Clinical Otoluryngology 1979, 4, 271-281

Berineural spread in squamous cell carcinomas of the head and neck: a clinicopathological study R . L . CARTER, N. S . B. TANNER, P. CLIFFORD AND H . J . SHAW Head and Neck Unit and Department of Pathology, Royal Marsden Hospital and Institute of Cancer Research, London and Sutton, and King's College Hospital, London

Accepted for publication 5 M a y 1978 CARTER R.L., TANNER N.S.B., CLIFFORD P. 8r SHAW H.J. (1979) Clinical Otolaryngology 4, 271-281 Perineural spread i n squamous cell carcinomas of the head and neck: a clinicopathological study Perineural spread of tumour has been demonstrated in 24% of an unselected series of 70 patients with squamous carcinomas of the head and neck treated by surgery. Slightly more than half the patients had primary tumours arising within the buccal cavity. Clinical features suggesting perineural infiltration were found in about two thirds of the cases; the symptoms and signs were usually sensory and occurred early in the disease. The mandibular division of the trigeminal nerve was most commonly affected. Perineural spread was more likely to occur with large carcinomas, moderate or poorly ditferentiated, showing local invasion and lymph node metastases. The associated pathological changes are described and a high incidence of damage to nerve fibres is recorded. The practical implications of perineural spread of tumour are discussed with reference to indications for more radical surgery or for more conservative measures supplemented with other modes of treatment. Keywords sgiramous carcinoma head and neck perineural spread

Extension along perineural spaces occurs in tumours of various histological types at several anatomical sites. 1-4 One important group of neoplasms which spread by this route consists of cancers of the head and neck. Though commented on by a few writer^,^ perineural spread of tumours in this region remained poorly documented until 1965 when Ballantyne and his colleagues6 published a retrospective account of 80 examples, nearly 70% of which were squamous cell carcinomas. The distribution of cases in this important series was unusual in that carcinoma at 2 sites (skin and lip) comprised over half the total. Tumours of the tongue, larynx and pharynx appeared to be markedly under-represented, though perineural spread of these lesions has been noted el~ewhere.~-'O T h e literature on perineural spread in head and neck carcinomas is still sparse and clear information on the incidence of perineural infiltration by tumours at various sites, the frequency with which it can be detected clinically, the associated pathological cha anges,nd * Address for correspondence: Dr R.L. Carter, Dept of Pathology, Royal Marsden Hospital and Inst. of Cancer Research, Sutton, Surrey SM2 5PX. 0307-7772/79/0800-027 I

$0~.00

0 1979 Blackwell Scientific Publications

271

272

R . I,. C A R T E R et

al.

the clinical implications is not available. These topics form the basis of a current investigation and, in this paper, we present initial findings on perineural spread in an unselected group of 70 patients.

Clinical material The patients, all with histologically proven primary squamous cell carcinomas of the head and neck, were drawn from the Royal Marsden Hospital and King's College Hospital, London. Standard methods of clinical investigation, T N M classification' and management were used. Formalin-fixed surgical specimens were processed and examined in the normal way. Additional sections were taken from material showing perineural spread and stained with methazole fast Blue to demonstrate myelin and with the Glees-Marsland silver impregnation technique for axons. T h e assessment of clinical and pathological evidence of perineural involvement in each case was, initially, made independently by the surgeons and the pathologist. Results Seventy patients with squamous cell carcinoma were treated surgically over a period of 18 months. Perineural spread was demonstrated in a group of 17 cases (24%) which form the basis for this paper. A further 8 patients had suggestive symptoms and signs but perineural spread of tumour was not found in the operation specimens; these cases are not considered further here but it is possible that definite evidence of perineural infiltration will emerge during follow-up. The main subgroup of 17 cases comprised 14 males and 3 females, aged between 28 and 79 years (mean 60). All of them received previous treatment, either radiotherapy alone (8) or combined radiotherapy and chemotherapy (9). Table

I Distribution of tumours in 70 cases of squamous carcinoma of the head and neck, showing the incidence of perineural spread

Site of primary tumour Lip, cheek Retromolar trigone, alveolus Tongue (anterior two-thirds), floor of mouth Tongue (posterior one-third), tonsil, soft palate Pyriform fossa, oesophagus (upper one-third) Larynx Nose, paranasal sinuses Nasopharynx Unknown primary Totals

Total no.

