by infiltration unresectable) or by lymph node involvement.
4 Any primary tumor with dissemination to distant lymph nodes, bone, bone marrow, liver, skin, and/or other organs (except as defined for stage 4S).
4S Localized primary tumor (as defined for stage 1, 2A, or 2B), with dissemination limited to skin, liver, and/or bone marrow.
P.369

P.370

Children with stage 1 and 2 disease have a 3-year survival rate of 90%; stage 4S, 80%; stage 3 and 4, 30%. The overall 3-year survival rate is 50% (277,284). Diagnosis of neuroblastoma at a younger age (particularly less than 1 year) also improves survival rates (77% vs. 38% for those with diagnosis at ages greater than 2 years) (284). Metastatic disease is present in nearly 70% of neuroblastoma patients at presentation, with bone, lymph node, liver, and skin the most frequent sites (281). Metastasis to bone (60% of cases) is a particularly ominous finding almost invariably associated with a fatal outcome (281). Spontaneous regression or maturation of neuroblastoma or ganglioneuroblastoma is reported in 1% to 2% of patients, usually in children younger than 1 year and in patients with stage 4S disease (277,281,284). Additional favorable prognostic factors include histologic tumor type and some laboratory and genetic markers (low serum ferritin, neuron-specific enolase, hyperdiploidy, lack of N-myc amplification gene sequence sites, high expression of TrkA [tyrosine kinase receptor A] nerve growth factor), and no allelic loss of 1p (277). Patients with neuroblastoma located in the retroperitoneum or adrenal have a worse prognosis that those with lesions in the pelvic or paraspinal (cervical/thoracic) regions.
Histologic features are also used to divide patients into those with and without differentiated stroma, with prognostic implications. These various factors can be used to divide patients into three risk groups: low, intermediate, and high. Low-risk patients are usually adequately managed with surgery alone. Intermediate-risk patients usually require chemotherapy and second-look surgery. High-risk patients receive dose-intensive chemotherapy; bone marrow and blood stem cell transplantation are also used. Patients with 4S disease are individualized for treatment depending on age (older or younger than 4 weeks) and for those older than 4 weeks of age, the presence of unfavorable biologic markers. Patients with stage 4 disease may be treated with radiation and chemotherapy as opposed to surgical excision (281). Ganglioneuroblastomas generally have a less malignant clinical course than neuroblastoma. Ganglioneuroma, in contrast, is almost invariably a benign tumor and surgical resection is curative. Ganglioneuromas only rarely undergo malignant transformation; this association has been reported in a patient with HIV.
Imaging of these neoplasms in extra-adrenal locations is often nonspecific. Location of the lesion along the sympathetic chain is important for diagnosis. Paraspinal lesions may extend into the spinal canal (Figs. 9.37 and 9.38). Calcification is seen in approximately 55% of retroperitoneal neuroblastomas on radiographs (85% by CT) (Fig. 9.39) and in 20% to 30% of ganglioneuromas (281,287). CT and MR imaging are best to evaluate and stage disease extent of the soft tissue mass (280,281,282,283,284,285). CT is superior to detect the presence of calcification (Figs. 9.37,9.38 and 9.39). Dietrich et al. suggested that MR imaging is superior to CT in determining tumor extent because of its multiplanar capabilities and superior contrast resolution (282). The MR imaging signal intensity of neuroblastoma is generally nonspecific with low-to-intermediate intensity on T1-weighting and intermediate-to-high signal on T2-weighting. Similarly, CT reveals a nonspecific soft tissue attenuation mass. Higher signal intensity on T2-weighted images likely corresponds to increased myxoid tissue and is more common in ganglioneuroma (Fig. 9.37) (288). Curvilinear bands of low signal intensity on long TR images causing a whorled appearance have been described in ganglioneuroma (289,290). Amundson et al., in reporting the ultrasound results in 10 cases, noted focal areas or lobules of increased echogenicity within 4 (40%) of ganglioneuroblastomas neuroblastomas in their series, and they believe this finding is characteristic because it was not seen in 43 other abdominal neoplasms (280,291,292). Ganglioneuromas tend to surround adjacent major vascular structures partially or completely without compromising blood flow. CT scanning of ganglioneuroma may show gradual delayed heterogeneous contrast enhancement. We are not aware of any radiologic findings that distinguish neuroblastoma, ganglioneuroblastoma, or ganglioneuroma. Paraspinal ganglioneuroblastoma and ganglioneuroma
P.371

typically are long, vertical, posterior mediastinal tumors (285,293,294). Radioiodine MIBG scintigraphy may demonstrate radionuclide uptake within both primary and metastatic primitive neuroectodermal neoplasms (295,296).
Primitive Neuroectodermal Tumor (PNET)/Extraskeletal Ewing Sarcoma
Extraskeletal Ewing sarcoma and primitive neuroectodermal tumor (PNET) are similar soft tissue sarcomas likely derived from neuroectodermal origin (279,297). Ewing sarcoma was originally described in 1921; PNET was reported 3 years earlier by Stout (298,299). Synonyms for PNET include peripheral neuroepithelioma and peripheral neuroblastoma. Extraskeletal Ewing sarcoma and PNET can occur in bone or soft tissue, although our discussion is limited to soft tissue lesions. These neoplasms previously were considered to be distinct lesions because of purported differences in clinical behavior and pathologic appearance (300). Although still controversial, newer studies suggest a common genetic abnormality and clinical behavior, and these lesions may well represent a single entity; therefore, it may be preferable to consider these lesions together in the same family of tumors (300).
Both extraskeletal Ewing sarcoma and PNET usually affect young patients between 10 and 30 years of age (277,301). However, in review of some studies, the age range of patients was reported to be wider in PNET: from birth to 81 years (301,302,303,304,305,306). Males are affected slightly more commonly than females. Extraskeletal Ewing sarcoma is rarely seen in blacks. The most frequently involved locations include the paravertebral region, chest wall (PNET Askin tumor; see later discussion), retroperitoneum, and lower extremities. Clinical presentation is usually that of a rapidly growing soft tissue mass, often 5 cm to 10 cm in size at diagnosis. No elevation of catecholamine levels is seen, unlike neuroblastoma. Interestingly, 90% to 95% of patients show a reciprocal translocation of the long arm of chromosomes 11 and 22 (q24;q12) with both extraskeletal Ewing sarcoma and PNET (307).
Pathologically, these lesions are small, round, blue cell neoplasms, usually rich in glycogen. Previous distinction of PNET from extraskeletal Ewing sarcoma required detection of rosette formation in the former. In addition, electron microscopy revealed filaments and microtubules, and immunohistochemical stains should be positive for neuron-specific enolase and at least one other neural marker in PNET, although this distinction has become obscured over time (277,301,302,303,304,305,306,307,308). Both lesions express the product of the MIC2 gene in up to 95% of cases, providing further credence to a histogenetic linkage.
Treatment usually involves chemotherapy followed by surgical resection, and prognosis has progressively improved (309,310). Radiation therapy may also be used in some cases as an adjunct. Several studies have shown that the overall prognosis of PNET is worse than that of extraskeletal Ewing sarcoma (311). Schmidt et al. reported a disease-free survival rate of only 45% with PNET versus 60% for extraskeletal Ewing sarcoma (311). Other reports suggest a survival rate of only 30% for PNET and 65% to 70% for extraskeletal Ewing sarcoma (3-year survival rate of 80% for smaller tumors vs. 32% for larger lesions) (277,303,306). However, other studies do not confirm this disparity in prognosis, with 5-year survival rates in PNET ranging from 56% to 68% (312). Factors that worsen prognosis are large tumor size at diagnosis and evidence of extensive necrosis. Metastases and local recurrence are usually seen in the first 2 years after diagnosis. Sites of metastatic involvement are most commonly lung and bone.
Imaging appearance of these lesions has not been extensively evaluated (309,313,314,315). Findings are usually nonspecific, and to the best of our knowledge, no imaging features are described to allow distinction of extraskeletal Ewing sarcoma from PNET. O’Keeffe et al. reported that most frequently the lesions are hypoechoic on ultrasound and low attenuation on CT (314). However, in our experience, CT scanning shows attenuation similar to that of muscle without evidence of calcification. MR imaging also reveals nonspecific features with low-to-intermediate signal intensity on T1-weighting and generally intermediate-to-high signal intensity on T2-weighting (Fig. 9.40). The high cellularity of these lesions likely accounts for the common appearance of intermediate signal intensity on long TR images. Areas of hemorrhage, with high signal on all MR imaging pulse sequences, are not infrequent. Neoplasm margins may be relatively well-defined with a pseudocapsule, or may appear infiltrative.
Figure 9.40 Extraskeletal Ewing sarcoma of the calf in a girl 5 years of age. A,B: Coronal T1-weighted (TR/TE; 633/16) (A) and axial turbo T2-weighted (TR/TE; 3000/102) (B) spin-echo MR images show a large soft tissue mass in the lateral aspect of the lower leg. The mass is relatively well-defined and homogeneous with nonspecific signal intensity. Flow voids (arrowhead) represent rapidly flowing blood in small peripheral vascular channels are also seen. C: Axial T1-weighted (TR/TE; 800/16) spin-echo MR image with fat suppression following contrast administration shows striking contrast enhancement, reflecting the lesion’s marked vascularity.
