TEMPORAL BONE RESECTION FOR CANCER

MILLAR

TEMPORAL BONE RESECTION FOR CANCER HUGHS. MILLAR Royal Melbourne Hospital a n d Peter MacCallum Hospital Fifteen patients with cancer Involving the temporal bone have beenconsidered for radical surgical treatment by partial resection of the temporal bone during the past 12years. All but one had undergone previous treatment by local surgery andlor irradlatlon. Two patients proved to be inoperable at surgical exploration. Three types of partial resection of the temporal bone and described to encompass disease involving the concha, the mastoid and squamous areas of the temporal bone, the ear canal, the middle ear, and the parotid gland. Closure of the surgical defect has been achieved in live cases using the residual pinna, inlour cases with scalp flaps, and in flve cases with a deltopectoral flap. Complications have been surprisingly few, with only one postoperative death. in one case communicating hydrocephalus persisted until death from residual disease many months later. Minor repair failure occurred in two patlents. No attempt has been made to restore facial nerve function by grafting procedures. Long-term survival has been disappointing: however, It is considered that such radical surgery remains justified in selected cases for relief of the pain and disfigurement caused by chronic ulcerating neoplastic disease.

PRIORt o 1950 the accepted method of treatment of cancer of the ear canal andlor middle ear was by radiation therapy, either alone or following radical mastoidectomy. Even so, the results of treatment were acceptable, for Lederman (1965) reported fiveyear survival figures of 24% for external auditory canal malignancies and 30% for petromastoid malignancies treated by radiation therapy only. In 1951 Ward et alii and also Campbell etaliidescribed the piecemeal removal of the temporal bone t o the depth of the carotid canal as an advance on previous surgical endeavours. It was not until 1954 that Parsons and Lewis first described the surgical technique of monoblock resection of the temporal bone, following which important contributions were made by Conley and Novack (1960) and Coleman (1966). Excellent descriptions are available in the surgical atlases of Lore (1962) and Montgomery (1971). Hilding and Selker (1969) introduced a minor modification which ensures that the body of the petrous bone medial to the middle ear cavity is included in the resected specimen for cancer involving the middle ear proper. INDICATIONS FOR SURGERY

Fifteen patients with malignant tumours involving the temporal bone have been considered suitable for radical partial resection of the monoblock type during the past 12 years. Six of these patients had Reprints Dr ti S Millar. F R A C S Parade Fitzroy. vic 3065

F R C S ,D L 0

183 Victoria

AUST. N.Z. J. SURG.,VOL. 48 - No 6, DECEMBER. 1978

received full-dosage radiotherapy as primary treatment and were referred for salvage surgery after a considerable delay. Eight patients had undergone various local resections of primary tumours involving the periaural skin or parotid gland. Only one patient was treated de novo without failure of previous modalities. Sixteen operations have been carried out on the 15 patients, the condition in two of whom proved inoperable at the initial exploration. Inoperability was determined in one case of primary middle ear cancer when there proved to be extensive involvement of the dura of the middle cranial fossa on primary exposure of that area. In the other patient there was extensive involvement of a residual parotid gland tumour which had infiltrated deeply along the base of the skull. The indications for attempting temporal bone resection were as follows: (1) Squamous cell or basal cell carcinoma of periaural skin: (a) Direct invasion of the temporal bone (three cases); (b) Metastases to upper jugular lymph nodes involving the mastoid tip or the parotid lymph nodes from remote controlled primary skin cancer (four cases); (2) Squamous cell or basal cell carcinoma of the bony section of the external auditory canal (three cases); (3) Squamous cell cancer of middle ear origin (two cases); 621

TEMPORAL BONE RESECTION FOR CANCER

.

i

/’

*

(4)

I

FIGURE1: incisions for preservation of pinna.

Recurrent parotid tumours (three cases, including two mixed salivary turnours and one carcinoma).