3 9 23 8

6 16

3 I I

Tumours shoming perineural spread : in cervical in primary node metastases 0

3 6 2

0

Perineural spread of tuinoirrs

273

The sites of the tumours in the entire series of 70 cases, and in the subgroup of 17 cases with proven perineural infiltration, are shown in Table I. Slightly more than half the patients with perineural spread had carcinomas arising in the buccal cavity, particularly from the tongue and the floor of the mouth. Perineural spread in the vicinity of metastatic disease in cervical lymph nodes was found in only 3 cases. T h e incidence of clinically-detected perineural involvement at each site is shown in Table 2. About two thirds of the cases had clinical features suggesting perineural infiltration. Table z Sites of 17 tumours showing perineural spread; incidence of clinical detection

Site of tumour Tongue, floor of mouth Alveolus Retromolar trigone Soft palate, tonsil Larynx, transglottic Nose Upper deep cervical lymph nodes Totals

Nerve involvemeiit : detected clinically not detected clinically

5

0

2 I I I

0 0 I 2 0

I

21-1

1%

5

I2

* Metastasis from carcinoma of oropharynx. -1 Metastasis from carcinoma of tongue.

1Metastasis from carcinoma of larynx.

Table 3 Sites of tumours showing clinically detectable perineural spread; nerves involved

Site of tumour Nose ( I ) Retromolar trigone ( I ) Tongue, floor of mouth, alveolus (7) Soft palate + tonsil, ( I ) Larynx ( I ) Upper deep cervical lymph nodes (I)"

Nerves involved V maxillary V ophthalmic VII VI V maxillary + V mandibular V mandibular, XI1

+

+

+

IX

x

X + cervical nerves

" Metastasis from carcinoma of oropharynx. Numbers of turnours at each site shown in parentheses. The main nerves involved are summarized in Table 3. T h e trigeminal nerve, particularly the mandibular division, was most commonly affected. Progressive involvement of cranial nerves was seen in one patient with centripetally spreading tumour which successively infiltrated V, maxillary division, VII and, within the skull, V, ophthalmic division, and V1. Predictably, the clinical features were mainly sensory with local or referred pain, or anaesthesia. Such changes were interpreted as perineural involvement only if they occurred within an appropriate anatomical distribution and unrelated causes such as local infection

274

R. L. C A R T E R ~ ~ U ~ .

Table 4 Perineural spread and TNM classification

Tz T, T.4

NO

N,

2 (1) 4 (1)

I (1)

Nz 3 (1) 3 (1)

N3 2

2

Notes: No distant metastases-'M' in the T N M classificationpresent in any of the patients. Numbers in parentheses indicate patients in whom perineural spread was clinically occult.

could be excluded (see also 'Discussion'). All but 3 of the patients showed symptoms and signs of neural involvement early in the disease, either at the time of presentation or shortly afterwards. The relation between perineural spread and the clinical extent of the disease, judged according to the TNM classification' I at the time of surgery, is shown in Table 4. Nerve involvement was more likely to occur, and more likely to be clinically apparent, with large primary tumours. PATHOLOGICAL F I N D I N G S

Perineural infiltration, usually in association with other modes of spread, was found more often with less differentiated tumours-occurring with 3 well differentiated tumours, 6 moderately differentiated tumours and 8 poorly differentiated lesions. Local infiltration, sometimes extensive, was present in all instances. Direct invasion of bone was found in 3 patients and of ossified laryngeal cartilage in 3. T h e relation between perineural spread, involvement of lymphatics and lymph nodes, and vascular invasion is shown in Table 5. Perineural infiltration Table 5 Patterns of dissemination of

17

tumours showing perineural spread

Spread : Site of tunrour Tongue, floor of mouth (5) Alveolus (2) Retromolar trigone ( I ) Soft palate, tonsil ( 2 ) Larynx (3) Nose ( I ) Upper deep cervical lymph nodes (3)

perinerrrnl

lymphatics, lymph nodes

blood vesdds

Figure I Squamous cell carcinoma of lower alveolus, invading perineural spaces in small branches of the inferior dental nerve (N). Haematoxylin and eosin. x zoo. Figure 2 Squamous cell carcinoma of nose infiltrating perineural spaces in a local cutaneous nerve (N), z cm from the main lesion. The normal structure of the nerve fibres is disorganized with loss of nuclei, swelling and fragmentation. Haematoxylin and eosin. x zoo.

