Reconsideration of Cavernous Sinus Surgeries
Liang Chen,Xiang Huang,Ying Mao,Liangfu Zhou
DOI: https://doi.org/10.1227/neu.0000000000000402
2014-01-01
Abstract:The optimal therapeutic modality for tumors in or around the cavernous sinus (CS) is one of the most debatable issues among neurosurgeons, otolaryngologists, ophthalmologists, and radiotherapy experts. This debate, however, is not bad. Thanks to the advancement of microneurosurgery and endoscopic techniques and the progression of anatomic knowledge, the CS is no longer a formidable place. We now have several approaches to enter the CS and remove tumor. Radiotherapy methods and radiosurgeries have also made great progress. The optimal therapy not only should achieve long-term tumor control but also should relieve or at least not aggravate preoperative symptoms. So far, there are no randomized trials to make comparisons among the various therapeutic strategies. Our consideration is based on literature review and our 30 years of experience with surgery and 20 years of experience with stereotactic radiosurgery for CS tumors. HISTORY OF CS SURGERIES It has been almost 280 years since Winslow1 christened this small lateral sellar compartment as the CS, presumably thinking that it would resemble the corpora cavernosa of the penis. In the following 200 years, no one looked into this small, mysterious, extradural compartment again. In 1967, Parkinson2 first reported a case of long-standing arteriovenous fistula in which the engorged and thickened “arterialized” vein was readily noted to be neither a CS nor a dural sinus but a plexus of veins. The presence of cranial nerves (CNs) III, IV, and V at the capsule of the sinus, in addition to CN VI, the cavernous carotid artery, and the surrounding carotid sympathetics in CS, made this location a challenging one in which to work.3,4 Thus, surgical access to the CS has been thought to be rather difficult and risky because of its deep location and complex neurovascular structures. From the 1970s to the 1980s, many pioneer doctors carefully dissected the CS, locating >10 triangles of anatomic landmarks for safely entering the CS.5-11 One of the great findings was that the CS is composed of venous plexus in the middle fossa extradural space and that the outer layer of the overlying dural structure could be peeled off, offering an interlayer approach to expose the CS. This is the so-called middle fossa epidural approach. The main advantages of this approach included a wide exposure of the CS with clear anatomic orientation and minimal invasion of subdural tissues. The greatly improved safety of open surgery encouraged neurosurgeons to perform tumor resection more aggressively and completely. One way was to perform cerebral revascularization before tumor resection with the encased segment of the carotid artery for intended cure of the tumors. Unfortunately, the outcome was not good. For advanced head and neck malignancies, a recent review from the Barrow Institute indicated a high mortality of 11.1% and morbidity of 50%. The median survival was only 8 months. This surgical method is no longer advocated for skull base malignancies.12 For benign meningiomas involving the CS, total resection rates varied from 12% to 80%, with tumor control rates reported at 86% to 90%.13-16 However, operative mortality in some cases was as high as 7.3%, with complications ranging up to 59.6%.17-19 Thus, despite all the advances and developments that optimized current surgical techniques, significant morbidity and mortality remain. In contrast, Gamma Knife surgery (GKS) has offered impressive local control rates from 91% to 100%, with complications rates ranging from 0 to 7%.20-28 With long-term follow-up, GKS for cavernous meningiomas results in up to a 93.1% local control rate29 (Table 1). A recent report of 100 cases with GKS showed 1-, 5-, and 10-year tumor control rates of 98.9%, 94.2%, and 91.6%, respectively. The morbidity was 6%, including hypopsia, diplopia, hypopituitarism, and brain edema.30 For recurrence, operation or repeat GKS can be performed. The more recently used fractionated proton radiotherapy has also demonstrated better tumor control. A report from Loma Linda University Medical Center involved 72 CS meningiomas with a mean tumor volume of 27.6 cm3. One prior operation was performed in 30 cases, and 2 operations were performed in 20 cases. The mean 5-year tumor control rate was 96%, with 99% for tumors of World Health Organization grade I or no pathology and 50% for tumors of World Health Organization grade II or higher.31 For CS hemangioma (CSH), GKS seems to have better results than surgery, although long-term results are not available32-39 (Table 2).TABLE 1: Different Treatment Outcomes of Cavernous Sinus MeningiomaaTABLE 2: Different Treatment Outcomes of Cavernous Sinus HemangiomaaThus, a limited tumor resection has been advocated recently. However, what extent of tumor resection would be best for patient outcome has not been determined. In our opinion, maximal safe tumor resection, a rationale often used for glioma surgery in functional brain area, should also be advocated