SURGICALTECHNIQUE The method of surgical resection is planned according to the locality and extent of involvement of the temporal bone depending on the findings on c l i n i c a l examination, examination under anaesthesia including microexamination of the ear canal and tympanic membrane, and multiple biopsies of the margins of the disease. Radiological studies help only to a limited extent in that negative findings on tomography do not necessarily exclude the possibility of bony involvement, whereas radiological evidence of bone erosion tends to occur only in advanced disease and is more useful in determining inoperability than the possibility of cure. Computerized axial tomography would now be used as a routine. Carotid angiography and retrograde jugular venography may be used in selected cases. The treatment of glomus tympanicum and jugulare turnours is not considered in this paper. Three variations in the technique of surgical resection have been practised in order to encompass the area of involvement by disease without resorting to radical monolock resection to the level of the carotid canal in every case. ( 1 ) Wedge Resection of the Ear Canal This procedure allows adequate resection of cancer involving the bony and cartilaginous sectors 622

MILLER

of the ear canal together with the tympanic membrane, but with preservation of the facial nerve after the method described by Conley and Novack (1960). Hutcheon (1966) has described experience with this techique in Brisbane. The-pinna is preserved on a temporal-based flap after wide circumferential incision of the concha and tragus (Figure 1). The pinna is retracted upward to expose widely the mastoid and squamous sectors of the temporal bone. A complete simple mastoidectomy is carried out to allow exposure of the attic of the middle ear, followed by a posterior tympanotomy to inspect the middle ear contents. The facial nerve is then exposed in its entire descending length down to the stylomastoid foramen, using the electric drill and the operating microscope. The lateral wall of the attic of the middle ear is completely removed until the tympanic plate is reached anteriorly. The capsule of the mandibular joint and the parotid fascia and gland are dissected away from the tragus and the cartilaginus meatus. Eventually it is possible to position a small curved osteotome superficial t o the facial nerve through the middle ear cavity t o fracture the tympanic plate and remove the ear canal all en bloc. This leaves in effect a large “radical mastoidectomy cavity” for which a wide meatoplasty has already been fashioned by virtue of the wide excision of the concha. The cavitv should be allowed to granulate and epithelializk from without rather then be skin-grafted to allow observation for recurrent disease. AUST.N.Z. J. SURG..VOL. 48

- No. 6, DECEMBER, 1978

TEMPORAL BONE RESECTION FOR CANCER

(2) Superficial Resection of the Temporal Bone This procedure allows resection of the mastoid and squamous segments of the temporal bone through the plane of the middle ear together with the parotid gland, mandibular joint and ramusof the mandible, i.e., it will encompass skin cancer invading the temporal bone or residual parotid gland tumour or parotid lymph node disease. The depth of resection is inadequate should there be irrvotvement of the middle ear cleft. The success of this type of resection is determined by the depth of thedisease in the neck at the level of the mastoid tip or parotid area. It is therefore preferable to commence with exploration of the upper neck region in order to ensure that there is a plane between the carotid sheath following ligation of the external carotid artery and that the disease is contained in the potential specimen. A formal radical neck dissection can be carried out simultaneously if indicated, but is not necessarily an advantage as the plane of separation of the temporal bone specimen is superficial to the internal jugular vein. Therefore, a neck dissection has only been done for therapeutic reasons in five cases in this series. Following preliminary exploration of the upper neck region the incisions are completed as indicated in Figure 1. If the pinna is to be preserved it is turned up with the temporal flap, or it may be resected partially or totally with a wide margin of adjacent mastoid, parotid, or upper cervical skin. The incision can be extended downward to allow for a modified or complete radical neck dissection. The dura of the middle cranial fossa is now widely exposed through the thin portion of the squamous temporal, using hammer and gouge, burr holes, or the drill, and this line is extended anteriorly in a circumferential fashion to the mandibular joint and posteriorly around behind the mastoid process. The emissary vein from the sigmoid sinus may give troublesome bleeding here. The lateral sinus is defined and carefully elevated from the bone and followed to the sigmoid portion. This dissection becomes difficult along the posterior surface of the petrous temporal. In the event of tearing, the lateral sinus must be ligated, so that at this stage it is helpful to have the assistance of a neurosurgical colleague. After wide removal of the squamous temporal and careful elevation of the dura and lateral and sigmoid sinuses, it is possible to define the arcuate eminence which marks the superior semicircular canal which is medial to the roof of the middle ear. Exposure of the petrous apex is facilitated by giving intravenous mannitol.