276

R . L. C A R T E R C ~ ~ ~ .

was usually found in combination with lymphatic and/or haematogenous invasion, lymphatic spread occurring in 13 of the cases (76%); Perineural spread, apparently alone, was demonstrated in the remaining 4 patients. The pathological changes associated with perineural spread of tumour are illustrated in Figures 1-5. Tumour cells fill the perineural spaces and infiltrate along them, encircling the nerves in a continuous concentric sheath. There is a little invasion inwards into the deeper layers of the nerves (Figures I & 2). Swelling and fragmentation of nerve fibres can often be seen in conventionally stained tissues, and special stains show profound disorganization of fibres and disruption of myelin sheaths and axons (Figure 3a, b). T h e perineural spaces are sometimes involved by tumour for distances of several centimetres, and they provide an important route for dissemination. T h e mandible and maxilla in 3 patients were invaded mainly as a result of tumour cells spreading along the perineural spaces of branches of the superior and inferior dental nerves. In one instance the inferior dental nerve, densely infiltrated by tumour within the inferior dental canal, was totally infarcted (Figure 4). Spread of carcinoma cells in the epineural spaces of X, in the carotid sheath, was found in one patient. I M P L I C A T I O N S FOR M A N A G E M E N T A N D P R O G N O S I S

Clinical and pathological evidence of perineural infiltration may provide a useful guide to management indicating, in different circumstances, the need for more extensive surgery or for conservative surgery supplemented by other modes of treatment. Two patients from the present series illustrate those contrasting situations. G.B., a 60-year-old farm manager, developed a well-differentiated squamous carcinoma of the nostril which had been treated since 1975 by several local excisions and by radiotherapy. An extended total rhinectomy was eventually performed. Tumour was present in the perineural spaces of cutaneous nerves at the resection lines. Clinical examination showed a sensory loss in the maxillary division of V and motor loss in the buccal branch of VII for about 2 cm from the original resection margins. One course of chemotherapy (vincristine, bleomycin, methotrexate) was given and, 7 days after rhinectomy, the anaesthetic skin and soft tissues were excised. A remarkable feature of the new operation specimen was the presence of degenerate and dead tumour cells in the perineural spaces (Figure 5). A symptom-free period of about 5 months ensued but the patient then developed a small local recurrence and additional ipsilateral involvement of V ophthalmic division, and VI. H e died 2 months later. P. O'D., a 61-year-old labourer, developed a moderately differentiated squamous carcinoma involving the base of the tongue, the anterior floor of the mouth and the mandibular ramus together with enlarged submandibular lymph nodes. Response to chemotherapy (vincristine, bleomycin and methotrexate) and radiotherapy was incomplete. Radical excision was contemplated but the patient continued to complain of increasing pain in the distribution Figure 3 Squamous cell carcinoma of tongue (T) infiltrating perineural spaces of local nerves. a Degeneration of myelin with swelling and fragmentation (methazole fast Blue) ; b Degeneration of axons (Glees-Marsland). x zoo.

278

R. L. C A R T E R C t

(11.

Figure 4 Infarcted nerve (N) in inferior dental canal, associated with perineural spread of squarnous cell carcinoma from the lower alveolus. Apart from occasional peripherally-placed blood vessels, the tissue is totally necrotic. Fragments of mgelin and axons were found within this lesion in special stains. Haernatoxylin and eosin. x 60.

of V, mandibular division, and was found to have decreased mobility of the tongue and local anaesthesia to pin-prick. A repeat wide biopsy of the buccal mass showed tumour infiltrating perineural spaces. The patient is now being maintained on palliative chemotherapy. The extent to which perineural spread may determine prognosis requires investigation of more patients over a longer period of time. Twelve patients in the present series have so far been followed up for between 4 and 14 months since perineural infiltration of tumour was detected. Of these, 5 died within 3-11 months from recurrent tumour and 4 are alive with locally recurrent disease.

