for Radical Resection of Skull Base Tumors

Michael T. Lawton, M.D., and Robert F. Spetzler, M.D.

Internal Carotid Artery Sacrifice for Radical Resection of

Skull Base Tumors

Anterior skull base tumors frequently encase the internal carotid artery (ICA).1-8 Because complete resection is the objective of most tumor operations, the surgeon must decide how to manage this artery. Some tumors can be dissected cleanly from the artery and resected completely with carotid preservation.' Other tumors, however, adhere to or invade the artery and cannot be resected totally without injuring it.3'5-8 The choice between preserving the ICA and resecting a tumor subtotally or sacrificing the ICA to obtain a complete resection is difficult and controversial. Furthermore, the issue of whether to reconstruct the ICA after sacrifice is also controversial. We review our experience with ICA sacrifice and present our strategy for managing the ICA in the radical resection of anterior skull base lesions.

MATERIALS AND METHODS In an experience with more than 300 anterior skull base tumors treated between 1988 and 1994, the ICA was sacrificed and revascularized in 10 patients. There were five women and five men with an average age of 48 years (range, 31 to 72 years). Four patients presented with untreated tumors. Six patients sought treatment for recurrent tumor after being treated elsewhere. All six of these patients previously underwent surgical resection of their tumors, with multiple resections in four patients. Five patients received radiation therapy, and none received chemotherapy. Five patients presented with symptoms of mass ef-

Skull Base Surgery, Volume 6, Number 2, April 1996 Division of Neurological Surgery, Barrow Neurological Institute, St. Joseph's Hospital and Medical Center, Phoenix, Arizona. Reprint requests: Dr. Spetzler,% Neuroscience Publications, Barrow Neurological Institute, 350 West Thomas Road, Phoenix, AZ 85013-4496. Copyright X) 1996 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.

119

SKULL BASE SURGERYNOLUME 6, NUMBER 2 APRIL 1996

fect and neural compression. Two patients had progressively worsening pain, two had episodes of excessive nasopharyngeal bleeding, and one had transient ischemic attacks related to occlusion of the ICA by tumor. Cranial nerve (CN) deficits were present in all 10 patients and included two patients with decreased vision from optic nerve compression, nine patients with extraocular muscle movement abnormalities (CN III palsy, n = 9; CN IV palsy, n = 6; CN VI palsy, n = 7), seven patients with sensory deficits in the trigeminal distributions, four patients with facial paresis or palsy, one patient with hearing loss, and two patients with deficits related to the accessory nerve. The motor examination in all 10 patients was normal. Based on Glasgow Outcome Score (GOS) criteria,9 all 10 patients had moderate preoperative disabilities (GOS 2). Eight tumors were located in and around the cavernous sinus, and two were in the infratemporal fossa. Eight tumors were located on the right side and two were on the left side. Angiography demonstrated encasement of the ICA with luminal narrowing in nine patients; one patient had complete ICA occlusion. Pathologic examination of tumor specimens revealed the following diagnoses: five squamous cell carcinomas, two adenoid cystic carcinomas, one mucoepidermoid carcinoma, one malignant meningioma, and one benign meningioma.

RESU LTS All lesions were approached through a standard frontotemporal craniotomy. Additional exposure was gained through an infratemporal fossa approach in three patients, through a radical neck dissection in two patients, and through a transfacial approach in one patient. Six patients underwent staged resections of tumor. The tumors were resected totally in five patients and subtotally in five patients. All 10 patients had their ICA sacrificed as part of a radical tumor resection; revascularization was performed in all patients. Five patients were bypassed before tumor resection during the same operation; five were bypassed first, followed by tumor resection in a second-stage operation. Revascularization procedures included five cervical ICA-middle cerebral artery (MCA) saphenous bypasses (including one on the contralateral side for a complication), one cervical-to-supraclinoid ICA saphenous bypass, three petrous-to-supraclinoid (C3-C5) bypasses, and two bonnet bypasses (contralateral superficial temporal artery-to-ipsilateral MCA).10 Postoperative GOS decreased by one point in four patients. One patient developed an orocutaneous fistula, and the pectoralis muscle flap used to obliterate the fistula necrosed. A similar patient developed a wound infection that progressed to sepsis and vascular insufficiency to the latissimus dorsi muscle flap, which then resulted in necro120 sis. Both patients had undergone previous radical neck