P.372

The angiographic appearance is described as hypervascular, although this is not an invariable finding. MR imaging frequently reveals definable high-flow (low intensity on all MR imaging pulse sequences) vascular channels within the mass, often more prominent peripherally (Fig. 9.40). This is not a unique feature of PNET or extraskeletal Ewing sarcoma, since it can also be seen with vascular neoplasms (hemangiopericytoma, hemangioendothelioma, and angiosarcoma), alveolar soft part sarcoma, and alveolar rhabdomyosarcoma. Prominent contrast enhancement is also a feature of these lesions, in our experience, on CT or MR imaging, although O’Keeffe et al. only reported enhancement in 4 of 11 patients on CT (314). Associated involvement of bone is unusual, and bone scintigraphy is usually normal.
Askin Tumor
Extraskeletal Ewing sarcoma/PNET occurring in the chest wall was formerly referred to as Askin tumor, and this term is often retained for lesions in the thoracopulmonary region (316). In 1979, Askin et al. described a small cell malignant tumor of the thoracopulmonary region in children and adolescents that is now generally believed to be equivalent to other extraskeletal Ewing sarcoma/PNET (316). These lesions are usually large and seen in young adults and children. A chest wall mass is the typical clinical presentation with or without pain. Lesions are most common in young females (3:1), and are almost exclusively unilateral (277,316). Associated rib destruction is common and seen in more than 50% of cases (317). There is typically a large associated pleural effusion, which may be loculated or form pseudotumors. Constitutional symptoms, including fever, anorexia, and weight loss, may also be present.
Grossly, the lesions are multilobulated, gray-white masses, with foci of hemorrhage and necrosis. The tumor is usually circumscribed but not encapsulated. On microscopy, the tumor is composed of small round cells with an identical appearance, as previously described, for extraskeletal Ewing sarcoma/PNET.
Figure 9.41 Primitive neuroectodermal tumor (PNET) of the chest wall (Askin tumor) in a girl 5 years of age. A: Chest radiograph shows opacification of the right hemithorax. B: Noncontrast axial CT reveals a large heterogeneous mass with low attenuation due to necrosis. C,D: Coronal T1-weighted (TR/TE; 483/20) (C) and axial T2-weighted (TR/TE; 2200/80) (D) spin-echo MR images show a large complex mass, with areas of necrosis. Foci of increased signal represent subacute hemorrhage. E: Axial T1-weighted (TR/TE; 800/20) spin-echo MR image following contrast administration shows solid areas of intense enhancement and low signal intensity nonenhancing necrotic regions.
P.373

Chest radiographs reveal a large pleural mass, which is usually a combination of pleural mass and fluid. Associated parenchymal disease is seen in approximately 25% of cases (Fig. 9.41) (318,319). Calcification is seen on radiographs in 10% of cases (318,319). Ipsilateral hilar and mediastinal adenopathy, as well as pneumothorax may also be present.
On CT scans, the lesion appears as a unilateral heterogeneous mass of mixed attenuation, often with associated rib destruction (25% to 63% of cases) and pleural fluid (318,319,320). In the study by Winer-Muram et al, lesions were commonly heterogeneous with signal intensity greater than that of muscle on T1-weighting (88%) (Fig. 9.41) (320). On T2-weighted MR images, high signal intensity was seen in all cases (Fig. 9.41) (320). Prominent contrast enhancement has been reported, reflecting the rich vascularity of these lesions on MR imaging (319,320,321).
Metastases are seen at presentation in 10% to 38% of patients (277). Recurrent thoracic disease is seen in more
P.374

than 50% of patients, either with local recurrent disease, mediastinal nodes, or pulmonary nodules. Bone metastases develop in approximately 25% of patients, and lesions may be osteoblastic (277,316). The prognosis is poor; the median survival in Askin’s original 20 patients was only 8 months (316).
References
1. Kransdorf MJ. Malignant soft-tissue tumors in a large referral population: distribution of diagnoses by age, sex, and location. AJR Am J Roentgenol. 1995;164:129–134.
2. Damjanov I. Anderson’s pathology. In: Kissane J, ed. Peripheral Nervous System. 10th ed. St. Louis: Mosby; 1996:2799–2803.
3. Fornage BD. Peripheral nerves of the extremities: imaging with US. Radiology. 1988;167:179–182.
4. Graif M, Seton A, Nerubai J, et al. Sciatic nerve: sonographic evaluation and anatomic-pathologic considerations. Radiology. 1991;181:405–408.
5. Ikeda K, Haughton VM, Ho KC, et al. Correlative MR-anatomic study of the median nerve. AJR Am J Roentgenol. 1996;167: 1233–1236.
6. Filler AG, Howe FA, Hayes CE, et al. Magnetic resonance neurography. Lancet. 1993;341:659–661.
7. Martinoli C, Bianchi S, Derchi LE. Tendon and nerve sonography. Radiol Clin North Am. 1999;37:691–711, viii.
8. Silvestri E, Martinoli C, Derchi LE, et al. Echotexture of peripheral nerves: correlation between US and histologic findings and criteria to differentiate tendons. Radiology. 1995;197: 291–296.
9. Enzinger F, Weiss SW. Benign tumors of peripheral nerves. In: Weiss SW, Goldblum JR, eds. Soft Tissue Tumors. 3rd ed. St. Louis: Mosby; 1995:821–888.
10. Das Gupta TK, Brasfield RD. Amputation neuromas in cancer patients. N Y State J Med. 1969;69:2129–2132.
11. Donnal JF, Blinder RA, Coblentz CL, et al. MR imaging of stump neuroma. J Comput Assist Tomogr. 1990;14:656–657.
12. Fisher GT, Boswick JA Jr. Neuroma formation following digital amputations. J Trauma. 1983;23:136–142.
13. Huber G, Lewis D. Amputation neuromas: their development and prevention. Arch Surg. 1920;1:85–113.
14. Singson RD, Feldman F, Slipman CW, et al. Postamputation neuromas and other symptomatic stump abnormalities: detection with CT. Radiology. 1987;162:743–745.
15. Singson RD, Feldman F, Staron R, et al. MRI of postamputation neuromas. Skeletal Radiol. 1990;19:259–262.
16. Spencer PS. The traumatic neuroma and proximal stump. Bull Hosp Joint Dis. 1974;35:85–102.
17. Enzinger F, Weiss SW. Malignant tumors of Peripheral Nerves. In: Soft Tissue Tumors. 3rd ed. St. Louis: Mosby; 1995:889–928.
18. Lee SH, Jung HG, Park YC, et al. Results of neurilemoma treatment: a review of 78 cases. Orthopedics. 2001;24:977–980.
19. Appiah-Anane S. Amputation neuroma: a late complication following sagittal split osteotomy of the mandible. J Oral Maxillofac Surg. 1991;49:1218–1220.
20. Detenbeck LC. Infrapatellar traumatic neuroma resulting from dashboard injury. J Bone Joint Surg Am. 1972;54:170–172.
21. Feller SR, Evans RD. Proper digital neuroma of the fifth toe. J Am Podiatr Med Assoc. 1996;86:187–188.
22. Gregg JM. Studies of traumatic neuralgias in the maxillofacial region: surgical pathology and neural mechanisms. J Oral Maxillofac Surg. 1990;48:228–237; discussion 238–239.
23. Peszkowski MJ, Larsson A. Extraosseous and intraosseous oral traumatic neuromas and their association with tooth extraction. J Oral Maxillofac Surg. 1990;48:963–967.
24. Rodeo SA, Sobel M, Weiland AJ. Deep peroneal-nerve injury as a result of arthroscopic meniscectomy. A case report and review of the literature. J Bone Joint Surg Am. 1993;75:1221–1224.
25. Sist TC Jr, Greene GW. Traumatic neuroma of the oral cavity. Report of thirty-one new cases and review of the literature. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 1981;51: 394–402.
26. Kransdorf MJ, Murphey M. Neurogenic tumors. In: Imaging of Soft Tissue Tumors. Philadelphia: WB Saunders; 1997: 235–273.
27. Attarian DE. Neuromas of the superficial radial nerve. Mil Med. 1988;153:393–394.
28. Kneeland JB, Kellman GM, Middleton WD, et al. Diagnosis of diseases of the supraclavicular region by use of MR imaging. AJR Am J Roentgenol. 1987;148:1149–1151.
29. Stahl S, Kaufman T, Ben-David B. Neuroma of the superficial branch of the radial nerve after intravenous cannulation. Anesth Analg. 1996;83:180–182.
30. Gupta RK, Mehta VS, Banerji AK, et al. MR evaluation of brachial plexus injuries. Neuroradiology. 1989;31:377–381.
31. Boutin RD, Pathria MN, Resnick D. Disorders in the stumps of amputee patients: MR imaging. AJR Am J Roentgenol. 1998;171: 497–501.
