We reviewed 234 benign solitary schwannomas treated between 1984 and 2004. The mean age of the patients was 45.2 years (11 to 82). There were 170 tumours (73%) in the upper limb, of which 94 (40%) arose from the brachial plexus or other nerves within the posterior triangle of the neck. Six (2.6%) were located within muscle or bone. Four patients (1.7%) presented with tetraparesis due to an intraspinal extension.
There were 198 primary referrals (19 of whom had a needle biopsy in the referring unit) and in these patients the tumour was excised. After having surgery or an open biopsy at another hospital, a further 36 patients were seen because of increased neurological deficit, pain or incomplete excision. In these, a nerve repair was performed in 18 and treatment for pain or paralysis was offered to another 14.
A tender mass was found in 194 (98%) of the primary referrals. A Tinel-like sign was recorded in 155 (81%). Persistent spontaneous pain occurred in 60 (31%) of the 194 with tender mass, impairment of cutaneous sensibility in 39 (20%), and muscle weakness in 24 (12%).
After apparently adequate excision, two tumours recurred. No case of malignant transformation was seen.
Benign schwannoma is the most common tumour of peripheral nerves.1,2 Malignant transformation is extremely rare.3 The tumours are composed of Schwann cells which support the peripheral nerve fibres and are neuroectodermal in origin. Their histological features have been clarified by transmission electron microscopy and by immunohistochemical staining which shows strong expression for the S-100 protein.4,5 Folpe and Gowan6 have stated that immunostaining for this protein is so consistent and of such intensity that it serves as an important diagnostic tool. Schwannomas may arise from Schwann cells investing the third to the twelfth cranial nerves close to their origin from the brain stem. Bilateral vestibular (eighth nerve) schwannomas are pathognomonic of neurofibromatosis type II and schwannomas are common in neurofibromatosis type I. Most, be they sporadic or an expression of neurofibromatosis type II, show deletion of the neurofibromatosis type II gene on chromosome 22.7,8 A schwannoma is enveloped by a true capsule which consists of the perineurium of the nerve bundle of origin, surrounded by a condensation of the deepest layers of the epineurium. The capsule is often covered by tortuous blood vessels and allows excision of the tumour without damage to the parent nerve. The rare plexiform variant of a schwannoma may infiltrate between nerve bundles and thus make excision difficult.9
Schwannomas characteristically form an eccentric oval swelling, less than 3 cm in diameter, with the attenuated nerve bundles (fascicles) of the parent nerve stretched and displaced over the dome of the mass. They grow slowly, but in sites such as the mediastinum and retroperitoneum, may attain a large size before presentation and their true nature may not be recognised.10,11 Such lesions often display extensive degenerative features including the formation of a cyst, fibrosis and calcification. Small tumours are usually uninodular but larger tumours may be multi-nodular. Occasionally, ‘dumbbell’ tumours encompass two aspects of a bony foramen.
The tumour contains varying proportions of two distinctive tissues, Antoni-A and Antoni-B. The former comprises compact spindle cells with indistinct cell boundaries and buckled or wavy nuclei. The cells are arranged in sheets or bundles and mitoses are sparse but normal. There are areas where nuclei are arranged in parallel rows termed ‘nuclear palisading’. Verocay bodies, in which two rows of palisading nuclei are separated by pink fibrillary material, are common. Antoni-B tissue is less cellular and lacks distinctive architectural features. The matrix is more delicate and contains sparse collagen, numerous blood vessels and a mixture of spindle and oval cells. Foci of inflammatory cells including histiocytes may be present.
Degenerative changes are frequent. The walls of blood vessels become hyalinised with associated haemorrhage and deposition of fibrin. There is formation of cysts, hyalinisation of the matrix and focal calcification. An inflammatory infiltrate is usually present including numerous histiocytes, some of which contain haemosiderin. The nuclei of the Schwann cells become hyperchromatic, enlarged and multilobed, but mitoses remain sparse. Although such nuclear atypia may cause concern, the association with other degenerative features and the lack of mitoses indicate a benign tumour.
The cellular schwannoma is composed almost entirely of Antoni-A tissue. The spindle cells are closely packed and may be arranged in a herringbone pattern. Nuclear palisading and Verocay bodies are absent. The mitotic rate is low. Staining with S-100 protein is uniformally positive. In contrast, the malignant peripheral nerve-sheath tumours shows scattered positive cells or is only focally positive.
The infrequency of schwannomas in general clinical practice causes difficulties in diagnosis and delay before recognition.12 Unnecessary resection of the parent nerve or nerves is a serious complication.9,13
Patients and Methods
We have reviewed 234 benign solitary schwannomas treated between 1984 and 2004. The mean age of the patients was 45.2 years (11 to 82) with 80% (188) aged between 30 and 69 years. They were followed up for between one and 20 years. Most were discharged after one year if free from symptoms.