dissection for tumor resection and radiation therapy, making the tissues susceptible to complications associated with wound healing. Both patients were treated with debridement and reconstruction of the wound with additional muscle flaps, and both patients recovered their baseline neurologic status before dying 1.5 and 1.3 years after surgery, respectively. Two patients suffered ischemic complications. A bypass graft occlusion in one patient resulted in infarction and new deficits despite immediate revision. This patient was lost to follow-up 2 months after surgery. Another patient ruptured a pseudoaneurysm that developed on the opposite cavernous ICA after a transfacial approach. The epistaxis was tamponaded and a contralateral cervical ICA-MCA saphenous-vein bypass excluded the injured segment of ICA from the circulation. Although this patient suffered a cerebral infarction, he functions independently and showed only moderate disabilities at his 2-year follow-up examination. His current GOS is at his preoperative level. One additional complication occurred without neurologic sequelae in a patient with a cerebrospinal fluid leak that resolved after placement of a lumboperitoneal shunt. Overall, all except the patient described in the previous paragraph (with only 2-months follow-up) returned to their preoperative neurologic status as determined by GOS. Currently, two patients are alive, three patients are dead, and five patients have been lost to follow-up (mean duration, 1.7 years). None of the surviving patients have symptoms or radiographic evidence of local recurrence. The additional death was caused by pneumonia 7 months after surgical resection of a squamous carcinoma.

DISCUSSION This small experience with carotid sacrifice for radical resection of skull base tumors reflects our practice of preserving the ICA whenever possible. Our protocol for managing these lesions is derived from a review of the literature and relates to the benignancy or malignancy of the tumor. We preserve the ICA with benign tumors because they do not invade the artery, or they invade it to a limited extent. In contrast, we recommend radical tumor resection and sacrifice of the ICA with malignant tumors because they directly threaten the integrity of the ICA and, thus, the patient's survival. The ICA should not be considered a limitation to radical tumor resection because it can be safely reconstructed with an appropriate bypass procedure.

Benign Skull Base Tumors The majority of anterior skull base neoplasms-. meningioma, nerve sheath tumors, pituitary adenoma,

INTERNAL CAROTID ARTERY SACRIFICE-LAWTON, SPETZLER

and juvenile angiofibroma-are benign.7 Many of these tumors can be resected completely without sacrifice of the ICA. DeMonte et al2 preserved the ICA in all 41 patients with cavernous sinus meningiomas and achieved total resection in 31 patients (76%). In his review of clinoidal meningiomas, Al-Meftyl concluded that the presence of an arachnoid membrane between the ICA and the tumor determines whether dissection of the vessel from the tumor is possible; complete ICA encasement by the tumor does not prevent total resection. However, this arachnoid layer is absent in extradural or interdural spaces such as the cavernous sinus, and it is also absent in patients who have undergone previous operations or irradiation. Meningiomas that encase the ICA without this intervening arachnoidal layer adhere to the adventitia and cannot be resected completely without sacrificing the ICA. The 76% total resection rate in DeMonte's series demonstrates that a protective arachnoidal membrane is often present around meningiomas; therefore, the majority of these benign tumors can be resected completely while preserving the ICA. Other benign tumors, such as pituitary adenomas, neurilemomas, and angiofibromas, are more likely than meningiomas to be dissected free of the ICA.4 Residual benign tumor left around the ICA to preserve it could potentially invade the vessel wall and compromise the integrity of the artery. However, benign tumors tend not to invade the ICA to such an extent. Kotapka et aP3 studied ICAs encased by cavernous sinus meningiomas in 19 patients. Luminal narrowing was present in all but two. Eight of these patients' ICAs were infiltrated by tumors-to the outer adventitia in five patients and to the media through adventitia in three patients. Apparently, meningiomas invade the carotid wall, but only to a limited extent. Therefore, residual tumor on the ICA after resection does not threaten the integrity of the vessel. Tumor debulking is often sufficient to restore lumenal patency. Residual benign tumor around the ICA could potentially increase the incidence of tumor recurrence. Because the risk of tumor recurrence decreases with more extensive surgical resection,11 the recurrence rate should be lower in patients whose carotid arteries are sacrificed than in patients whose carotid arteries are preserved. However, two representative series of patients with cavernous sinus meningiomas that were aggressively resected, either with8 or without ICA resection,2 had recurrence rates that were strikingly similar. Sen and Sekhar8 treated 17 patients with tumor and ICA resection, totally resecting the tumors of 76% of the patients. The symptomatic and radiographic recurrence rate was 8% after total tumor resection and 25% after subtotal tumor resection. In comparison, DeMonte et a12 achieved the same gross total resection rate in 41 patients without carotid resection. In this series, the rates of recurrence were 3% and 20% after total and subtotal resections, respectively. Therefore, ICA resection in the treatment of meningiomas of the anterior skull base does not appear to lower the rate of recuffence.