32. Beggs I. Pictorial review: imaging of peripheral nerve tumours. Clin Radiol. 1997;52:8–17.
33. Bodner G, Huber B, Schwabegger A, et al. Sonographic detection of radial nerve entrapment within a humerus fracture. J Ultrasound Med. 1999;18:703–706.
34. Provost N, Bonaldi VM, Sarazin L, et al. Amputation stump neuroma: ultrasound features. J Clin Ultrasound. 1997;25:85–89.
35. Peer S, Bodner G, Meirer R, et al. Examination of postoperative peripheral nerve lesions with high-resolution sonography. AJR Am J Roentgenol. 2001;177:415–419.
36. Henrot P, Stines J, Walter F, et al. Imaging of the painful lower limb stump. Radiographics. 2000;20(special issue):S219–S235.
37. Huang LF, Weissman JL, Fan C. Traumatic neuroma after neck dissection: CT characteristics in four cases. AJNR Am J Neuroradiol. 2000;21:1676–1680.
38. Yabuuchi H, Kuroiwa T, Fukuya T, et al. Traumatic neuroma and recurrent lymphadenopathy after neck dissection: comparison of radiologic features. Radiology. 2004;233:523–529.
39. Burchiel KJ, Johans TJ, Ochoa J. The surgical treatment of painful traumatic neuromas. J Neurosurg. 1993;78:714–719.
40. Tada K, Nakashima H, Yoshida T, et al. A new treatment of painful amputation neuroma: a preliminary report. J Hand Surg [Br]. 1987;12:273–276.
41. Ernberg LA, Adler RS, Lane J. Ultrasound in the detection and treatment of a painful stump neuroma. Skeletal Radiol. 2003;32: 306–309.
42. Thomas AJ, Bull MJ, Howard AC, et al. Peri operative ultrasound guided needle localisation of amputation stump neuroma. Injury. 1999;30:689–691.
43. Reed RJ, Bliss BO. Morton’s neuroma. Regressive and productive intermetatarsal elastofibrositis. Arch Pathol. 1973;95:123–129.
44. Morton T. A peculiar and painful affection of the fourth metatarsophalangeal articulation. Am J Med Sci. 1876;71:37–45.
45. Alexander IJ, Johnson KA, Parr JW. Morton’s neuroma: a review of recent concepts. Orthopedics. 1987;10:103–106.
46. Mann RA, Reynolds JC. Interdigital neuroma—a critical clinical analysis. Foot Ankle. 1983;3:238–243.
47. Lassmann G. Morton’s toe: clinical, light and electron microscopic investigations in 133 cases. Clin Orthop. 1979;142:73–84.
48. Mulder JD. The causative mechanism in Morton’s metatarsalgia. J Bone Joint Surg Br. 1951;33-B:94–95.
49. Torriani M, Kattapuram SV. Technical innovation. Dynamic sonography of the forefoot: The sonographic Mulder sign. AJR Am J Roentgenol. 2003;180:1121–1123.
50. Bossley CJ, Cairney PC. The intermetatarsophalangeal bursa—its significance in Morton’s metatarsalgia. J Bone Joint Surg Br. 1980;62-B:184–187.
51. Zanetti M, Strehle JK, Zollinger H, et al. Morton neuroma and fluid in the intermetatarsal bursae on MR images of 70 asymptomatic volunteers. Radiology. 1997;203:516–520.
52. Bencardino J, Rosenberg ZS, Beltran J, et al. Morton’s neuroma: is it always symptomatic? AJR Am J Roentgenol. 2000;175:649–653.
53. Zanetti M, Strehle JK, Kundert HP, et al. Morton neuroma: effect of MR imaging findings on diagnostic thinking and therapeutic decisions. Radiology. 1999;213:583–588.
54. Zanetti M, Ledermann T, Zollinger H, et al. Efficacy of MR imaging in patients suspected of having Morton’s neuroma. AJR Am J Roentgenol. 1997;168:529–532.
P.375

55. Redd RA, Peters VJ, Emery SF, et al. Morton neuroma: sonographic evaluation. Radiology. 1989;171:415–417.
56. Kaminsky S, Griffin L, Milsap J, et al. Is ultrasonography a reliable way to confirm the diagnosis of Morton’s neuroma? Orthopedics. 1997;20:37–39.
57. Turan I, Lindgren U, Sahlstedt T. Computed tomography for diagnosis of Morton’s neuroma. J Foot Surg. 1991;30:244–245.
58. Mendicino SS, Rockett MS. Morton’s neuroma. Update on diagnosis and imaging. Clin Podiatr Med Surg. 1997;14:303–311.
59. Resch S, Stenstrom A, Jonsson A, et al. The diagnostic efficacy of magnetic resonance imaging and ultrasonography in Morton’s neuroma: a radiological-surgical correlation. Foot Ankle Int. 1994;15:88–92.
60. Sartoris DJ, Brozinsky S, Resnick D. Magnetic resonance images. Interdigital or Morton’s neuroma. J Foot Surg. 1989;28:78–82.
61. Shapiro PP, Shapiro SL. Sonographic evaluation of interdigital neuromas. Foot Ankle Int. 1995;16:604–606.
62. Sobiesk GA, Wertheimer SJ, Schulz R, et al. Sonographic evaluation of interdigital neuromas. J Foot Ankle Surg. 1997;36:364–366.
63. Oliver TB, Beggs I. Ultrasound in the assessment of metatarsalgia: a surgical and histological correlation. Clin Radiol. 1998;53: 287–289.
64. Quinn TJ, Jacobson JA, Craig JG, et al. Sonography of Morton’s neuromas. AJR Am J Roentgenol. 2000;174:1723–1728.
65. Pollak RA, Bellacosa RA, Dornbluth NC, et al. Sonographic analysis of Morton’s neuroma. J Foot Surg. 1992;31:534–537.
66. Read JW, Noakes JB, Kerr D, et al. Morton’s metatarsalgia: sonographic findings and correlated histopathology. Foot Ankle Int. 1999;20:153–161.
67. Erickson SJ, Canale PB, Carrera GF, et al. Interdigital (Morton) neuroma: high-resolution MR imaging with a solenoid coil. Radiology. 1991;181:833–836.
68. Hoskins CL, Sartoris DJ, Resnick D. Magnetic resonance imaging of foot neuromas. J Foot Surg. 1992;31:10–16.
69. Terk MR, Kwong PK, Suthar M, et al. Morton neuroma: evaluation with MR imaging performed with contrast enhancement and fat suppression. Radiology. 1993;189:239–241.
70. Williams JW, Meaney J, Whitehouse GH, et al. MRI in the investigation of Morton’s neuroma: which sequences? Clin Radiol. 1997;52:46–49.
71. Theodoresco B, Lalande G. Magnetic resonance and Morton’s neuroma [in French]. Rev Chir Orthop Reparatrice Appar Mot. 1991;77:273–275.
72. Unger HR Jr, Mattoso PQ, Drusen MJ, et al. Gadopentetate-enhanced magnetic resonance imaging with fat saturation in the evaluation of Morton’s neuroma. J Foot Surg. 1992;31:244–246.
73. Weishaupt D, Treiber K, Jacob HA, et al. MR imaging of the forefoot under weight-bearing conditions: position-related changes of the neurovascular bundles and the metatarsal heads in asymptomatic volunteers. J Magn Reson Imaging. 2002;16:75–84.
74. Weishaupt D, Treiber K, Kundert HP, et al. Morton neuroma: MR imaging in prone, supine, and upright weight-bearing body positions. Radiology. 2003;226:849–856.
75. Beskin J. Primary and salvage procedures for Morton’s interdigital neuroma. In: Myerson M, ed. Current Therapy in Foot and Ankle Surgery. St. Louis: Mosby-Yearbook; 1994:183–187.
76. Biasca N, Zanetti M, Zollinger H. Outcomes after partial neurectomy of Morton’s neuroma related to preoperative case histories, clinical findings, and findings on magnetic resonance imaging scans. Foot Ankle Int. 1999;20:568–575.
77. Wu J, Chiu DT. Painful neuromas: a review of treatment modalities. Ann Plast Surg. 1999;43:661–667.
78. Ruuskanen MM, Niinimaki T, Jalovaara P. Results of the surgical treatment of Morton’s neuralgia in 58 operated intermetatarsal spaces followed over 6 (2–12) years. Arch Orthop Trauma Surg. 1994;113:78–80.
79. Weiss S, Goldblum J. Benign tumors of peripheral nerves. In: Weiss SW, Goldblum JR, eds. Enzinger and Weiss’s Soft Tissue Tumors. 4th ed. St. Louis: Mosby; 2001:1111–1207.
80. Miettinen M. Nerve sheath tumors. In: Diagnostic Soft Tissue Pathology. New York: Churchill Livingstone; 2003:343–378.
81. Kransdorf MJ. Benign soft-tissue tumors in a large referral population: distribution of specific diagnoses by age, sex, and location. AJR Am J Roentgenol. 1995;164:395–402.