Of the 234 cases, 198 were primary referrals, 19 of whom had undergone a needle biopsy at the referring unit, and a further 36 secondary referrals who had undergone previous surgery or open biopsy at another hospital.
Data were held in tumour files recording both clinical and operative findings with radiological and histopathological features. The records made at the first attendance included age, gender, location and the characteristics of the tumour. Of particular interest were the presence of a Tinel-like sign,9 spontaneous pain and disturbance of sensation or power within the distribution of the parent nerve. The patients were carefully examined for other, unnoted tumours and were questioned about any possible familial history of neurofibromatosis.
Tissue removed in other hospitals was re-examined by a histopathologist (JP) with a special interest in nerve tumours. Ancillary investigations included plain radiography and MRI. Imaging of major blood vessels was performed in the massive, more centrally-located tumours by digital subtraction angiography or magnetic resonance angiography. CT with contrast enhancement was carried out in those patients in whom the tumour had extended into the spinal canal. In three with massive mediastinal tumours, it was necessary to perform a pre-operative CT-guided or ultrasound-guided trucut needle biopsy because there was uncertainty about the nature of the lesion.
The operations were performed under general anaesthesia. Exposure was planned to allow display of the parent nerve above and below the swelling and also of the adjacent axial structures. Nerve conduction was used to define conducting bundles displaced by the tumour and also to demonstrate conduction through the nerve bundle of origin (Fig. 1⇓).
Tumours arising within the brachial plexus were exposed by a transverse supraclavicular incision. This was extended to allow display of the accessory, the supraclavicular, the phrenic and the spinal nerves as well as their trunks, and divisions of the brachial plexus. Anterior scalenotomy improved exposure of C7, C8 and T1 and the adjacent subclavian artery and vein. The transclavicular exposure14 was used in 13 cases of schwannoma arising in the superior mediastinum, and the posterior subscapular approach in two lesions in the posterior mediastinum. In three of the four cases of intra-extra spinal tumour, the intraspinal component was removed at the first operation through a hemilaminectomy. The residual tumour was excised through an anterolateral approach at a second operation. In the fourth case corpectomy of C4, C5 and C6 was necessary. All histopathological material was reviewed by the surgeon (RB) and pathologist (JP).
The duration of symptoms before presentation was a mean of 35 months (one month to 30 years); 168 tumours (72%) presented within 36 months of the onset of symptoms.
The distribution of all 234 tumours is given in Table I⇓. Table II⇓ gives the clinical signs and symptoms in 191 of the 198 primary referrals in which the information was recorded. The records for seven patients were incomplete.
The maximum diameter of the tumours ranged between 2 mm, arising from a digital nerve, and 120 mm from a massive mediastinal tumour (mean 34 mm).
Findings and outcome in the 198 primary referrals.
Plexiform infiltration between the bundles of the nerve was shown in three tumours. In two of these, excision was incomplete, but the residual tumour appeared to regress.
Significant complications occurred in five patients. In the first, a 55-year-old woman, the schwannoma arose centrally within the ulnar nerve and removal was possible only by excision of three bundles. There was impairment of cutaneous sensation, but no demonstrable loss of power and no pain. In the second, a 56-year-old woman, a haematoma developed within the median nerve in the arm after excision of a schwannoma measuring 6 cm at its greatest diameter. A dense nerve palsy developed at about 24 hours after operation. The nerve was re-explored after 48 hours and the haematoma removed. Pain was relieved and function recovered to nearly normal levels over the subsequent 24 months. In the third patient, a 24-year-old man, the schwannoma involved the lower and middle trunks of the brachial plexus in an infiltrative manner. Exposure was by the transclavicular route. There was a laceration of the first part of the subclavian artery which was controlled, but repair of this vessel required division of the left brachiocephalic vein. This was later repaired after completion of successful suturing of the artery. In the fourth patient with a recurrence of a schwannoma of C6, severe neuropathic pain was not improved by resection of the nerve or by subsequent graft repair. Figure 2⇓ demonstrates the difficulties encountered in the fifth case following recurrence of a schwannoma arising from C6. There were no other examples of recurrence apart from these two cases.
Pain, sensory symptoms and weakness were abolished in the remaining 193 patients and no added neurological deficit occurred. Most tumours showed the characteristic formation of Antoni-A and Antoni-B tissue with strong S-100 positivity (Fig. 3⇓). Ancient change was common in the larger tumours (Fig. 4⇓). There were three cellular schwannomas. Another showed Antoni-B tissue only and one contained melanin pigment. A small number showed an area of epithelioid Schwann cells within a conventional schwannoma (Figs 5⇓ to 8⇓⇓⇓).
Findings and outcome in the 36 secondary referrals following primary surgery or open biopsy.
