Department of Neurosurgery, University of Cincinnati College of Medicine, and Mayfield Clinic, Cincinnati, Ohio
METHODS: In this retrospective review, 39 patients, including 27 women and 12 men ranging in age from 24 to 73 years (median, 48 yr), underwent surgical treatment for this condition. Completeness of tumor resection, cranial nerve morbidity, general morbidity, and long-term outcome were studied. The cavernous internal carotid artery was partially encased in 15 patients, totally encased in 11 patients, and narrowed by tumor in 13 patients.
RESULTS: Of eight patients who underwent complete tumor resection, seven had partial encasement of the internal carotid artery. Of 31 patients who underwent subtotal resection, 11 underwent postoperative radiotherapy. There were no deaths in the series. Morbidity was 17.9% for cranial nerves controlling extraocular motor function. Trigeminal nerve function did not improve after surgical treatment. The median follow-up period was 2 years (range, 6 mo-5.3 yr). Symptomatic and radiographic recurrence occurred in two patients who underwent complete tumor resection and in two patients who underwent subtotal resection.
CONCLUSION: Based on our findings and a review of the literature, we conclude the following: 1) the resectability of meningiomas of the cavernous sinus depends on the degree of internal carotid artery involvement; 2) total excision of cavernous sinus meningiomas is possible but rarely achieved in holocavernous meningiomas; 3) cranial nerve morbidity is significant; and 4) subtotal excision with or without postoperative radiotherapy is an effective short-term oncological strategy.
(Neurosurgery 40:238-247, 1997)
Key words: Cavernous sinus, Complication, Cranial nerve, Meningioma, Outcome, Radiotherapy
Since Cushing and Eisenhardt (8) first published their treatise on meningiomas, tremendous advances have been made in all aspects of the management of meningiomas. However, meningiomas arising at the cranial base, particularly those involving the cavernous sinus region, still present a formidable surgical challenge. There has been a general reluctance to explore the cavernous sinus for fear of causing uncontrollable hemorrhage, internal carotid artery (ICA) injury, and cranial nerve injury. New insights into the microsurgical anatomy of the cavernous sinus together with improvements in imaging and anesthesia have led several groups to explore the cavernous sinus, attempting radical tumor excision and repair of vascular lesions (3, 6, 7, 9-12, 16, 18-21, 24-26, 31-38, 41, 43, 44). In experienced surgical hands, uncontrollable hemorrhage and ICA injury are rare.
For treatment of cavernous sinus meningiomas, several questions remain unanswered. Are meningiomas of the cavernous sinus curable? Is the cranial nerve morbidity associated with cavernous sinus meningioma resection acceptable? What are the indications for partial resection? Recent reports (9, 12, 17, 24, 32, 34-37) have presented an optimistic outlook for patients harboring this pathological abnormality, but at least one senior neurosurgeon (30) has expressed his reservations concerning the morbidity associated with aggressive surgical treatment of tumors that affect the cranial base.
Our retrospective review of 39 patients with meningiomas involving the cavernous sinus who underwent surgical treatment at our institution from 1985 to 1994 differs from recently reported results in terms of resectability and morbidity. We outline the difficulties that our cranial base team faced in treating these tumors to continue the debate regarding the proper indications for cavernous sinus surgery.
The intracavernous component of the tumor was classified according to the criteria of Hirsch et al. (17) (Table 1). The ICA involvement was Grade 1 in 15 patients, Grade 2 in 11, and Grade 3 in 13 (Fig. 1). Serial postoperative MRI scans were assessed for degree of tumor resection (Table 2) (9), rate of tumor recurrence, and extent of tumor progression. When a discrepancy occurred between the surgical impression and the radiological study, the latter was used.