82. Scheithauer BW, Woodruff JM, Erlandson RA. Tumors of peripheral nervous system. In: Rosai J, Sobin LH, eds. Atlas of Tumor Pathology. Washington, DC: Armed Forces Institute of Pathology; 1999;1–415.
83. Stout A. The peripheral manifestations of the specific nerve sheath tumor (neurilemmoma). Am J Cancer. 1935;24:751–796.
84. Sheikh S, Gomes M, Montgomery E. Multiple plexiform schwannomas in a patient with neurofibromatosis. J Thorac Cardiovasc Surg. 1998;115:240–242.
85. Ogose A, Hotta T, Morita T, et al. Multiple schwannomas in the peripheral nerves. J Bone Joint Surg Br. 1998;80:657–661.
86. Alberghini M, Zanella L, Bacchini P, et al. Cellular schwannoma: a benign neoplasm sometimes overdiagnosed as sarcoma. Skeletal Radiol. 2001;30:350–353.
87. Russell D, Rubinstein L. Pathology of Tumors of the Nervous System. Baltimore: William and Wilkins; 1989.
88. Scheithauer B, Woodruff JM, Erlandson RA. Atlas of Tumor Pathology: Tumors of the Peripheral Nervous System. Bethesda, MD: Armed Forces Institute of Pathology; 1999.
89. Walker CW, Adams BD, Barnes CL, et al. Case report 667. Fibrolipomatous hamartoma of the median nerve. Skeletal Radiol. 1991;20:237–239.
90. Shaw RJ, Paez JG, Curto M, et al. The Nf2 tumor suppressor, Merlin, functions in Rac-dependent signaling. Dev Cell. 2001;1:63–72.
91. Ferner RE, O’Doherty MJ. Neurofibroma and schwannoma. Curr Opin Neurol. 2002;15:679–684.
92. Kehoe NJ, Reid RP, Semple JC. Solitary benign peripheral-nerve tumours. Review of 32 years’ experience. J Bone Joint Surg Br. 1995;77:497–500.
93. Donner TR, Voorhies RM, Kline DG. Neural sheath tumors of major nerves. J Neurosurg. 1994;81:362–373.
94. von Recklinghausen F. Ueber die multiplen Fibrome der Haut und igre Beziehung zu den multiplen Neuromen. Berlin: August Hirschwald; 1882.
95. Cnossen MH, de Goede-Bolder A, van den Broek KM, et al. A prospective 10 year follow up study of patients with neurofibromatosis type 1. Arch Dis Child. 1998;78:408–412.
96. Friedman DP. Segmental neurofibromatosis (NF-5): a rare form of neurofibromatosis. AJNR Am J Neuroradiol. 1991;12: 971–972.
97. Evans DG, Huson SM, Donnai D, et al. A genetic study of type 2 neurofibromatosis in the United Kingdom. I. Prevalence, mutation rate, fitness, and confirmation of maternal transmission effect on severity. J Med Genet. 1992;29:841–846.
98. Neurofibromatosis. Conference statement. In: National Institutes of Health Consensus Development Conference. Arch Neurol. 1988;575–578.
99. Neurofibromatosis 1 (Recklinghausen disease) and neurofibromatosis 2 (bilateral acoustic neurofibromatosis). An update. In: ed. National Institutes of Health Conference. Ann Intern Med. 1990;39–52.
100. Feldman F. Tuberous sclerosis, neurofibromatosis, and fibrous dysplasia. In: Diagnosis of Bone and Joint Disorders. 3rd ed. Philadelphia: WB Saunders; 1995:4361–4379.
101. Gutmann DH, Aylsworth A, Carey JC, et al. The diagnostic evaluation and multidisciplinary management of neurofibromatosis 1 and neurofibromatosis 2. JAMA. 1997;278:51–57.
102. Huson SM, Hughes RAC. Neurofibromatosis 1: a clinical and genetic overview. In: SM H, RAC H, eds. The Neurofibromatoses: A Pathogenetic and Clinical Overview. Cambridge: Cambridge University Press; 1994;60–203.
103. Mulvihill JJ, Parry DM, Sherman JL, et al. NIH conference. Neurofibromatosis 1 (Recklinghausen disease) and neurofibromatosis 2 (bilateral acoustic neurofibromatosis). An update. Ann Intern Med. 1990;113:39–52.
104. Holt JF. 1977 Edward B. D. Neuhauser lecture: neurofibromatosis in children. AJR Am J Roentgenol. 1978;130:615–639.
105. Riccardi VM. Von Recklinghausen neurofibromatosis. N Engl J Med. 1981;305:1617–1627.
106. Ruggieri M, Huson SM. The clinical and diagnostic implications of mosaicism in the neurofibromatoses. Neurology. 2001;56: 1433–1443.
107. Ingordo V, D’Andria G, Mendicini S, et al. Segmental neurofibromatosis: is it uncommon or underdiagnosed? Arch Dermatol. 1995;131:959–960.
P.376

108. Ogose A, Hotta T, Imaizumi S, et al. Deep-seated segmental neurofibromatosis without cafe au lait spots. Skeletal Radiol. 2000;29:543–547.
109. Sieb JP, Schultheiss R. Segmental neurofibromatosis of the sciatic nerve: case report. Neurosurgery. 1992;31:1122–1125; discussion 1125.
110. Harris WC Jr, Alpert WJ, Marcinko DE. Elephantiasis neuromatosa in von Recklinghausen’s disease. A review and case report. J Am Podiatry Assoc. 1982;72:70–72.
111. Hertzanu Y, Hirsch M, Peiser J, et al. Computed tomography of elephantiasis neuromatosa. J Comput Assist Tomogr. 1989;13:156–158.
112. Roy SM, Ghosh AK. Elephantiasis neuromatosa: a clinicopathologic study of four cases. J Indian Med Assoc. 1992;90: 185–187.
113. Stevens KJ, Ludman CN, Sully L, et al. Magnetic resonance imaging of elephantiasis neuromatosa. Skeletal Radiol. 1998;27: 696–701.
114. Packer RJ, Gutmann DH, Rubenstein A, et al. Plexiform neurofibromas in NF1: toward biologic-based therapy. Neurology. 2002;58:1461–1470.
115. Evans DG, Baser ME, McGaughran J, et al. Malignant peripheral nerve sheath tumours in neurofibromatosis 1. J Med Genet. 2002;39:311–314.
116. Sorensen SA, Mulvihill JJ, Nielsen A. Long-term follow-up of von Recklinghausen neurofibromatosis. Survival and malignant neoplasms. N Engl J Med. 1986;314:1010–1015.
117. Kleihues P, Cavenee W. World Health Organization Classification of Tumors. In: Fletcher CDM, Unni KK, Mertens F, eds. Pathology and Genetics Tumors of the Nervous System. Lyon, France: IARC Press; 2000.
118. Weiss S, Goldblum J. Malignant tumors of the peripheral nerves. In: Weiss SW, Goldblum JR, eds. Enzinger and Weiss’s Soft Tissue Tumors. 4th ed. St. Louis: Mosby; 2001:1209–1263.
119. Ducatman BS, Scheithauer BW, Piepgras DG, et al. Malignant peripheral nerve sheath tumors. A clinicopathologic study of 120 cases. Cancer. 1986;57:2006–2021.
120. Wanebo JE, Malik JM, VandenBerg SR, et al. Malignant peripheral nerve sheath tumors. A clinicopathologic study of 28 cases. Cancer. 1993;71:1247–1253.
121. Kourea HP, Cordon-Cardo C, Dudas M, et al. Expression of p27(kip) and other cell cycle regulators in malignant peripheral nerve sheath tumors and neurofibromas: the emerging role of p27(kip) in malignant transformation of neurofibromas. Am J Pathol. 1999;155:1885–1891.
122. Nielsen GP, Stemmer-Rachamimov AO, Ino Y, et al. Malignant transformation of neurofibromas in neurofibromatosis 1 is associated with CDKN2A/p16 inactivation. Am J Pathol. 1999; 155:1879–1884.
123. Hruban RH, Shiu MH, Senie RT, et al. Malignant peripheral nerve sheath tumors of the buttock and lower extremity. A study of 43 cases. Cancer. 1990;66:1253–1265.
124. Biondetti PR, Vigo M, Fiore D, et al. CT appearance of generalized von Recklinghausen neurofibromatosis. J Comput Assist Tomogr. 1983;7:866–869.
125. Berlin O, Stener B, Lindahl S, et al. Vascularization of peripheral neurilemomas: angiographic, computed tomographic, and histologic studies. Skeletal Radiol. 1986;15:275–283.
126. Levine E, Huntrakoon M, Wetzel LH. Malignant nerve-sheath neoplasms in neurofibromatosis: distinction from benign tumors by using imaging techniques. AJR Am J Roentgenol. 1987;149:1059–1064.
127. Stener B, Angervall L, Nilsson L, et al. Angiographic and histologic studies of the vascularization of peripheral nerve tumors. Clin Orthop. 1969;66:113–124.