Discussion In an unsclected series of 70 consecutive cases of squamous carcinoma of the head and neck, treated by surgery, we have demonstrated perineural spread of tumour in 24%. Slightly

Perineural spread of tuntours

279

Figure 5 Squamous cell carcinoma of nose. Dead tuniour cells (T) in perineural spaces of a cutaneous nerve (N), 7 days after chemotherapy. Haematoxylin and eosin. x zoo.

more than half these cases had primary tumours arising within the buccal cavity. The trigeminal nerve was most commonly affected. Though based on comparatively small numbers, these results confirm the general conclusion of Ballantyne and his colleagues,6 that the site of the primary tumour is of major importance in determining the pattern of perineural involvement. The anatomical location of the lesion will also determine the extent to which perineural infiltration can be diagnosed clinically. Strict diagnostic criteria were used in the present series and, for tumours at certain sites, it is questionable whether a clinical diagnosis of perineural spread can be made. In carcinoma of the larynx, for example, it is usually impossible on clinical grounds alone to decide whether a fixed cord is due to local fibrosis, neoplastic infiltration of the arytenoid muscles or perineural invasion; a clinical diagnosis of perineural spread only becomes feasible when the tumour has extended outside the larynx. The present results differ from those of Ballantyne in that perineural spread has been found to occur early in the disease, a discrepancy which may reflect the fact that his series included a large number of skin carcinomas. We also find that perineural infiltration is more likely with large tumours, moderately or poorly differentiated, which are showing signs of lymphatic and/or vascular invasion. The detailed pathological changes of perineural spread have received little attention. The perineural spaces are no longer regarded as lymphatic^,^.^ but it is still unclear whether

280

R . I,. CARTER

et al.

they are merely unlined clefts or represent more organized structures.’ T h e characteristic pattern of tumour cell infiltration within perineural spaces is familiar but the accompanying nerve damage has not previously been described. Willis4 claimed that ‘medullated nerves enjoy a decided immunity from invasion by epidermoid carcinoma’, and he went on to comment that he had ‘yet to see an example of epidermoid carcinomatous invasion and disruption of a microscopic nerve bundle’. Deep infiltration into the substance of nerves was not found in the present study but disorganization of fibres was regularly seen. Although special histological techniques emphasise such abnormalities, some alterations can usually be detected in routinely stained material. T h e changes were observed only in nerves infiltrated by tumour so there is no reason to implicate previous radiotherapy or chemotherapy. The type of damage sustained by infiltrated nerves is non-specific and probably reflects a combination of mechanical trauma and hypoxia. Nerves are normally vascularized by a series of segmental arteriolar meshworks in the perineurium from which capillary plexuses pass inwards;’ 3 * 1 4 such vessels are likely to be compressed by columns of tumour cells tracking along the perineural spaces, resulting in hypoxia or even infarction. Once within perineural spaces, tumour cells can invade axially in either direction. Intracranial extension has been stressed previously;6 but centrifugal spread along the nerve periphery also occurs, providing an important route of invasion into adjacent tissues such as bone. This pattern of peripheral spread has received less attention’5*16 though it has important surgical implications. The traditional approach of excising the primary tumour, associated fascia1 spaces and regional lymph nodes may be inadequate and, in selected cases, complete clearance may require resection of nerves traversing the tumour site.” Equally, unequivocal evidence of neural spread may be an indication for less, rather than more, radical surgery. Examples of both situations are described in this paper. Experience in the present series is too small and too recent to comment on prognostic aspects; but with the possible exception of some squamous carcinomas of the skin (where perineural spread may occur as the only sign of dissemination), it appears to be a feature of aggressive, infiltrative disease and indicates a poor prognosis, particularly when the tumour has gained access to the cranial cavity.

Acknowledgements We are greatly indebted to Mrs Diana Mitchell and her staff, Department of Pathology, Royal Marsden Hospital, Sutton, for the histology; and to Mr K. G. Moreman for the photomicrographs.