TABLE 1. Radiological Grade of Cavernous Sinus Tumors as Determined by Magnetic Resonance Imaging and Angiography According to the Criteria of Hirsch et al. (17)
|1||Touch or partially encircle the cavernous internal carotid artery|
|2||Completely encircle but do not narrow the lumen of the internal carotid artery|
|3||Encircle and narrow the lumen of the cavernous internal carotid artery|
TABLE 2. Extent of Tumor Removal: A Grading System According to De Monte et al. (9)
|1||Complete microscopic removal of tumor and its dural attachment with any abnormal bone|
|2||Complete microscopic removal of tumor with diathermy coagulation of its dural attachment|
|3a||Complete microscopic removal of intra- and extradural tumor without resection or coagulation of its dural attachment|
|3b||Complete microscopic removal of intradural tumor without resection or coagulation of its dural attachment or any extradural tumor|
|4a||Intentional subtotal removal to preserve cranial nerves or blood vessels; complete microscopic removal of tumor dural attachment|
|4b||Partial removal leaving tumor <10% in volume|
|5||Partial removal leaving >10% in volume or decompression with or without biopsy|
Of 28 patients who had extensive extracavernous tumor involvement, none underwent total resection. This was often predetermined by the surgical strategy, which focused on decompression and preservation of the brain and brain stem as well as "containment" of the meningioma to the cavernous sinus to preserve existing cranial nerve function. Of these 28 patients, 13 underwent De Monte Grade 4b resection (Fig. 3), leaving less than 10% tumor volume, and 15 underwent De Monte Grade 5 resection, leaving more than 10% of tumor volume. Twenty-three of these patients began with tumors graded either Hirsch Grade 2 (ICA encased) or Grade 3 (ICA narrowed) with cavernous sinus involvement. The resectability of the cavernous sinus portion of the tumor and the cranial nerve morbidity related more closely to the Hirsch grade of the cavernous portion than to the extracavernous portion; thus, resectability did not differ for tumors confined to an expanded cavernous sinus. Noncranial nerve morbidity related more closely to the extracavernous extent and preoperative brain stem compression. Ten patients who underwent subtotal resection elected postoperative radiotherapy; six of these patients underwent stereotactic radiosurgery, receiving a mean dose to the tumor of 1200 cGy, and four underwent external beam radiation, receiving a mean dose of 5000 cGy.
Of 17 patients who had preoperative optic nerve dysfunction, 4 (24%) experienced significant improvement. Of 17 patients with normal preoperative optic nerve function, only 1 (6%) suffered a new visual deficit. Of the 24 extraocular nerves with preoperative deficits, only 5 (21%) improved. Of the 78 extraocular nerves that were normal before surgery, new deficits developed in 14 (18%). In functional terms, a significant new permanent deficit occurred in 7 of 10 patients who had normal preoperative oculomotor functions. However, only 2 of 17 patients with preoperative oculomotor dysfunction improved. Preoperative trigeminal deficits never improved after surgery; of 97 trigeminal branches that were normal before surgery, new deficits developed in 8 (8%). Considering that a total of 192 nerves (optic, extraocular, and trigeminal) with normal preoperative function were at risk, 23 new cranial nerve deficits developed after cavernous sinus surgery for meningioma, constituting an overall 12% risk of new cranial nerve morbidity.
In two of the eight patients who underwent total resection, tumor recurrence that was revealed symptomatically and radiologically developed at 16 and 18 months after surgery. Histological findings revealed a hemangiopericytoma in one patient and atypical features with increased mitoses in the other; these findings may be irrelevant to the tumor recurrence. Both patients underwent stereotactic radiosurgery for the recurrence and remain clinically and radiographically stable at 13 and 9 months after recurrence, respectively. The recurrence rate of benign meningiomas after total resection (De Monte Grade 3a or higher) remains 0%. If the two meningiomas with atypical features are included, the recurrence rate is 25%.
TABLE 3. Preoperative and Postoperative Cranial Nerve Function in 26 Patients Who Underwent Subtotal Resection and in 8 Patients Who Underwent Total Resection of Meningiomas Involving the Cavernous Sinusa
|Cranial Nerves||Preoperative Function
aIn five cases, the data were insufficient.