128. Littlewood AH, Stilwell JH. The vascular features of plexiform neurofibroma with some observations on the importance of pre-operative angiography and the value of pre-operative intra-arterial embolisation. Br J Plast Surg. 1983;36:501–506.
129. Patel YD, Morehouse HT. Neurofibrosarcomas in neurofibromatosis: role of CT scanning and angiography. Clin Radiol. 1982;33:555–560.
130. Hammond JA, Driedger AA. Detection of malignant change in neurofibromatosis (von Recklinghausen’s disease) by gallium-67 scanning. Can Med Assoc J. 1978;119:352–353.
131. Kaplan IL, Swayne LC, Baydin JA. Uptake of Ga-67 citrate in a benign neurofibroma. Clin Nucl Med. 1989;14:224.
132. Kobayashi H, Kotoura Y, Sakahara H, et al. Schwannoma of the extremities: comparison of MRI and pentavalent technetium-99m-dimercaptosuccinic acid and gallium-67-citrate scintigraphy. J Nucl Med. 1994;35:1174–1178.
133. Eary JF, Conrad EU. Positron emission tomography in grading soft tissue sarcomas. Semin Musculoskelet Radiol. 1999;3:135–138.
134. Ahmed AR, Watanabe H, Aoki J, et al. Schwannoma of the extremities: the role of PET in preoperative planning. Eur J Nucl Med. 2001;28:1541–1551.
135. Beaulieu S, Rubin B, Djang D, Conrad E, et al. Positron emission tomography of schwannomas: emphasizing its potential in preoperative planning. AJR Am J Roentgenol. 2004;182:971–974.
136. Shah N, Sibtain A, Saunders MI, et al. High FDG uptake in a schwannoma: a PET study. J Comput Assist Tomogr. 2000;24:55–56.
137. Ferner RE, Lucas JD, O’Doherty MJ, et al. Evaluation of (18) fluorodeoxyglucose positron emission tomography ((18) FDG PET) in the detection of malignant peripheral nerve sheath tumours arising from within plexiform neurofibromas in neurofibromatosis 1. J Neurol Neurosurg Psychiatry. 2000;68:353–357.
138. Solomon SB, Semih Dogan A, Nicol TL, et al. Positron emission tomography in the detection and management of sarcomatous transformation in neurofibromatosis. Clin Nucl Med. 2001;26: 525–528.
139. Hems TE, Burge PD, Wilson DJ. The role of magnetic resonance imaging in the management of peripheral nerve tumours. J Hand Surg [Br]. 1997;22:57–60.
140. Kransdorf MJ, Jelinek JS, Moser RP Jr, et al. Soft-tissue masses: diagnosis using MR imaging. AJR Am J Roentgenol. 1989;153: 541–547.
141. Lin J, Martel W. Cross-sectional imaging of peripheral nerve sheath tumors: characteristic signs on CT, MR imaging, and sonography. AJR Am J Roentgenol. 2001;176:75–82.
142. Soderlund V, Goranson H, Bauer HC. MR imaging of benign peripheral nerve sheath tumors. Acta Radiol. 1994;35:282–286.
143. Kilcoyne RF, Richardson ML, Porter BA, et al. Magnetic resonance imaging of soft tissue masses. Clin Orthop. 1988;228:13–19.
144. Sundaram M, McGuire MH, Herbold DR. Magnetic resonance imaging of soft tissue masses: an evaluation of fifty-three histologically proven tumors. Magn Reson Imaging. 1988;6:237–248.
145. Murphey MD, Smith WS, Smith SE, et al. From the archives of the AFIP. Imaging of musculoskeletal neurogenic tumors: radiologic-pathologic correlation. Radiographics. 1999;19:1253–1280.
146. Beaman FD, Kransdorf MJ, Menke DM. Schwannoma: radiologic-pathologic correlation. Radiographics. 2004;24:1477–1481.
147. Hoddick WK, Callen PW, Filly RA, et al. Ultrasound evaluation of benign sciatic nerve sheath tumors. J Ultrasound Med. 1984;3: 505–507.
148. King AD, Ahuja AT, King W, et al. Sonography of peripheral nerve tumors of the neck. AJR Am J Roentgenol. 1997;169: 1695–1698.
149. Cerofolini E, Landi A, DeSantis G, et al. MR of benign peripheral nerve sheath tumors. J Comput Assist Tomogr. 1991;15:593–597.
150. Li MH, Holtas S. MR imaging of spinal neurofibromatosis. Acta Radiol. 1991;32:279–285.
151. Bass JC, Korobkin M, Francis IR, et al. Retroperitoneal plexiform neurofibromas: CT findings. AJR Am J Roentgenol. 1994;163: 617–620.
152. Burk DL Jr, Brunberg JA, Kanal E, et al. Spinal and paraspinal neurofibromatosis: surface coil MR imaging at 1.5 T1. Radiology. 1987;162:797–801.
153. Demachi H, Takashima T, Kadoya M, et al. MR imaging of spinal neurinomas with pathological correlation. J Comput Assist Tomogr. 1990;14:250–254.
154. Hu HP, Huang QL. Signal intensity correlation of MRI with pathological findings in spinal neurinomas. Neuroradiology. 1992;34:98–102.
155. Lewis TT, Kingsley DP. Magnetic resonance imaging of multiple spinal neurofibromata-neurofibromatosis. Neuroradiology. 1987; 29:562–564.
156. Suh JS, Abenoza P, Galloway HR, et al. Peripheral (extracranial) nerve tumors: correlation of MR imaging and histologic findings. Radiology. 1992;183:341–346.
P.377

157. Khong PL, Goh WH, Wong VC, et al. MR imaging of spinal tumors in children with neurofibromatosis 1. AJR Am J Roentgenol. 2003;180:413–417.
158. Inaoka T, Takahashi K, Hanaoka H, et al. Paravertebral neurinoma associated with aggressive intravertebral extension. Skeletal Radiol. 2001;30:286–289.
159. Jee WH, Oh SN, McCauley T, et al. Extraaxial neurofibromas versus neurilemmomas: discrimination with MRI. AJR Am J Roentgenol. 2004;183:629–633.
160. Sullivan TP, Seeger LL, Doberneck SA, et al. Case report 828: plexiform neurofibroma of the tibial nerve invading the medial and lateral gastrocnemius muscles and plantaris muscle. Skeletal Radiol. 1994;23:149–152.
161. Gossios KJ, Guy RL. Case report: imaging of widespread plexiform neurofibromatosis. Clin Radiol. 1993;47:211–213.
162. Stull MA, Kransdorf MJ, Devaney KO. Langerhans cell histiocytosis of bone. Radiographics. 1992;12:801–823.
163. Hallisey MJ, Nelson MC, Muraki A, et al. Case report 547: plexiform neurofibroma involving a dorsal root ganglion. Skeletal Radiol. 1989;18:314–317.
164. Poussaint TY, Jaramillo D, Chang Y, et al. Interobserver reproducibility of volumetric MR imaging measurements of plexiform neurofibromas. AJR Am J Roentgenol. 2003;180:419–423.
165. Katsumi K, Ogose A, Hotta T, et al. Plexiform schwannoma of the forearm. Skeletal Radiol. 2003;32:719–723.
166. Yamamoto T, Maruyama S, Mizuno K. Schwannomatosis of the sciatic nerve. Skeletal Radiol. 2001;30:109–113.
167. Peh WC, Shek TW, Yip DK. Magnetic resonance imaging of subcutaneous diffuse neurofibroma. Br J Radiol. 1997;70:1180–1183.
168. Fortman BJ, Kuszyk BS, Urban BA, et al. Neurofibromatosis type 1: a diagnostic mimicker at CT. Radiographics. 2001;21:601–612.
169. Reinbold WD, Wimmer B, Adler CP, et al. Radiologic findings in peripheral neurilemoma. Eur J Radiol. 1987;7:268–273.
170. Cohen LM, Schwartz AM, Rockoff SD. Benign schwannomas: pathologic basis for CT inhomogeneities. AJR Am J Roentgenol. 1986;147:141–143.
171. Feyerabend T, Schmitt R, Lanz U, et al. CT morphology of benign median nerve tumors. Report of three cases and a review. Acta Radiol. 1990;31:23–25.
172. Kumar AJ, Kuhajda FP, Martinez CR, et al. Computed tomography of extracranial nerve sheath tumors with pathological correlation. J Comput Assist Tomogr. 1983;7:857–865.
173. Pierallini A, Bastianello S, Antonini G, et al. CT findings in peripheral mononeuropathies. Zentralbl Neurochir. 1993;54: 66–71.
174. Thiebot J, Laissy JP, Delangre T, et al. Benign solitary neurinomas of the sciatic popliteal nerves CT study. Neuroradiology. 1991;33:186–188.
175. Zingale A, Consoli V, Tigano G, et al. CT morphology of a median nerve neurilemmoma at the arm. Case report and review. J Neurosurg Sci. 1993;37:57–59.