References I

ERNSTP. (1907) Uber das Waclistuni und die Verbreitung bosartiger Geschwiilste insbesondere des Krebs in den Lymphbahnen der Nerven. Beitrage mu pathologisches Anatomie Suppl. 7, 29.

Perineural spread of tuniours

281

LARSOND.L., RODINA.E., ROBERTS D.K., O’STEENW.K., RAPPERPORT A.S. & LEWISS.R. (1966) Perineural lymphatics: myth or fact? American Journal of Surgery 112, 488. 3 RODINA.E., LARSON D.L. & ROBERTS D.K. (1967) Nature of the perineural space invaded by prostate carcinoma Cancer 20, 1772. 4 WILLISK.A. (1973) The Spread of Tumours in the Human Body, 3rd edn. Butterworths, London. 5 WILLISR.A. (1930) Epidermoid carcinoma of the head and neck, with special reference to metastatis. Journal of PatholoRy and Bacteriology 33, 5 0 1 . 6 BALLANTYNE A.J., MCCARTEN A.B. & IBANEZ M.L. (1963) The extension of cancer of the head and neck through peripheral nerves. American Journal of Surgery 106, 651. M.H., BAUERW.C. & OGURA J.H. (1961) The incidence of cervical lymph node metastases 7 MCGAVRAN z

from epidermoid carcinoma of the larynx and their relationship to certain characteristics of the primary tumour. Cancer 14, 55. 8 MACCOMB W.S. & FLETCHER G.H. (1967) Cancer of the Head and Neck. Williams & Wilkins, Baltimore. 9 SESSIONS D.G. (1976) Surgical pathology of cancer of the larynx and hypopharynx. Laryngoscope 86, 814. 1 0 MCDONALD T.J., DESANTO L.W. & WEILAND L.H. (1976) Supraglottic larynx and its pathology as studied by whole laryngeal sections. Laryngoscope 86, 635. 1 1 T N M CLASSIFICATION OF MALIGNANT TuhroURs (1974) 2nd edn. UICC, Geneva. 1 2 CLIFFORD P., O’CONNORA.D., DURDEN-SMITH J., HOLLISB.A.B., EDWARDS W.G. & DALLEY V.M. (1978) Synchronous multiple drug chemotherapy and radiotherapy for advanced (stage I11 & IVj squamous carcinoma of the head and neck. Antibiotics and Chemotherapy 24, 60. 13 SHANTHAVEERAPPA T.R. & BOURNEG.H. (1962) The ‘perineural epithelium’, a metabolically active, continuous, protoplasmic cell barrier surrounding peripheral nerve fasciculi. Journal of Anatomy 96, 527. 1 4 ADAMW.E. (1942) The blood supply of nerves. I . Historical review. Jozirnal oj’Anatonzy 76, 323. 1 5 ADAhIs W.E. (1943) The blood supply of nerve;, part 2. Effects of exclusion of its regional sources of supply on the sciatic nerve of the rabbit. Journal of Anatomy 77, 243. 16 DODD, G.D., DOLAN P.A. BALLANTYNE A.J., IBANEZ M.L. & CHAUP. (1970) The dissemination of tuniours of the head and neck via the cranial nerves. Radiologic Clinics ojNorth America 8, 445. 17 CALCATERRA J.C., CHERNEY E.F. & HANAFEE W. (1973) Normal variations in size and neoplastic changes of skull foramina. Laryngoscope 83, 1385. 18 SOUTHAM J.C. (1970) The extension of squamous carcinoma along the inferior dental neurovascular bundle. British Journal of Oral Surgery 7, 137. 19 LIGGINS D.F. & MEHROTRA P.N. (1978) The spread of facial tumours along bony canals. British 3ournal of Plastic Surgery 31, 54. 20 LEEE.S. & WILSON J.S.P. (1973) Carcinoma involving the alveolus. An appraisal of past results and an account of current management. British Journal of Surgery 60,85.

Perineural spread in squamous cell carcinomas of the head and neck: a clinicopathological study.

Clinical Otoluryngology 1979, 4, 271-281 Berineural spread in squamous cell carcinomas of the head and neck: a clinicopathological study R . L . CART...
992KB Sizes 0 Downloads 0 Views