AUST N.Z. J. SURG,VOL 48 - No 6, DECEMBER, 1978

MILLER

Attention is then directed inferiorly to divide the attachments of the sternomastoid and digastric muscles to the mastoid process. The external carotid artery has been ligated in the preliminary exploration and the carotid sheath is already freed. The internal jugular vein is dissected on its superficial surface and gently elevated from the lateral wall of the jugular foramen. It is usually easier and often necessary to remove the entire parotid gland with all or part of the ramus of the mandible and the mandibular joint. The mandible is now divided with a Gigli saw at the angle or higher, and the parotid gland is mobilzed along its anterior border without any attempt at preservation of the branches-of the facial nerve. A segment of the zygomatic arch is removed to allow access to the mandibular joint which may or may not be included in the monoblock. Now the line of resection in. the anterior squamous temporal is extended inferiorly either through the joint or anterior to it with an osteotome. For superficial resection of the temporal bone the plane of removal is through the middle ear where there is a natural cleavage (Line A, Figure 2). Furthermore, this is well superficial to the jugular foramen and carotid canal, so that these vessels are not in any particular jeopardy. When all areas of dissection have been completed i n this circumferential fashion, an osteotome is first placed

gh petrous apex.

623

MILLER

TEMPORAL BONE RESECTION FOR CANCER

behind the mastoid process in the digastric groove to commence separation of the specimen. A second cut is then made through the petrous bone just lateral to the arcuate eminence through the tegmen tympani or roof of the middle ear. At this point the block usually becomes loose, so that the cut can be continued anteriorly through the middle ear to divide the tympanic plate. Thespecimen can now be retracted superiorly to allow direct inspection of the deep surface and division of the soft tissue attachments such as the styloid muscles and the medial pterygoid. Brisk bleeding may occur from the venous plexus of the pterygoid fossa. but this is readily controlled with pressure and with diathermy. (3) Monoblock Resection of the Temporal Bone through the Carotid Canal This procedure is necessary for neoplastic disease involving the middle ear proper so that the line of resection must be through the body of the petrous bone medial to the arcuate eminence (Line B, Figure 2). The success of this type of procedure depends on whether there has been extension of disease through the tegmen tympani to involve the middle or posterior cranial fossa dura, so that in this situation it is advisable to expose the middle cranial fossa in the first instance in order to determine operability.

FIGURE 3: Seventy-year-old lady with huge recurrent malignant salivary tumour.

624

Once operability has been determined superiorly, the procedure is the same as described in the superficial resection, except that it may be necessary to remove more bone from the squamous temporal to obtain exposure medial to the arcuate eminence. Also, the sigmoid sinus must be elevated from the posterior surface of the petrous apex as medially as possible; otherwise it will be torn in the process of the transsection. As previously, the anterior line of resection can be either through the mandibular joint or anterior to it, this decision depending on whether or not there is extratemporal extension of disease. The difficult part of this operation is to obtain the correct line of cleavage through the petrous apex without fracturing it into the middle ear. The bone medial to the arcuate eminence is intensely hard, so that Stryker saw cuts on each of the three surfaces are helpful i f this instrument con be introduced sufficiently into the area. These cuts are followed by judicious use of a curved osteotome until the correct fracture line is obtained, after which the temporal bone can be rocked free. The line of resection should run through the jugular foramen lateral to the jugular bulb and immediately lateral to the carotid canal. It includes the whole tympanic plate together with the styloid process. In the event of tearing or the necessity to resect dura mater, repair is achieved with either temporal fascia or fascia lata. Figure 3 shows a 70-year-old lady with a huge recurrent malignant salivary tumour and Figure 4 illustrates diagramatically the anatomy of the bed of the dissection following deep temporal bone resection.

FIGURE 4:

AUST

Diagram of surgical defect following deep temporal bone resection.

N.Z.J . SURG, VOL

48 - No 6, DECEMBER, 1978

MILLER

TEMPORAL BONE RESECTION F O R CANCER

FIGURE 5 : Twelve-year survivor following deep temporal bone resection with preservation of pinna.