The lack of ICA invasion by residual tumor and the lack of improvement in the tumor recurrence rate with carotid resection argue for preserving the ICA when dealing with benign skull base tumors. ICA preservation is the practice at our institution. Furthermore, the risk of complication from elective ICA occlusion is 7% to 10% in our experience, even with revascularization.12 This risk cannot be justified in the treatment of benign lesions. Although total resection is the goal of this surgery, subtotal resection of a benign tumor is considered safe. Nonetheless, patients with residual tumor are followed carefully for recurrence with serial neurologic and imaging examinations. An exception to this protocol is a benign tumor that completely occludes the ICA. One such patient in this series had a meningioma that completely occluded the ICA and caused transient ischemic attacks. Complete ICA occlusion permits resection of the artery with the tumor, but the ischemic symptoms indicate that cerebrovascular reserve may be inadequate and necessitate a bypass graft.

Malignant Skull Base Tumors An ICA surrounded by a malignant tumor is managed differently from one surrounded by a benign tumor. First, because malignant tumors can invade adjacent structures, metastasize, and cause death, oncological principles of radical resection with tumor-free margins must be followed more strictly. Second, the risks of ICA sacrifice are justified when compared to the dismal prognosis of patients with malignant neoplasms. Our treatment strategy is based on the invasive behavior of malignant tumors. Unlike malignant neoplasms which tend to invade the ICA wall directly and can lead to frank ICA rupture,13-16 benign neoplasms tend to invade the ICA superficially and do not precipitate rupture. Because ICA rupture can cause potentially life-threatening hemorrhage, an artery involved with a malignant tumor is best resected. The risk of an ICA rupture invaded by malignant intracranial tumor is unknown. However, it is likely to be similar to the risk of rupture of an ICA invaded by malignant cervical tumor-which is around 18%.13,14,17 Furthermore, most of these ICAs rupture within 6 months.13'14 Therefore, when dealing with malignant neoplasms, the risk of ICA preservation and subsequent rupture exceeds the risk of ICA sacrifice with revascularization. The effect of ICA resection on tumor recurrence and the survival of patients is difficult to determine because no prospective, randomized studies have analyzed this issue. Malignant tumors are more difficult to resect than benign tumors, with total resection rates of 55% and 86%, respectively.7 Recurrence after total resection is also increased with malignant tumors (25% vs 6%). ICA sacrifice permits en bloc resection of the cavernous sinus and,

121

SKULL BASE SURGERYNOLUME 6, NUMBER 2 APRIL 1996

therefore, a better chance at disease-free margins of resec- in these patients would have been more devastating withtion.18 However, it is unclear whether aggressive resec- out revascularization, and revascularization contributed tion extends a patient's survival significantly.'9 At the to the recovery in the second patient. least, radical tumor resection should attempt to relieve local symptoms and to decrease the rate of intracranial recurrence.