in CS tumor resection. This is determined largely by the tumor origin and its growth pattern. ORIGIN AND GROWTH PATTERN OF CS TUMORS Our experience comes from >400 cases of trigeminal schwannomas, meningiomas, pituitary adenomas, metastasis, hemangiomas, chondromas, chordomas, lymphomas, and sarcomas. They have different origins and growth patterns. Neuropathological study discovered that the schwannoma of the middle fossa expands in the potential interspace of the lateral wall of the CS, mostly from the trigeminal nerve and with a tendency to grow into the posterior fossa via the Meckel cave and extracranial extension through the foramen rotundum and ovale, or it tends to the orbit through the superior orbital fissure.40-43 Thus, total tumor resection with minimal nerve injury is achievable after circumambient dissection of the tumor within the interspace of the CS. Pure CS meningiomas are tumors growing from the inner layer of the CS dural wall, not invading from outside. These kinds of meningiomas have intrinsic preference for spreading along and penetrating across the membrane layer, so they often invade into the substantive space of the CS; encase the internal carotid artery (ICA) and CN VI; fill the nerve bundle of CNs III, IV, and V in the lateral wall; and sometimes have subdural extension. Blindly pursued tumor resection might put the nerves at high risk of unexpected injury. Thus, maximal safe tumor resection and optic decompression followed by radiosurgery are recommended. For meningiomas initially outside the CS such as a sphenoid ridge meningioma and invading the CS, resection of the tumor outside the CS while sparing the part in the CS is reasonable and advisable. Unlike these 2 tumors, the hemangioma derives from the matrix of the CS and has a specific growth pattern. Typically, the CSH grows along the venous pathway, expands in the CS, and then extends to the sella turcica via the intercavernous sinus. Rarely, the tumor might protrude into the orbit via the communicating venous channels through the superior orbital fissure. The intracavernous structures like the ICA and CN VI are often encapsulated by the tumor, and the lateral wall of the CS bulges outward, stretching the CNs on the surface of the tumor; ie, CN III is distorted upward and CNs IV and V are distorted laterally. The extradural location of the CSH suggests the rationality of the extradural transcavernous route to remove the tumor by following its pathway. The other tumors include osteogenic tumors from beneath the CS, pituitary adenomas from the medial side of the CS, and tumors that invaded from the temporal lobe. SURGICAL ROUTES TO THE CS There are generally 3 routes to enter the CS. The middle fossa epidural approach first exposes the lateral wall and its intradural space and then enters the CS through the lateral wall or goes beneath the CS after petrous apex drilling. Spaces medial to the CS are exposed last. The entire horizontal portion of the petrous carotid is surrounded by a periosteal sheath that has to be opened to expose the artery. A venous plexus often surrounds the artery inside the sheath. So this approach, via the Glasscock triangle,11 provides the best exposure of the lateral surface of the posterior vertical and horizontal segments of the ICA, the posterior bend, and the lower part of the anterior bend. It also provides the best visualization of the origin of the meningohypophyseal trunk (MHT) and the artery of the inferior CS.44 A subtemporal or transsylvian intradural approach first exposes the upper wall of the CS. We can go into the anterior medial CS through the Dolenc triangle6 or to the medial side of the CS via the Hakuba triangle.8 These approaches allow good access to the superior space of the CS. The lateral wall of the CS can be exposed and entered via the Parkinson triangle.5 However, for full and clear exposure of the CS and its contents, the intradural approach is inferior to the epidural approach. The transsphenoidal approach first goes to the medial side of the CS and then can enter the CS from the medial side. After removal of the mucosa and bone in the lateral wall of the sphenoid sinus, the dural that forms the medial wall of the CS will be exposed, which is removed to expose the medial venous space and the junction of the CS with the basilar sinus. Medial retraction of the cavernous carotid provides exposure of the lateral venous space and access to the origin of the artery of the inferior CS and the abducent nerve.45,46 Thus, the transsphenoidal approach may be used to expose the medial venous space, the anterior and medial surfaces of the anterior bend, and the medial surface of the horizontal segment.47 However, hemostasis is still a problem for this approach. Thus, for tumors in the lateral wall of the CS such as schwannomas and epidermoid cysts, the middle fossa epidural approach should be the primary choice of surgical route. For the intra-CS tumors such as the hemangioma, the middle fossa epidural approach is also optimal. For tumors beneath the CS, especially at the posterior side, the transpetrous apex middle fossa approach can be