176. Beggs I. Sonographic appearances of nerve tumors. J Clin Ultrasound. 1999;27:363–368.
177. Reynolds DL Jr, Jacobson JA, Inampudi P, et al. Sonographic characteristics of peripheral nerve sheath tumors. AJR Am J Roentgenol. 2004;182:741–744.
178. Mann FA, Murphy WA, Totty WG, et al. Magnetic resonance imaging of peripheral nerve sheath tumors. Assessment by numerical visual fuzzy cluster analysis. Invest Radiol. 1990;25: 1238–1245.
179. Banks KP. The target sign: extremity. Radiology. 2005;234: 899–900.
180. Bhargava R, Parham DM, Lasater OE, et al. MR imaging differentiation of benign and malignant peripheral nerve sheath tumors: use of the target sign. Pediatr Radiol. 1997;27:124–129.
181. Ogose A, Hotta T, Morita T, et al. Tumors of peripheral nerves: correlation of symptoms, clinical signs, imaging features, and histologic diagnosis. Skeletal Radiol. 1999;28:183–188.
182. Stuart RM, Koh ES, Breidahl WH. Sonography of peripheral nerve pathology. AJR Am J Roentgenol. 2004;182:123–129.
183. Chui MC, Bird BL, Rogers J. Extracranial and extraspinal nerve sheath tumors: computed tomographic evaluation. Neuroradiology. 1988;30:47–53.
184. Damascelli B, Landoni L, Spreafico C. Magnetic resonance imaging of extracranial neurogenic tumors. J Med Imaging. 1987; 1:27–37.
185. Declercq H, De Man R, Van Herck G, et al. Case report 814: fibrolipoma of the median nerve. Skeletal Radiol. 1993;22: 610–613.
186. Glasier CM, Williamson MR, Lange TA. MRI of peripheral neurofibromas in children. Orthopedics. 1989;12:269–272.
187. Nichols J, Tehranzadeh J. Benign neurilemmoma in the posterior soft tissues of the right thigh. Case report 325. Skeletal Radiol. 1985;14:136–140.
188. Petasnick JP, Turner DA, Charters JR, et al. Soft-tissue masses of the locomotor system: comparison of MR imaging with CT. Radiology. 1986;160:125–133.
189. Roberts CS, Fetto JF, Fay CM. Neurilemoma in the distal part of the thigh. A case report. J Bone Joint Surg Am. 1989;71:1082–1083.
190. Santi MD, Mitsunaga MM, Lockett JL. Total sacrectomy for a giant sacral schwannoma. A case report. Clin Orthop. 1993;294: 285–289.
191. Stull MA, Moser RP Jr, Kransdorf MJ, et al. Magnetic resonance appearance of peripheral nerve sheath tumors. Skeletal Radiol. 1991;20:9–14.
192. Varma DG, Moulopoulos A, Sara AS, et al. MR imaging of extracranial nerve sheath tumors. J Comput Assist Tomogr. 1992; 16:448–453.
193. Moon WK, Im JG, Han MC. Malignant schwannomas of the thorax: CT findings. J Comput Assist Tomogr. 1993;17:274–276.
194. Friedman DP, Tartaglino LM, Flanders AE. Intradural schwannomas of the spine: MR findings with emphasis on contrast-enhancement characteristics. AJR Am J Roentgenol. 1992;158: 1347–1350.
195. Gouliamos AD, Kontogiannis DS, Androulidakis J, et al. Spinal neurilemmomas and neurofibromas: central dot sign in postgadolinium MRI. J Comput Assist Tomogr. 1993;17:446–448.
196. Lee JA, Boles CA. Peripheral schwannoma lacking enhancement on MRI. AJR Am J Roentgenol. 2004;182:534–535.
197. Nemeth A, Patel S. Peripheral schwannoma lacking enhancement on MRI [Letter to Editor]. AJR Am J Roentgenol. 2004;182:534–535.
198. Graviet S, Sinclair G, Kajani N. Ancient schwannoma of the foot. J Foot Ankle Surg. 1995;34:46–50.
199. Hide IG, Baudouin CJ, Murray SA, et al. Giant ancient schwannoma of the pelvis. Skeletal Radiol. 2000;29:538–542.
200. Loke TK, Yuen NW, Lo KK, et al. Retroperitoneal ancient schwannoma: review of clinico-radiological features. Australas Radiol. 1998;42:136–138.
201. Isobe K, Shimizu T, Akahane T, et al. Imaging of ancient schwannoma. AJR Am J Roentgenol. 2004;183:331–336.
202. Schultz E, Sapan MR, McHeffey-Atkinson B, et al. Case report 872. “Ancient” schwannoma (degenerated neurilemoma). Skeletal Radiol. 1994;23:593–595.
203. Lazarus SS, Trombetta LD. Ultrastructural identification of a benign perineurial cell tumor. Cancer. 1978;41:1823–1829.
204. Heilbrun ME, Tsuruda JS, Townsend JJ, et al. Intraneural perineurioma of the common peroneal nerve. Case report and review of the literature. J Neurosurg. 2001;94:811–815.
205. Aaron AD, Nelson MC, Azumi N, et al. Case report 820: Granular cell tumor of striated muscle. Skeletal Radiol. 1994;23:63–66.
206. Hurrell MA, McLean C, Desmond P, et al. Malignant granular cell tumour of the sciatic nerve. Australas Radiol. 1995;39: 86–89.
207. Mukherji SK, Castillo M, Rao V, et al. Granular cell tumors of the subglottic region of the larynx: CT and MR findings. AJR Am J Roentgenol. 1995;164:1492–1494.
208. Reuter KL, Raptopoulos V, DeGirolami U, et al. Ultrasonography of a plexiform neurofibroma of the popliteal fossa. J Ultrasound Med. 1982;1:209–211.
209. Shimamura K, Osamura RY, Ueyama Y, et al. Malignant granular cell tumor of the right sciatic nerve. Report of an autopsy case with electron microscopic, immunohistochemical, and enzyme histochemical studies. Cancer. 1984;53:524–529.
210. Fanburg-Smith JC, Meis-Kindblom JM, Fante R, et al. Malignant granular cell tumor of soft tissue: diagnostic criteria and clinicopathologic correlation. Am J Surg Pathol. 1998;22: 779–794.
P.378

211. Chinn DH, Filly RA, Callen PW. Unusual ultrasonographic appearance of a solid schwannoma. J Clin Ultrasound. 1982;10:243–245.
212. Hughes DG, Wilson DJ. Ultrasound appearances of peripheral nerve tumours. Br J Radiol. 1986;59:1041–1043.
213. Kudawara I, Ueda T, Yoshikawa H. Granular cell tumor of the subcutis: CT and MRI findings. A report of three cases. Skeletal Radiol. 1999;28:96–99.
214. Blacksin MF, White LM, Hameed M, et al. Granular cell tumor of the extremity: magnetic resonance imaging characteristics with pathologic correlation. Skeletal Radiol. 2005;34:625–631.
215. Elkousy H, Harrelson J, Dodd L, et al. Granular cell tumors of the extremities. Clin Orthop. 2000;380:191–198.
216. Tsuchida T, Okada K, Itoi E, et al. Intramuscular malignant granular cell tumor. Skeletal Radiol. 1997;26:116–121.
217. Krompecher E. Zur histogenese und morphologie der adamantinome und sonstiger kiefergeschwulste. Pathol Res Pract. 1918;64:169–197.
218. Atkinson GO Jr, Davis PC, Patrick LE, et al. Melanotic neuroectodermal tumor of infancy. MR findings and a review of the literature. Pediatr Radiol. 1989;20:20–22.
219. Cutler LS, Chaudhry AP, Topazian R. Melanotic neuroectodermal tumor of infancy: an ultrastructural study, literature review, and reevaluation. Cancer. 1981;48:257–270.
220. Kozlowski K, Masel J, Sprague P, et al. Mandibular and para-mandibular tumors in children. Report of 16 cases. Pediatr Radiol. 1981;11:183–192.
221. Mirich DR, Blaser SI, Harwood-Nash DC, et al. Melanotic neuroectodermal tumor of infancy: clinical, radiologic, and pathologic findings in five cases. AJNR Am J Neuroradiol. 1991;12: 689–697.
222. Enzinger FM. Clear-cell sarcoma of tendons and aponeuroses. An analysis of 21 cases. Cancer. 1965;18:1163–1174.
223. Chung EB, Enzinger FM. Malignant melanoma of soft parts. A reassessment of clear cell sarcoma. Am J Surg Pathol. 1983;7: 405–413.
224. Eckardt JJ, Pritchard DJ, Soule EH. Clear cell sarcoma. A clinicopathologic study of 27 cases. Cancer. 1983;52:1482–1488.
225. Lucas DR, Nascimento AG, Sim FH. Clear cell sarcoma of soft tissues. Mayo Clinic experience with 35 cases. Am J Surg Pathol. 1992;16:1197–1204.
226. Andrew TA. Clear cell sarcoma of the hand. Hand. 1982;14: 200–203.
227. Limon J, Debiec-Rychter M, Nedoszytko B, et al. Aberrations of chromosome 22 and polysomy of chromosome 8 as non-random changes in clear cell sarcoma. Cancer Genet Cytogenet. 1994;72:141–145.