Repair of the Defect The pinna can usually be preserved and replaced in resections of cancer localized to the ear canal and/or middle ear. Figure 5 illustrates a 12-year survivor who underwent a monoblock resection for middle ear cancer with preservation of the pinna. Otherwise some form of skin replacement is

necessary. Split skin grafts have been used by many authors (Lewis, 1975), but this is considered inadequate following previous radiotherapy, and particularly if there has been interference with the continuity of the dura mater. Various plastic surgical colleagues have cooperated in devising remote flaps for repair as follows: (i) anterior based scalp flap - 1 case; (ii) posterior based scalp flap - 3 cases. Figure 6 illustrates the minimal deformity which may occur following a repair of this type when the hair grows to camouflage the loss of the pinna. Based on the occipital artery, this flap is always dependable, but the donor site on the vertex requires a split skin graft. A secondary procedure is necessary to tidy up the pedicle; (iii) deltopectoral flap -5cases. Figure 7 illustrates the repair following excision of an extensive eroding skin cancer with radical neck dissection. The tip of the pinna has been preserved for support of the spectacles. A relatively long and broad deltopectoral flap easily reaches to the level of the zygomatic process. It may be convenient to delay the flap one week previously at the time of a preliminary examination and biopsy under anaesthetia. Conley and Schuller (1977) have devised a posterior cervical pedicle flap based superiorly on the occiput and receiving its blood supply from the p o s t a u r i c u l a r a n d o c c i p i t a l vessels a n d contralateral scalp for repairs of this nature. However, it may be very often impossible to preserve the ipsilateral postauricular and occipital arteries, so that this flap may not always be applicable. The same authors recommend immediate repair of the divided facial nerve by the technique of hypoglossal crossover, and point out the close anatomical proximity of the hypoglossal nerve to the cut peripheral end of the facial nerve.

FIGURE6: Use of posterior based scalp flap to repair defect of deep temporal bone resection (courtesy of Mr W. Wilson).

AUST. N.Z. J. SURG..Voc. 48 - No. 6, DECEMBER. 1978

625

TEMPORAL BONE RESECTION FOR CANCER

FIGURE 7: Useof

deltopectoral flap to repair defect of superficial temporal bone resection (courtesy of Mr Hunter Fry).

When total parotidectomy is included, they use the peripheral branches of the hypoglossal nerve to the tongue to anastomose to the peripheral branches of the seventh nerve. Lower motor neurone facial paralysis was present before operation in five of the patients in this series, and the facial nerve was sacrificed in another six, it being preserved only in the three wedge resections of the ear canal. No attempt has been made to carry out any form of facial nerve grafting or anastomosis, on the basis that the prognosis has been limited in all patients undergoing such surgery. Adequate facial rehabilitation has been achieved with lateral tarsorhaphies in every case and a delayed fascia lata facial sling in one case. Complications Complications have been surprisingly few, considering the relatively major extent of the surgery. Significant complications were as follows: (I) lntraoperative haemorrhage - (a) severe bleeding occured in one case from the torn sigmoid sinus and necessitated ligature. This was presumably responsible for the subsequent development of communicating hydrocephalus; (b) severe bleeding occurred in one case from the torn jugular bulb and was controlled by packing with absorbable gauze. This probably led to the loss of 626

MILLER

the tip of thescalp rotation flap; (c) difficult bleeding occurred in one case from a torn carotid bifrucation when the heavily irradiated external carotid artery was avulsed in the process of ligature. This was successfully repaired without neurological deficit by a vascular surgical colleague. (2) Postoperative death - This occurred in one patient after 48 hours due t o pulmonary complications when intensive care assistance was unavailable. He had known residual disease at the petrous apex. (3) Cerebrospinal fluid leak - This occurred in one patient and lasted for three weeks without morbidity. (4) Vertigo due to destruction ofthe labyrinth This has not seemed to disturb these patients unduly. The vertigo improves rapidly in the course of the first postoperative week and equilibrium soon returns to near normal. (5) Unilateral deafness - This is not important when there is normal hearing in the other ear. (6) Communication hydrocephalus - This was the only severe complication in one patient and was due to thrombosis extending around the transverse sinus to the sagittal sinus. This patient had residual disease involving the dura of the middle cranial fossa and subsequently died after three months. AUST. N.Z. J. SURG,VOL 48

- No. 6, DECEMBER. 1978

MILLER

TEMPORAL BONE RESECTION FOR CANCER

(7) Repair failure - Failure of repair occurred i n two patients, but was not serious. In one the tip of the scalp flap necrosed; in the second patient difficulties were encountered with infection of the split skin applied to the denuded calvarium at the donor site of the scalp flap.