At last follow-up, Karnofsky scores equaled or exceeded preoperative scores in 35 of the 38 surviving patients. Eighteen months after undergoing surgery, one patient who required a postoperative ventriculoperitoneal shunt suffered a shunt malfunction with herniation before revision and now persists in a vegetative state. Two patients have Karnofsky scores of 70.
In review of the neurosurgical literature, 10 authors have reported their results of surgical treatment for 252 cases of meningiomas involving the cavernous sinus (3, 6, 7, 9, 12, 16, 20, 24, 31, 35, 36, 38, 44). For those authors who reported their results sequentially, only the most recent report is included. A study by Kleinpeter and Bock (21) was excluded, because it summarized the results for all tumors, including those outside the cavernous sinus region.
TABLE 4. Summary of Reported Cases of Meningiomas of the Cavernous Sinus
|Series (Ref. No.)||No. of Cases||No. of Total Resections (%)||Follow-up (mo)|
|Cioffi et al., 1987 (6)||12||1 (8.3)||24|
|De Monte et al., 1994 (9)a||41||31 (76)||45|
|Dolenc et al., 1987 (12)||40||Unknown||Unknown|
|Hakuba et al., 1989 (16)||4||Unknown|
|Kawase et al., 1987 (20)||9||7 (77.8)||Unknown|
|Lesoin and Jomin, 1987 (25)||16||0||Unknown|
|Risi et al., 1994 (31)||15||4 (26.7)||Unknown|
|Sekhar et al., 1992 (36)a||70||61 (87.1)||36|
|Sepehrnia et al., 1991 (38)||36||18 (50)|
|Von Wild and Eskinja, 1989 (44)||9||4 (44.4)||10|
|O'Sullivan et al. (present study)||39||8 (20.5)||24|
aRefers to most recent reports by authors and included previously reported data.
The most common complication of cavernous sinus surgery is extraocular paralysis. We have only two categories for extraocular nerve function, normal and deficit, because partial function of Cranial Nerve III still results in diplopia. In 34 patients with sufficient follow-up to determine the final outcome of extraocular function, 102 extraocular nerves were at risk. Of 78 extraocular nerves normal preoperatively, 5 (21%) improved and 14 (18%) deteriorated. Hirsch et al. (17) reported that 5 of 13 patients with poor extraocular motility preoperatively improved postoperatively, although the degree of improvement was not defined. Of 41 patients with good or excellent preoperative function, 31 (74%) eventually had good or excellent results postoperatively, implying a 26% rate of functional deterioration. De Monte et al. (9) reported no improvement in patients with poor preoperative extraocular motility. Four patients with fair preoperative function experienced improvement, all patients with excellent preoperative function remained unchanged, and one patient's function deteriorated from good to fair. The rate of extraocular deterioration is largely determined by the aggressiveness of the surgical strategy. Even with a conservative strategy such as ours, the extraocular morbidity remains significant and should weigh heavily against surgery in many cases.
The trigeminal nerve is often overlooked in consideration of cavernous sinus surgery. In 34 patients with sufficient data for analysis, 102 trigeminal branches were at risk. Of 97 trigeminal branches that were normal before surgery, eight (8%) incurred new permanent deficits (Table 3). Hirsch et al. (17) reported that neurotrophic keratitis developed in 14 (21.5%) of 65 patients, but they did not elaborate. De Monte et al. (9) reported that of 31 patients, 6 developed new trigeminal neuropathies and 1 developed neurotrophic keratitis.
General morbidity is significant, but the worst complications (i.e., hemiparesis) tend to relate to the extracavernous extension of the tumor and the need to decompress the brain stem justifies the incidence of serious complications. In our study, five (12.8%) patients suffered permanent hemiparesis. Sekhar et al. (36) summarized the complications for all tumors, but it is impossible to determine those attributable to meningioma surgery. De Monte et al. (9) reported the expected types of complications, including stroke (two cases), cerebrospinal fluid leak (two cases), and transient pituitary dysfunction (three cases). Although there were no deaths in our series, others have reported mortality in 1 of 70 patients (36), 3 of 41 patients (9), 1 of 15 patients (31), and no patients (6, 25). These results confirm the low mortality rates in experienced hands.