228. Morishita S, Onomura T, Yamamoto S, et al. Clear cell sarcoma of tendons and aponeuroses (malignant melanoma of soft parts) with unusual roentgenologic findings. Case report. Clin Orthop. 1987;216:276–279.
229. Raynor AC, Vargas-Cortes F, Alexander RW, et al. Clear-cell sarcoma with melanin pigment: a possible soft-tissue variant of malignant melanoma. Case report. J Bone Joint Surg Am. 1979;61:276–280.
230. Sartoris DJ, Haghighi P, Resnick D. Case report 423: clear-cell sarcoma plantar aspect of right foot. Skeletal Radiol. 1987;16: 325–332.
231. Wetzel LH, Levine E. Soft-tissue tumors of the foot: value of MR imaging for specific diagnosis. AJR Am J Roentgenol. 1990;155: 1025–1030.
232. Deenik W, Mooi WJ, Rutgers EJ, et al. Clear cell sarcoma (malignant melanoma) of soft parts: a clinicopathologic study of 30 cases. Cancer. 1999;86:969–975.
233. Graadt van Roggen JF, Mooi WJ, Hogendoorn PC. Clear cell sarcoma of tendons and aponeuroses (malignant melanoma of soft parts) and cutaneous melanoma: exploring the histogenetic relationship between these two clinicopathological entities. J Pathol. 1998;186:3–7.
234. Speleman F, Delattre O, Peter M, et al. Malignant melanoma of the soft parts (clear-cell sarcoma): confirmation of EWS and ATF-1 gene fusion caused by a t(12;22) translocation. Mod Pathol. 1997;10:496–499.
235. De Beuckeleer LH, De Schepper AM, Vandevenne JE, et al. MR imaging of clear cell sarcoma (malignant melanoma of the soft parts): a multicenter correlative MRI-pathology study of 21 cases and literature review. Skeletal Radiol. 2000;29:187–195.
236. Gomori JM, Grossman RI, Shields JA, et al. Choroidal melanomas: correlation of NMR spectroscopy and MR imaging. Radiology. 1986;158:443–445.
237. Schnarkowski P, Peterfy CG, Johnston JO, et al. Clear cell sarcoma mimicking peripheral nerve sheath tumor. Skeletal Radiol. 1996;25:197–200.
238. Bishop GB Jr, Urist MM, el Gammal T, et al. Paragangliomas of the neck. Arch Surg. 1992;127:1441–1445.
239. Rao AB, Koeller KK, Adair CF. From the archives of the AFIP. Paragangliomas of the head and neck: radiologic-pathologic correlation. Armed Forces Institute of Pathology. Radiographics. 1999;19:1605–1632.
240. Weiss S, Goldblum J. Paraganglioma. In: Weiss SW, Goldblum JR, eds. Enzinger and Weiss’s Soft Tissue Tumors. 4th ed. St. Louis: Mosby; 2001:1323–1360.
241. Enzinger F, Weiss SW. Paraganglioma. In: Weiss SW, Goldblum JR, eds. Soft Tissue Tumors. 3rd ed. St. Louis: Mosby; 1995:965–990.
242. van Gils AP, Falke TH, van Erkel AR, et al. MR imaging and MIBG scintigraphy of pheochromocytomas and extraadrenal functioning paragangliomas. Radiographics. 1991;11:37–57.
243. Williams E, Siebenmann R, Sobin L. Histological typing of endocrine tumors. In: Fletcher CDM, Unni KK, Mertens F, eds. WHO International Histological Classification of Tumors. Geneva: World Health Organization; 1980:33–39.
244. Leonetti JP, Donzelli JJ, Littooy FN, et al. Perioperative strategies in the management of carotid body tumors. Otolaryngol Head Neck Surg. 1997;117:111–115.
245. Duncan AW, Lack EE, Deck MF. Radiological evaluation of paragangliomas of the head and neck. Radiology. 1979;132: 99–105.
246. Olsen WL, Dillon WP, Kelly WM, et al. MR imaging of paragangliomas. AJR Am J Roentgenol. 1987;148:201–204.
247. Kawai A, Healey JH, Wilson SC, et al. Carotid body paraganglioma metastatic to bone: report of two cases. Skeletal Radiol. 1998;27:103–107.
248. Grimley P. Multicentric paragangliomas and associated neuroendocrine tumors. In: Brunson K, ed. Local Invasion and Spread of Cancer. Dordrecht, The Netherlands: Kluwer Academic; 1989:49–61.
249. Zaslav AL, Myssiorek D, Mucia C, et al. Cytogenetic analysis of tissues from patients with familial paragangliomas of the head and neck. Head Neck. 1995;17:102–107.
250. Bosniak MA, Seidenberg B, Rubin IC, et al. Angiographic demonstration of bilateral carotid body tumors. Am J Roentgenol Radium Ther Nucl Med. 1964;92:850–854.
251. Farr HW. Carotid body tumors. A thirty year experience at Memorial Hospital. Am J Surg. 1967;114:614–619.
252. Jacobs JB, Chretien PB, Sugarbaker E, et al. Arteriographic discovery of a small contralateral carotid body tumor. Radiology. 1969;93:837–838.
253. Lack EE, Cubilla AL, Woodruff JM, et al. Paragangliomas of the head and neck region: a clinical study of 69 patients. Cancer. 1977;39:397–409.
254. Oberman HA, Holtz F, Sheffer LA, et al. Chemodectomas (nonchromaffin paragangliomas) of the head and neck. A clinicopathologic study. Cancer. 1968;21:838–851.
255. Rice RP, Holman CB. Roentgenographic manifestations of tumors of the glomus jugulare (chemodectoma). Am J Roentgenol Radium Ther Nucl Med. 1963;89:1201–1208.
256. Shugar MA, Mafee MF. Diagnosis of carotid body tumors by dynamic computerized tomography. Head Neck Surg. 1982;4: 518–521.
257. Weissman JL. Case 21: glomus vagale tumor. Radiology. 2000; 215:237–242.
258. Falke TH, van Gils AP, van Seters AP, et al. Magnetic resonance imaging of functioning paragangliomas. Magn Reson Q. 1990;6:35–64.
259. Schmedtje JF Jr, Sax S, Pool JL, et al. Localization of ectopic pheochromocytomas by magnetic resonance imaging. Am J Med. 1987;83:770–772.
260. Velchik MG, Alavi A, Kressel HY, et al. Localization of pheochromocytoma: MIBG [correction of MIGB], CT, and MRI correlation. J Nucl Med. 1989;30:328–336.
P.379

261. Derchi LE, Serafini G, Rabbia C, et al. Carotid body tumors: US evaluation. Radiology. 1992;182:457–459.
262. Jansen JC, Baatenburg de Jong RJ, Schipper J, et al. Color Doppler imaging of paragangliomas in the neck. J Clin Ultrasound. 1997;25:481–485.
263. Makarainen H, Paivansalo M, Hyrynkangas K, et al. Sonographic patterns of carotid body tumors. J Clin Ultrasound. 1986;14: 373–375.
264. Mukherji SK, Kasper ME, Tart RP, et al. Irradiated paragangliomas of the head and neck: CT and MR appearance. AJNR Am J Neuroradiol. 1994;15:357–363.
265. Som PM, Braun IF, Shapiro MD, et al. Tumors of the parapharyngeal space and upper neck: MR imaging characteristics. Radiology. 1987;164:823–829.
266. van den Berg R, van Gils AP, Wasser MN. Imaging of head and neck paragangliomas with three-dimensional time-of-flight MR angiography. AJR Am J Roentgenol. 1999;172:1667–1673.
267. Vogl T, Bruning R, Schedel H, et al. Paragangliomas of the jugular bulb and carotid body: MR imaging with short sequences and Gd-DTPA enhancement. AJR Am J Roentgenol. 1989;153: 583–587.
268. Vogl TJ, Juergens M, Balzer JO, et al. Glomus tumors of the skull base: combined use of MR angiography and spin-echo imaging. Radiology. 1994;192:103–110.
269. Idbohrn H. Angiographical diagnosis of carotid body tumours. Acta Radiol. 1951;35:115–123.
270. Chatal JF, Charbonnel B. Comparison of iodobenzylguanidine imaging with computed tomography in locating pheochromocytoma. J Clin Endocrinol Metab. 1985;61:769–772.
271. Quint LE, Glazer GM, Francis IR, et al. Pheochromocytoma and paraganglioma: comparison of MR imaging with CT and I-131 MIBG scintigraphy. Radiology. 1987;165:89–93.
272. van Gils AP, van der Mey AG, Hoogma RP, et al. Iodine-123-metaiodobenzylguanidine scintigraphy in patients with chemodectomas of the head and neck region. J Nucl Med. 1990; 31:1147–1155.
273. Welch TJ, Sheedy PF II, van Heerden JA, et al. Pheochromocytoma: value of computed tomography. Radiology. 1983;148:501–503.