Reconstruction of the ICA ICA resection produces a risk of ischemic complications. Morbidity from elective ICA resection without reconstruction for neck carcinomas ranges from 0% to 45%, and mortality ranges from 0% to 31%.20 Consequently, the surgeon must determine that the patient has either the collateral circulation to tolerate ICA sacrifice, or the surgeon must restore cerebral blood flow with a bypass graft. Diagnostic studies such as the balloon occlusion test with xenon computed tomographic-cerebral blood flow (Xe CT-CBF) identify patients who are unlikely to tolerate ICA sacrifice and who will require revascularization.5'6'8 These provocative tests may cause temporary or permanent ischemic deficits ranging from 0.7% to 7%,21,22 and their false negative rates range from 2% to 22%.8,22-24 In addition, such tests do not predict the risk of delayed ischemic complications or de novo aneurysm formation. Therefore, there is considerable cumulative risk from a balloon occlusion test, immediate ischemic complications, delayed ischemic complications, and de novo aneurysm formation. We tend to revascularize patients whose ICA has been sacrificed because this approach has a surgical morbidity rate between 7% and 10% and a small delayed stroke risk from symptomatic graft failure.12 The indications for revascularization remain controversial. The revascularization procedure is performed immediately before the tumor is resected or in a preceding surgical stage. Intraoperative somatosensory evoked potentials (SSEPs) are used routinely during the anastomosis. Changes in SSEPs alert the surgeon that the patient is not tolerating the temporary ICA occlusion, and ICA perfusion should be maintained during proximal anastomosis by suturing the bypass graft to the external carotid artery. Vascular complications occurred in 20% of the patients in this series, a value similar to the stroke rate following ICA occlusion.22 Although this similarity might suggest that the bypass provided no benefit, the bypasses in these two patients likely protected them from devastating complications. The one patient who developed an infarction after occlusion of the graft was dependent on it and would not have tolerated ICA sacrifice without it. The other patient who developed an infarction after pseudoaneurysm rupture would probably have developed bilateral infarctions from bilateral carotid occlusions and died had it not been for the revascularization performed prior to tumor resection. Instead, he returned to his pre122 operative level of function. The ischemic complications

Limitations of the Current Study The protocol at this institution for managing tumors that encase the ICA favors conservation rather than sacrifice of the vessel. This series of patients is therefore small, and unfortunately does not lend itself to statistically meaningful conclusions about the efficacy of treatment. The impact of ICA sacrifice on local recurrence and patient survival is a difficult and controversial topic. This study is retrospective and did not include a control group of patients with malignant tumors resected without ICA sacrifice for comparison. Furthermore, follow-up is limited. A randomized, prospective study would generate recurrence rates and survival data for patients whose tumors were resected with and without ICA sacrifice. That only 10 patients underwent ICA sacrifice at this institution in 6 years indicates that a multicenter study would be required to obtain enough patients for a robust statistical analysis. The current study does not resolve this controversial issue, but it does present an approach to treatment and results, as well as justification based on a review of the literature. Regarding the literature review, the analysis relied heavily on several articles.2,3'68 Unfortunately, few articles provide data relevant to this issue. Many authors discuss recurrence rates and survival after tumor resection,11'25-28 but these studies include tumors in diverse locations that do not relate to the carotid artery. The very limitations of this study emphasize the need for further investigations or for a prospective, randomized trial to provide definitive answers to whether the ICA should be preserved or sacrificed in patients with tumors.

REFERENCES 1. Al-Mefty 0: Clinoidal meningiomas. J Neurosurg 73:840-849,

1990 2. DeMonte F, Smith HK, Al-Mefty 0: Outcome of aggressive removal of cavernous sinus meningiomas. J Neurosurg 81:245251, 1994 3. Kotapka MJ, Kalia KK, Martinez AJ, Sekhar LN: Infiltration of the carotid artery by cavernous sinus meningioma. J Neurosurg 81:252-255, 1994 4. Lanzino G, Hirsch WL, Pomonis S, Sen CN, Sekhar LN: Cavernous sinus tumors: Neuroradiologic and neurosurgical considerations on 150 operated cases. J Neurosurg Sci 36:183-196, 1992 5. Linskey ME, Sekhar LN, Sen C: Cerebral revascularization in cranial base surgery. In: Sekhar LN, Janecka IP (eds): Surgery of Cranial Base Tumors, New York, Raven Press, 1993, p 45 6. Sekhar LN, Sen CN, Jho HD: Saphenous vein graft bypass of the cavernous internal carotid artery. J Neurosurg 72:35-41, 1990 7. Sekhar LN, Ross DA, Sen C: Cavernous sinus and sphenocavernous neoplasms: Anatomy and surgery. In: Sekhar LN, Janecka IP (eds): Surgery of Cranial Base Tumors, New York, Raven Press, 1993, p 521