Short-term recurrence data after radiosurgery are available in the literature. Duma et al. (13), reporting their results for stereotactic radiosurgery of cavernous sinus meningiomas in 34 patients, observed that all patients showed tumor control and 56% had tumor regression at median follow-up of 26 months. Three patients had permanent morbidity, including two with optic neuropathy and one in whom a preexisting ptosis worsened. Recent advances in radiation technology and imaging continue to reduce the associated morbidity (27). Many studies have shown that postoperative radiotherapy prevents or delays the regrowth of incompletely resected meningiomas (4, 14-16, 22, 27, 40, 42). Maroon et al. (28) and Wilson (45) have reported successful strategies combining surgery and radiotherapy or radiosurgery. These results are comparable with the short-term results in our series. However, the true comparability of these modalities alone or in combination can be assessed only in long-term follow-up.
Based on our experience, we recommend consideration of the following indications for surgery: progressive optic neuropathy, progressive extraocular paresis, young patients with Hirsch Grade 1 tumors, or cytoreductive surgery in preparation for radiosurgery. In a patient with a significant extracavernous portion, surgery should be directed at the extracavernous portion of the tumor for brain and brain stem decompression. The surgical strategy should be designed to "contain" the meningioma within the cavernous sinus in patients whose cavernous sinus involvement is Hirsch Grade 2 or 3 and to consider total resection of the cavernous sinus component in Hirsch Grade 1 tumors, especially in young patients.
The field of cranial base surgery and, in particular, surgery of the cavernous sinus is just evolving. These results, recommendations, and opinions are presented in the hope of contributing to an ongoing debate of appropriate surgical indications.
Michael O'Sullivan was supported by an Ainsworth Scholarship from University College Cork and by a Glaxo Travelling Grant from the Royal College of Surgeons in Ireland.
Received, October 13, 1995.
Accepted, September 6, 1996.
Reprint requests: Harry R. van Loveren, M.D., Editorial Office, Department of Neurosurgery, University of Cincinnati College of Medicine, 231 Bethesda Avenue, P.O. Box 670515, Cincinnati, OH, 45267-0515.
1) The authors are correct in their conclusion that some meningiomas involving the CS infiltrate the wall of the internal carotid artery (ICA) and infiltrate cranial nerves. The former was well established in a study by Shaffrey et al. (Shaffrey M, Dolenc V, unpublished data) and another study by our group (5). We see this repeatedly in the intracavernous ICA specimens that we send for pathological examination. The infiltration of cranial nerves by tumors was reported by the Cincinnati group (6) and has also been observed by us in a few patients.
2) We do not use the same grading systems for tumors and resections as that used by O'Sullivan et al. Although the senior author was instrumental in developing the gradation system described by Hirsch et al. (4), the version used in this study was constricted. We prefer to classify tumors invading the CS as extensive (>3 cm, extending to the multiple areas of the cranial base) or confined (<3 cm, involving the CS and surrounding areas) based on tumor extent and Grades I through V based on the extend of CS invasion, the involvement of the intracavernous ICA, and the involvement of the contralateral CS. We think that the distinction between Grades I and II is important, because Grade I tumors can be removed with a small opening of the CS whereas Grade II lesions often require a more extensive opening. The distinction between Grades III and IV is based on the surgical observation that the intracavernous ICA can be dissected free of tumor in some patients with Grade III meningiomas but none of the patients with Grade IV lesions. Grade V lesions involve both CSs. When the invasion of the contralateral CS is limited, the tumor can be resected totally. However, total resection is not attempted when there is extensive invasion of the contralateral CS. We also prefer a different categorization of the degree of resection: gross total, incomplete, or biopsy only. Gross total resection is defined as the removal of all of the visible tumor, including dural attachments and, if invaded, the intracavernous ICA. Gross total resection is based on the surgeon's intraoperative assessment and postoperative magnetic resonance imaging (MRI). Anything less than this is defined as incomplete resection. We recognize that total cytological resection of basal meningiomas is usually impossible, because the cytopathological changes extend well beyond the resection area.