274. Krenning EP, Kwekkeboom DJ, Bakker WH, et al. Somatostatin receptor scintigraphy with [111In-DTPA-D-Phe1]- and [123I-Tyr3]-octreotide: the Rotterdam experience with more than 1000 patients. Eur J Nucl Med. 1993;20: 716–731.
275. Kwekkeboom DJ, van Urk H, Pauw BK, et al. Octreotide scintigraphy for the detection of paragangliomas. J Nucl Med. 1993; 34:873–878.
276. Myssiorek D, Palestro CJ. 111 Indium pentetreotide scan detection of familial paragangliomas. Laryngoscope. 1998;108:228–231.
277. Enzinger F, Weiss SW. Primitive neuroectodermal tumors and related lesions. In: Weiss SW, Goldblum JR, eds. Soft Tissue Tumors. 3rd ed. St. Louis: Mosby; 1995:929–964.
278. Stout A. Ganglioneuroma of the sympathetic nervous system. Surg Gynecol Obstet. 1947;84:101–110.
279. Miettinen M. Neuroectodermal and neural tumors. In: Diagnostic Soft Tissue Pathology. New York: Churchill Livingstone; 2003:379–401.
280. Amundson GM, Trevenen CL, Mueller DL, et al. Neuroblastoma: a specific sonographic tissue pattern. AJR Am J Roentgenol. 1987;148:943–945.
281. David R, Lamki N, Fan S, et al. The many faces of neuroblastoma. Radiographics. 1989;9:859–882.
282. Dietrich RB, Kangarloo H, Lenarsky C, et al. Neuroblastoma: the role of MR imaging. AJR Am J Roentgenol. 1987;148:937–942.
283. Evans AE. Staging and treatment of neuroblastoma. Cancer. 1980;45:1799–1802.
284. Stark DD, Moss AA, Brasch RC, et al. Neuroblastoma: diagnostic imaging and staging. Radiology. 1983;148:101–105.
285. Armstrong EA, Harwood-Nash DC, Ritz CR, et al. CT of neuro-blastomas and ganglioneuromas in children. AJR Am J Roentgenol. 1982;139:571–576.
286. Page D, DeLellis R, Hough A. Tumors of the adrenal. In: Atlas of Tumor Pathology, 2nd Series, Fascicle 23. Washington, DC: Armed Forces Institute of Pathology; 1985.
287. Rha SE, Byun JY, Jung SE, et al. Neurogenic tumors in the abdomen: tumor types and imaging characteristics. Radiographics. 2003;23:29–43.
288. Ichikawa T, Koyama A, Fujimoto H, et al. Retroperitoneal ganglioneuroma extending across the midline: MR features. Clin Imaging. 1993;17:19–21.
289. Ichikawa T, Ohtomo K, Araki T, et al. Ganglioneuroma: computed tomography and magnetic resonance features. Br J Radiol. 1996;69:114–121.
290. Zhang Y, Nishimura H, Kato S, et al. MRI of ganglioneuroma: histologic correlation study. J Comput Assist Tomogr. 2001;25:617–623.
291. Jasinski RW, Samuels BI, Silver TM. Sonographic features of retroperitoneal ganglioneuroma. J Ultrasound Med. 1984;3: 413–415.
292. White SJ, Stuck KJ, Blane CE, et al. Sonography of neuroblastoma. AJR Am J Roentgenol. 1983;141:465–468.
293. Otal P, Mezghani S, Hassissene S, et al. Imaging of retroperitoneal ganglioneuroma. Eur Radiol. 2001;11:940–945.
294. Radin R, David CL, Goldfarb H, et al. Adrenal and extra-adrenal retroperitoneal ganglioneuroma: imaging findings in 13 adults. Radiology. 1997;202:703–707.
295. Harbert JC, Robertson JS, Held KD. Meta (131) iodobenzylguanidine therapy of malignant pheochromocytomas and other neuroendocrine lesions. In: Nuclear Medicine Therapy. New York: Thieme; 1987:99–108.
296. Munkner T. 131I-meta-iodobenzylguanidine scintigraphy of neuroblastomas. Semin Nucl Med. 1985;15:154–160.
297. Weiss S, Goldblum J. Primitive neuroectodermal tumors and related lesions. In: Weiss SW, Goldblum JR, eds. Enzinger and Weiss’s Soft Tissue Tumors. 4th ed. St. Louis: Mosby; 2001: 1265–1321.
298. Ewing J. Diffuse endothelioma of bone. Proc NY Soc Pathol. 1921;21:17–24.
299. Stout A. A tumor of the ulnar nerve. Proc NY Soc Pathol. 1918;12:2–12.
300. Parham DM, Hijazi Y, Steinberg SM, et al. Neuroectodermal differentiation in Ewing’s sarcoma family of tumors does not predict tumor behavior. Hum Pathol. 1999;30:911–918.
301. Dehner LP. Primitive neuroectodermal tumor and Ewing’s sarcoma. Am J Surg Pathol. 1993;17:1–13.
302. Cavazzana AO, Ninfo V, Roberts J, et al. Peripheral neuroepithelioma: a light microscopic, immunocytochemical, and ultrastructural study. Mod Pathol. 1992;5:71–78.
303. Gillespie JJ, Roth LM, Wills ER, et al. Extraskeletal Ewing’s sarcoma. Histologic and ultrastructural observations in three cases. Am J Surg Pathol. 1979;3:99–108.
304. Hashimoto H, Enjoji M, Nakajima T, et al. Malignant neuroepithelioma (peripheral neuroblastoma). A clinicopathologic study of 15 cases. Am J Surg Pathol. 1983;7:309–318.
305. Kushner BH, Hajdu SI, Gulati SC, et al. Extracranial primitive neuroectodermal tumors. The Memorial Sloan-Kettering Cancer Center experience. Cancer. 1991;67:1825–1829.
306. Shimada H, Newton WA Jr, Soule EH, et al. Pathologic features of extraosseous Ewing’s sarcoma: a report from the Intergroup Rhabdomyosarcoma Study. Hum Pathol. 1988;19: 442–453.
307. Turc-Carel C, Aurias A, Mugneret F, et al. Chromosomes in Ewing’s sarcoma. I. An evaluation of 85 cases of remarkable consistency of t(11;22)(q24;q12). Cancer Genet Cytogenet. 1988;32: 229–238.
308. Meister P, Gokel JM. Extraskeletal Ewing’s sarcoma. Virchows Arch. 1978;378:173–179.
309. Hoffer FA, Gianturco LE, Fletcher JA, et al. Percutaneous biopsy of peripheral primitive neuroectodermal tumors and Ewing’s sarcomas for cytogenetic analysis. AJR Am J Roentgenol. 1994; 162:1141–1142.
310. Marina NM, Etcubanas E, Parham DM, et al. Peripheral primitive neuroectodermal tumor (peripheral neuroepithelioma) in children. A review of the St. Jude experience and controversies in diagnosis and management. Cancer. 1989;64: 1952–1960.
311. Schmidt D, Herrmann C, Jurgens H, Harms D. Malignant peripheral neuroectodermal tumor and its necessary distinction from Ewing’s sarcoma. A report from the Kiel Pediatric Tumor Registry. Cancer. 1991;68:2251–2259.
P.380

312. Jurgens H, Exner U, Gadner H, et al. Multidisciplinary treatment of primary Ewing’s sarcoma of bone. A 6-year experience of a European Cooperative Trial. Cancer. 1988;61:23–32.
313. Allam K, Sze G. MR of primary extraosseous Ewing sarcoma. AJNR Am J Neuroradiol. 1994;15:305–307.
314. O’Keeffe F, Lorigan JG, Wallace S. Radiological features of extraskeletal Ewing sarcoma. Br J Radiol. 1990;63:456–460.
315. Rose JS, Hermann G, Mendelson DS, et al. Extraskeletal Ewing sarcoma with computed tomography correlation. Skeletal Radiol. 1983;9:234–237.
316. Askin FB, Rosai J, Sibley RK, et al. Malignant small cell tumor of the thoracopulmonary region in childhood: a distinctive clinicopathologic entity of uncertain histogenesis. Cancer. 1979;43: 2438–2451.
317. Faubert C, Inniger R. MRI and pathological findings in two cases of Askin tumors. Neuroradiology. 1991;33:277–281.
318. Fink IJ, Kurtz DW, Cazenave L, et al. Malignant thoracopulmonary small-cell (“Askin”) tumor. AJR Am J Roentgenol. 1985; 145:517–520.
319. Saifuddin A, Robertson RJ, Smith SE. The radiology of Askin tumours. Clin Radiol. 1991;43:19–23.
320. Winer-Muram HT, Kauffman WM, Gronemeyer SA, et al. Primitive neuroectodermal tumors of the chest wall (Askin tumors): CT and MR findings. AJR Am J Roentgenol. 1993;161: 265–268.
321. Burge HJ, Novotny DB, Schiebler ML, et al. MRI of Askin’s tumor. Case report at 1.5 T. Chest. 1990;97:1252–1254.