INTERNAL CAROTID ARTERY SACRIFICE-LAWTON, SPETZLER 8. Sen C, Sekhar LN: Direct vein graft reconstruction of the cavernous, petrous, and upper cervical internal carotid artery: Lessons learned from 30 cases. Neurosurgery 30:732-743, 1992 9. Jennett B, Bond M: Assessment of outcome after severe brain damage. A practical scale. Lancet 1:480-484, 1975 10. Spetzler RF, Roski RA, Rhodes RS, Modic MT: The "bonnett bypass." Case report. J Neurosurg 53:707-709, 1980 11. Adegbite AB, Khan MI, Paine KW, Tan LK: The recurrence of intracranial meningiomas after surgical treatment. J Neurosurg 58:51-56, 1983 12. Lawton MT, Hamilton MG, Morcos JJ, Spetzler RF: Revascularization and aneurysm surgery: Current techniques, indications, and outcome. Neurosurgery 38:83-94, 1996 13. Huvos AG, Leaming RH, Moore OS: Clinicopathologic study of the resected carotid artery. An analysis of sixty-four cases. Am J Surg 126:570-574, 1973 14. Kennedy JT, Krause CJ, Loevy S: The importance of tumor attachment to the carotid artery. Arch Otolaryngol 103:70-73, 1977 15. McCready RA, Miller SK, Hamaker RC, Singer MI, Herod GT: What is the role of carotid arterial resection in the management of advanced cervical cancer? J Vasc Surg 10:274-280, 1989 16. Suarez CN, Estervan Solano JM, Buron Martinez G, Fuente Martin E, Mendez Colunga JC, Abril Garcia A: Invasion of the carotid artery in tumours of the head and neck. Clin Otolaryngol 6:29-37, 1981 17. Hiranandani LH: The management of cervical metastasis in head and neck cancers. J Laryngol Otol 85:1097-1126, 1971 18. Al-Mefty 0: Management of the cavernous sinus and carotid siphon. Otolaryngol Clin North Am 24:1523-1533, 1991

19. Prasad S, Janecka IP: Efficacy of surgical treatments for squamous cell carcinoma of the temporal bone: A literature review. Otolaryngol Head Neck Surg 110:270-280, 1994 20. Brennan JA, Jafek BW: Elective carotid artery resection for advanced squamous cell carcinoma of the neck. Laryngoscope 104:259-263, 1994 21. Tarr RW, Jungreis CA, Horton JA, Pentheny S., SekharLN, Sen C: Complications of preoperative balloon test occlusion of the internal carotid arteries: Experience in 300 cases. Skull Base Surg 1:240-244, 1991 22. Origitano TC, Al-Mefty 0, Leonetti JP, DeMonte F, Reichman OH: Vascular considerations and complications in cranial base surgery. Neurosurgery 35:351-363, 1994 23. McIvor NP, Willinsky RA, TerBrugge KG, Rutka JA, Freeman JL: Validity of test occlusion studies prior to internal carotid artery sacrifice. Head Neck 16:11-16, 1994 24. Drake CG, Peerless SJ, Ferguson GG: Hunterian proximal arterial occlusion for giant aneurysms of the carotid circulation. J Neurosurg 81:656-665, 1994 25. Lesoin F, Jomin M: Direct microsurgical approach to intracavernous tumors. Surg Neurol 28:17-22, 1987 26. Marks SM, Whitwell HL, Lye RH: Recurrence of meningiomas after operation. Surg Neurol 25:436-440, 1986 27. Mirimanoff RO, Dosoretz DE, Linggood RM, Ojemann RG, Martuza RL: Meningioma: Analysis of recurrence and progression following neurosurgical resection. J Neurosurg 62:18-24, 1985 28. Simpson D: The recurrence of intracranial meningiomas after surgical treatment. J Neurol Neurosurg Psychiat 20:22-39, 1957

123