3) We agree with O'Sullivan et al. about the cranial nerve morbidity of resecting meningiomas invading the CS. The authors acknowledge that even conservative strategies have significant cranial nerve morbidity. It seems that the morbidity of partial resections and gross total resection is similar. For extensive tumors, the cranial nerve morbidity is the same whether the intracavernous tumor is resected. The occurrence of neurotrophic keratitis after CS resection was a real problem in our prior experience and was most likely due to a combination of corneal numbness (V1 neuropathy) and dryness (section of the greater superficial petrosal nerve). This has been greatly minimized by exposing the cervical rather than the petrous ICA for proximal control, which avoids sectioning the greater superficial petrosal nerve. Also, in patients who have only Grade I or II tumors, I often avoid actually exposing the proximal carotid artery but keep the neck prepared in a sterile fashion. The listing of individual cranial nerve neuropathies is interesting but actually underestimates the functional deficit. This is why we have used the system developed by Biglan et al. (1), which is a classification of binocular function. In this context, the surgeon should not overlook that patients with minor deficits of extraocular motility can often experience considerable improvement with minor eye muscle surgery.
4) We do not share the pessimism expressed by O'Sullivan et al. concerning resectability. There seems to be a fundamental difference in surgical technique. The Cincinnati team presumably uses the Dolenc technique. We use a technique that is somewhat different in many respects and very different in terms of the management of the intracavernous ICA. If the tumor cannot be easily dissected from the intracavernous ICA, we replace it with a saphenous vein graft bypass. The surgeon can then work in the CS without worrying about ICA injury. This allows the surgeon to concentrate on the cranial nerves, the pituitary gland, and achieve a complete resection, including the involved dura. To minimize the ischemia time, we have been using predominantly long saphenous vein bypass grafts, from the cervical external carotid artery or ICA to the M2 segment of the middle cerebral artery. With the use of intraoperative angiography, the patency rate of this bypass is approximately 98%. Of patients with purely intracavernous lesions, none has suffered a stroke with disability. We rarely expose the petrous ICA, because the ICA in the neck is much easier to expose and the greater superficial petrosal nerve does not need to be divided, which reduces the corneal morbidity in patients with trigeminal denervation. I think that Dolenc has also modified his original technique to perform more carotid resections for unresectable tumors with vein graft replacements of the artery. We now have much longer follow-up of patients who have undergone surgery for intracavernous meningiomas (3). Using life-table analysis, we found that the 5-year recurrence-free survival for completely resected tumors was 80.7% and the progression-free survival for incompletely resected tumors was 61.5%. Interestingly, total resection was accomplished in 16 of the 20 patients with saphenous vein graft bypass. There have been no recurrences among those who underwent total resection. Of the four incompletely resected tumors, two have progressed, both of which are biologically aggressive tumors.
5) Despite these results, we agree that the extent of resection needs to be moderated in some patients because of the invasion of cranial nerves by tumor, the age of the patient, or the patient's wishes. In such cases, we have treated the tumor remnants with radiosurgery. However, the efficacy of this strategy is unknown at present. We have operated on patients with lesions previously treated by external beam radiotherapy and a few patients treated previously with radiosurgery, and in these cases, the operations are much more difficult and the complications and hospital costs are much higher. Radiotherapy undoubtedly controls the growth of a number of lesions, but when lesions do recur, the patient and the surgeon pay the price for the past treatment modality.
6) One of the key elements that has been lacking in all of our treatments to date is an attack on the biological mechanisms of tumor replication. An effective nonsurgical means of treating meningiomas would dramatically change our management of these lesions and should be the goal.