In 28 of these patients there was significant neurological deficit and/or severe neuropathic pain after the first operation. The main nerve(s) had been excised or partially lacerated in 25. The damaged nerves were re-explored in 22 patients and repairs were performed in 18. Pain was significantly relieved in 17 of these 22 patients, musculotendinous transfer was performed in three because nerve repair was impracticable, and in two more because recovery after repair was poor. Useful recovery of function was achieved in 14. Neuropathic pain, in five of six patients eased with a combination of physiotherapy and drug treatment without recourse to further operation. In the remaining eight, some but not all of the tumour was excised at the first operation without inflicting any significant lesion upon conducting tissues.
Biopsies were performed in 53 patients at the hospital of first presentation. Our information about the methods used was incomplete, but fine-needle biopsy was the most common method. Eight biopsies failed to yield diagnostic material. Significant amounts of peripheral-nerve tissue were seen adjacent to lesional material in six open biopsies and in four core needle procedures. Misinterpretation of two core needle biopsies led to erroneous diagnosis of a soft-tissue sarcoma. This was followed by excision of the upper trunk of the brachial plexus in one patient. In the second, wide resection of L4/5 and partial excision of S1 were followed by chemo- and radiotherapy. The biopsy tissue was not subjected to immunohistochemical examination in either of these cases. Neuropathic pain was a frequent complication of biopsy and added to the difficulty of excision of the tumour because of fibrosis and obliteration of tissue planes within the nerve trunk.
We have encountered errors in the treatment of schwannomas and suggest that most of these arose because the diagnosis was not considered. Accurate diagnosis is essential in the treatment of any tumour. Sondack and Chang15 state that a properly performedbiopsyis the critical first step in a multimodality approach to treatment since if it is done improperly it can complicate the care of the patient and sometimes even eliminate options for treatment. In expert hands, needle biopsy has impressive accuracy.16,17 However, Geisinger and Abdul-Karim18 have set out the limitations of fine-needle aspiration biopsy of soft-tissue tumours in an informative manner. They state that its use for soft-tissue neoplasms has important limitations which must be recognised, and that the sample may not be representative of the tumour and be insufficient to make a diagnosis.
The diagnosis of a schwannoma arising within a nerve trunk of the upper or lower limb is usually straightforward. There is a swelling which is painful to pressure and is mobile from side to side, but not in the vertical axis of the limb. Percussion induces painful paraesthesiae in the area of the nerve of origin similar to Tinel’s sign. We suggest that this finding is the single most useful sign in the diagnosis of a schwannoma. Progressive neurological deficit only occurs when the tumour arises in a confined space. A mass arising from a nerve trunk which causes pain and which is attended by deepening loss of function is a malignant tumour until proven otherwise.
We have found that MRI is particularly helpful.19 The schwannoma is seen as an encapsulated structure eccentric to the axis to the nerve, displacing bundles (fascicles) to one side. We think that the heterogenous appearance is consistent with the mixed nature of the tissue within the schwannoma and also with the development of cysts. However, Moser and Parrish20 emphasise that in most cases radiological features cannot reliably distinguish between benign and malignant soft-tissue tumours. This important caveat cannot be over emphasised and they go on to state that it is imperative that a final interpretation of CT or MR findings is not given until the cross-sectional imaging has been correlated with plain radiography. Failure to adhere to this principle results in serious mistakes in management. Filler et al21 defined magnetic resonance neurography as “tissue-selected imaging directed at identifying and evaluating characteristics of nerve morphology, internal fascicular pattern, longitudinal variations in signal intensity and calibre, and connections and relations to other nerves or plexuses”. This method seems to be particularly informative for the diagnosis and accurate measurement of the extent of tumours in or near nerves, and may improve the accuracy and safety of needle biopsy. Refinements in ultrasonography offer possibilities towards more accurate diagnosis. Martinoli et al22 demonstrated a distinction between extrinsic and intrinsic lesions, and between tumours and neuromas. Chan23 used ultrasound to define intraneural pathology in an impressive manner.
We suggest that the clinical diagnosis for schwannomas arising from nerve trunks within the limbs is straightforward and that the proper treatment is excision with the object of removing the tumour while preserving nerve function. Biopsy is essential in cases in which there is any doubt about diagnosis. We think that such biopsy is best performed within the institution where definitive treatment will take place. It is for the responsible surgeon to decide whether biopsy is needed and which method is used. It is the responsibility of that surgeon to examine the biopsied tissue with an experienced pathologist. Confirmation of the diagnosis is the first step in definitive treatment.
G. Bonney provided us with details of his cases, and commented in an extensive and helpful manner on earlier drafts. Mrs M. Taggart and Mrs D. Shah were responsible for collation of information and the maintenance of the tumour files. Photographs were prepared by Mr D. De Camp, from the audiovisual department at the RNOH. Mrs B. Patel of the PNI Unit prepared the manuscript. Some of the cases described were complex and we are grateful for assistance given by Mr S. Edmondson, Mr J. Lehovsky and Dr A. Valentine.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
- Received May 15, 2006.
- Accepted November 1, 2006.
- © 2007 British Editorial Society of Bone and Joint Surgery