adopted. However, for tumors medial to the CS, the subtemporal or transsylvian intradural approach or the transsphenoidal approach can be used. For tumors communicating the intradural and extradural CS space, tumor resection should be from the side where the tumor enters the CS. CS HEMANGIOMA With an extradural approach, the CSH is exposed and removed through the interspace of the CN III, CN IV, and 3 divisions of CN V. Because hemostasis remains the main difficulty in tumor removal and bleeding is often uncontrollable with a bipolar coagulator, various methods have been developed to reduce blood loss during operation, including preoperative embolization48,49 and balloon test occlusion,48 intraoperative induced hypotension32,50,51 and hypothermia,52 and intratumoral injection of adhesive material.53,54 The ultrasonic surgical aspirator has hemostatic effects on small vessels, thus facilitating intracapsular removal of a CSH.52 However, Kudo et al55 reported a case with ICA laceration during the application of the instrument. More recently, radiofrequency thermocoagulation has also been used to facilitate surgical excision.56 However, life-threatening bleeding might still occur before blood supply from the ICA is secured. Digital subtraction angiography has revealed that the blood supply of the tumor is primarily from branches of the cavernous segment of the ICA such as the MHT and inferior lateral trunk, and after the tumor grows, there will be blood supply from external carotid artery branches such as the middle meningeal artery or accessory middle meningeal artery.48 For early control of the tumor blood supply, an early medial dissection to reach the ICA has been advocated. This would not be difficult for a small tumor. For a large tumor, however, medial dissection seems to be difficult or impossible, especially via the intradural approach. With an epidural approach, the middle meningeal artery could be interrupted at the foramen spinum, and dissection of the lateral wall of the CS would further minimize blood supply from the external carotid artery. Questions exist as to whether the optimal method of tumor resection is piecemeal or en bloc.34 This might be decided by the conditions we encounter during the operation. CSHs could be divided into subtypes of sponge-like hemangiomas or mulberry-like hemangiomas.32 For a sponge-like CSH, uncontrollable bleeding might occur if we incise into the tumor. Fortunately, this kind of tumor is soft and can be compressed to get to the surrounding dissection space. Bipolar shrinkage of the capsule of the tumor is also useful. The MHT can be reached after medial dissection of the capsule between CNs IV and VI and sometimes between CNs VI and VII.34 Then a medial-to-lateral dissection would be easy. For a mulberry-like CSH, if the tumor is very large, it might be very difficult to reach the MHT if the tumor is not soft enough, and excessive retraction might hurt the surrounding CNs. In this situation, the operator should be well oriented and cut into and piecemeal resect the tumor to reach the MHT as quickly as possible.34 Another problem is the frequently occurring iatrogenic abducens nerve damage. CN VI penetrates through the CS and is often encased by the tumor.57 If that is the case, it could hardly be functionally preserved. After operation, diplopia fails to improve because of abducens palsy. However, in cases when CN VI is not visualized, the nerve could be well preserved with intact eyeball movement such as in the case reported by Sepehrnia et al.58 Suri et al34 reported only 1 permanent abducens nerve palsy in 7 cases. This indicates that CN VI is not always encased by the tumor. To the best of our knowledge, the abducens nerve goes inside the posterior part of the CS and is vulnerable to being encased by the tumor. However, if the tumor derives from the anterior part of the CS, the CN VI is more likely to be compressed by the tumor. We cannot anticipate this situation preoperatively. CONCLUSION For surgery of the CS, the first priority is to avoid aggravation of preoperative disabilities. Advancing dissection along the membrane plate can best protect the neurovascular structures underneath. However, when the dissection membrane has been destroyed, one should identify the ICA and CNs early and then remove the tumor. For nerve sheath tumors, hemangiomas, chordomas, and some meningiomas, it is important to choose the optimal approach. For malignant invasive tumors, decompressive resection followed by more intensive multiple therapy should be used. Maximal safe tumor resection, the rationale often used for glioma surgery in functional brain area, should also be advocated in CS tumor resection. Disclosure The authors have no personal, financial, or institutional interests in any of the drugs, materials, or devices described in this article. Acknowledgments We thank the surgeons and pathologists in the Department of Neurosurgery at Huashan Hospital. We also thank Dr Gerald Grant, MD, for inviting us to write this paper for Clinical Neurosurgery. The study was presented in part at the 2013 CNS Annual Meeting in San Francisco, California.