This article presents a thorough analysis of meningiomas involving the CS. It is very important to distinguish, as the authors do, between meningiomas that are confined to the CS and those that involve the CS secondarily. I emphasize the authors' conclusion that the degree of involvement of the ICA, which is also an indicator of the growth behavior of the tumor, is the major factor in determining the resectability of the meningioma. A meningioma that infiltrates every surrounding structure will always be very difficult to cure regardless of its location and will be associated with considerable morbidity even if a conservative surgical approach is chosen. I agree with the authors that a more aggressive approach to meningiomas of the CS will lead to more surgical morbidity without the guarantee of a cure. A tumor remnant in the CS does not inflict a major risk for a patient. I have followed a large number of patients with known tumor residuals in the CS for years without radiological or neurological evidence of progression. Why should we risk a carotid artery or eye function for a lesion that hardly ever endangers the patient's life as long as those tumor parts involving other structures, such as the brain stem, are resected and controlled? Authors recommending aggressive surgical strategies still have to prove that such strategies improve long-term outcomes.
This report from the University of Cincinnati Skull Base Center provides an excellent summary of the surgical management of CS meningiomas. I agree with the approach the authors outline. As they note, the optimal management for patients with these tumors remains controversial.
The authors define a program of surgical management. They used radiosurgery in selected cases. The challenge for neurosurgeons in the future will be selecting the best treatment program for each patient, whether it be surgery, one of the radiation therapies, a combination of surgery and radiation therapy, or observation. The decision is often difficult, because the natural history is variable; some tumors grow very slowly, if at all, over prolonged periods of time, and long-term results of the newer surgical and radiation therapies are unknown.
The problem is illustrated in our series of 14 consecutive patients who had proven or presumed CS meningiomas and who have been followed for a minimum of 3 years. The tumors were confined to the CS or adjacent parasellar area and did not include those with large middle or posterior fossa growth, as illustrated in Figure 1, A and B, of the article. Three patients underwent surgery only, one with an apparent total removal and two with radical subtotal removal. Further tumor growth has not been demonstrated in these three patients. Two underwent surgery plus fractionated radiation therapy. Further tumor growth has not been demonstrated in these two patients. In the surgical group, one preoperative third nerve paresis was worse and one had no improvement in a preoperative trigeminal nerve deficit. There were no new deficits. Three patients were treated with fractionated radiation therapy based on MRI findings. Clinical status and follow-up MRI were unchanged. Six patients are being observed. These decisions were based on the patients' ages and/or minimal symptoms in five patients and the presence of an asymptomatic tumor in one patient. MRI revealed slight growth of the tumor and worsening third nerve palsy in one of the six patients who has refused treatment recommendation. The other five have shown no radiographic or clinical change. All 14 patients continue to be fully active at their previous levels of activity. Only long-term follow-up of patients treated with the currently available surgical and radiation therapies and those being observed without treatment will provide the information that will enable us to choose the optimum treatment.
The authors provide a detailed appraisal of their contemporary series of patients with meningiomas involving the CS. I agree that the results of CS meningioma surgery have probably been presented in overly optimistic terms. Meningiomas are an entirely different entity, with respect to cranial nerve morbidity, when compared with other benign neoplastic processes in the CS. Aggressive attempts at meningioma resection usually result in at least transient ocular motor palsies and trigeminal hypesthesia. Because of the complexity of the anatomy involved, the surgeon's perception of a "complete" resection is typically inaccurate. This is well demonstrated by the recurrence rates seen in short follow-up intervals in patients in whom resection was considered to be "total." I suspect that with longer follow-up periods, the recurrence rates will be seen to be much higher than originally expected without adjuvant radiation therapy.
I agree with the authors' indications for surgical treatment of CS meningiomas. Clearly, there are cases of CS meningiomas in younger patients that are separable from surrounding structures. This type of lesion may be excised completely, or nearly so. Rapidly progressive visual loss and/or ocular motility deficits are also clear indications for surgical decompression and an attempt at resection. The initial enthusiasm for aggressive resection of these lesions, generated by recent advances in cranial base techniques, is now tempered by the demonstrated risk of cranial nerve morbidity. However, surgical decompression has a primary role in offering the best chances for relief of symptoms and, coupled with radiation, long-term control of tumor growth. The definitive conclusions regarding the treatment of these difficult lesions will come with the continued routine handling of such patients in established cranial base centers.
Return to Neurosurgery Home Page