RESULTS In 1975 Lewis reported on 100 cases of temporal bone resections and found a five-year cure rate of 27%, with an operative mortality of 5%. These results do not suggest any significant advantage over those of radiotherapy alone as reported by Lederman (1965); however, many of these operations were secondary procedures following failed radiotherapy. Conley and Schuller (1977) have obtained a five-year survival of 61% for ear canal cancer and 36.8% for middle ear cancer, using a programme of radical surgical ablation followed by postoperative irradiation. The number of patients in this series is too small and the follow-up time insufficient for one to make any conclusions concerning cure rates. Of the three patients undergoing wedge resections of the auditory canal, two are free of disease for over two years. The third patient developed rapid evidence of residual disease in the mastoid cavity, and despite postoperative radiotherapy and then a radical monoblock resection died after 33 months. Of the 11 patients undergoing radical resections, one died after operation (with residual disease involving the middle fossa dura) and sixothers have since died of disease after periods varying from three to 27 months, with an average survival of 16 months. Of these six, two developed recurrence in the temporal bone, one in the deep parotid areaand two in the neck. Of the other four patients, one died from natural causes, clinically free from disease after 14 months, and the other three are alive and well, the longest survival being 12 years following the first monoblock temporal bone resection in 1966. COMMENTS Once the patient has survived the major resection there has been minimum morbidity following the surgery. A significant advantage is the relief of pain resulting from the widespread division of the cutaneous nerve supply t o the area and the

AUST.N.Z. J. SURG., VOL. 48 - No. 6,DECEMBER, 1978

coverage by large remote skin flaps. The cosmetic disfigurement of facial paralysis and loss of one pinna have not seemed to matter to these patients. Unilateral deafness is not important provided that normal hearing is present in the other ear. The loss of one labyrinth causes only minor disturbances of equilibrium in the long term. Thus these procedures do seem to have some merit, even if applied for only palliative or succorsive indications such as ulcerating upper cervical node metastases at the mastoid tip and residual neoplastic disease in the deep lobe of the parotid gland. The main hope of improvement in results would appear to be by earlier surgery. In this regard it would be preferable to undertake radical surgery much earlier in the course of the disease, either soon after failed primary radiotherapy or as the primary modality, rather than deliver radiotherapy when cure by that method alone would seem unlikely. Then, postoperative radiotherapy as advocated by Conley and Schuller (1977) as a planned sequential programme would seem logical. '

ACKNOWLEDGEMENTS

I wish to thank colleagues of the Peter MacCallum Hospital who referred many of the patients for treatment; Mr Frank Ham and members of the Department of Plastic Surgery, Royal Melbourne Hospital, who provided invaluable help in the repair procedures; and Mr David Brownhill, Neurosurgeon, Royal Melbourne Hospital, for his advice and support. REFERENCES CAMPBELL, E.. BOLK.B.M. and BURKLUND. C.W. (1951, Ann. Surg.. 134: 397. COLEMAN. C. C. (1966), Amer. J. Surg., 112: 583. CONLEY, J. J. and NOVACK. A. J. (19W), Arch. Otolaryng., 71: 635. CONLEY, J. J. and SCHULLER. D. E. (1977), Arch. Otolaryng., 103: 34. HILDING.D.A.and SELKER,R. (1969),Arch. Otolaryng., 89: 636. HUTCHEON. J. R. (1966), Med. J. Aust., 2: 406. LEDERMAN. M. (1965). J. Laryng., 79: 85. LEWIS. J. S. (1962). Atlas of Head and Neck Surgery, edited by LORE. J. M.,W. 5.SAUNDERS CO.. Philadelphia: 204. LEWIS.J. S. (1975), Arch. Otolbryng., 101: 23. LEWIS. J. S. and PAGE.R. (1966), Arch. Otolaryng., 83: 114. MONTGOMERY. W. W. (1971), Surgery of the Upper Respiratory System, Lea 8 Febiger, Philadelphia, volume 1: 465. PARSONS. H and LEWIS, J. S. (1954). Cancer IPhilad.), 7: 995. WARD. G. E., LOCH. W. E. and LAWRENCE. W. (1951), Amer. J. Surg., 82: 169

627

Temporal bone resection for cancer.

TEMPORAL BONE RESECTION FOR CANCER MILLAR TEMPORAL BONE RESECTION FOR CANCER HUGHS. MILLAR Royal Melbourne Hospital a n d Peter MacCallum Hospital F...
679KB Sizes 0 Downloads 0 Views