Cavernous sinus (CS) involvement by sinonasal and ventral skull base malignancies is infrequently encountered in neurosurgical practice. Despite advancements in skull base microneurosurgery and endoscopic techniques, detailed knowledge and experience of the surgical management of these lesions are limited. This article elaborates on surgical strategies and approaches for CS involvement of malignant ventral skull base tumors. The article discusses the indications, techniques, nuances, advantages, limitations, and complications of minimally invasive CS biopsy, transcranial microscopic, and transfacial endoscopic approaches to the CS using illustrative diagrams and operative videos. The principles and nuances of a high-flow cerebral revascularization procedure are mentioned.
Key points
- •
For a malignant tumor involving the cavernous sinus, the approach must be individualized to optimize the treatment strategy. Often a combination of surgical approaches is necessary for optimal resection of aggressive ventral skull base malignancies with cavernous sinus involvement, including frequent collaboration between the disciplines of otorhinolaryngology and neurosurgery.
- •
To confirm the diagnosis in cases of inconclusive radiological impressions or suspicious-looking lesions, use of minimally invasive approaches aided with stereotactic neuronavigation is helpful.
- •
Once the diagnosis is confirmed, and if the patient is healthy and has no metastatic disease, aggressive surgical resection may be indicated, especially if the removal of the cavernous sinus lesion may result in total tumor resection.
- •
For patients with more advanced and recurrent malignant disease, whereby carotid preservation would prevent a meaningful resection, en bloc resection of the tumor and cavernous sinus with cerebrovascular revascularization may be justified.
- •
Frequently, en bloc resection will require cavernous sinus exenteration, including sacrifice of the cavernous internal carotid artery, possible high-flow extracranial-to-intracranial bypass, and placement of a vascularized pedicled flap for ventral skull base reconstruction along with adjuvant chemotherapy and radiotherapy to effectively eradicate the microscopic tumor remnants.
Video content accompanies this article at http://www.oto.theclinics.com .
Introduction
Sinonasal and ventral skull base malignancy refers to tumors arising from the nasal cavity, paranasal sinuses, orbit, salivary glands, and soft tissue and bone along the ventral skull base. This broad spectrum of malignant diseases includes, but is not limited to, nasopharyngeal squamous cell carcinoma, adenoid cystic carcinoma, lymphoma, chordoma, chondrosarcoma, hemangiopericytoma, malignant meningioma, osteosarcoma, rhabdomyosarcoma, adenocarcinoma, mucoepidermoid carcinoma, acinic cell carcinoma, undifferentiated carcinoma, clear cell carcinoma, liposarcoma, and esthesioneuroblastoma. These lesions spread to the cavernous sinus (CS) either directly (83.2%) via the superior orbital fissure, inferior orbital fissure, foramen rotundum, and foramen ovale or through metastasis (16.8%) via perineural extension or hematogenous or lymphatic spread. Malignant tumors of the nasal cavity and paranasal sinuses are rare, accounting for 0.2% to 0.5% of all cancer cases and only 3.0% of malignant tumors in the head and neck region. According to the Surveillance, Epidemiology, and End Results database of all reported sinonasal malignancies between 1973 and 2006, the cumulative incidence of sinonasal malignancy was 0.556 cases per 100,000 population per year; the most common sites of origin were the nasal cavity (43.9%) and the maxillary sinus (35.9%). CS involvement by sinonasal and ventral skull base malignancies carries dismal prognosis overall, because it precludes radical oncological resection in many instances. However, the overall and progression-free survival in a particular malignant ventral skull base tumor case with CS involvement depends on many factors besides the extent of resection, which include tumor burden, pathologic conditions, presence of metastasis, positive tumor margins, and age of patients. Aggressive surgical resection followed by adjuvant chemotherapy and radiotherapy offers the best possible chance of prolonging overall survival in most sinonasal and ventral skull base malignancies.
Since the seminal articles by Ketcham and colleagues in 1963 and 1966 detailing the role of surgery in intracranial involvement of head and neck malignancy, innovation of radical surgical resection procedures for the eradication of malignant ventral skull base tumors is arguably one of the most important advancements in the treatment of head and neck malignancy in the past half century. Besides the use of conventional open transcranial and transfacial approaches to achieve the surgical goal of oncologic resection, minimally invasive endoscopic approaches have been applied more recently to selected patients with reasonable outcomes. It was not uncommon in the past that malignant skull base tumors were considered inoperable. With advancement in skull base microneurosurgery, availability of better endoscopic devices, expertise in cerebral revascularization, more effective hemostatic agents, and improved ventral skull base reconstruction techniques, the collective opinion of the neurosurgical community has changed from this conventional thinking. More aggressive groups preferred piecemeal removal of tumor, including parts extending into the CS proper ; however, the oncological principle of en bloc resection was still not realized until the pioneering efforts of Sekhar and Moller, Saito and colleagues, and others. This review discusses the current surgical strategies, their indications, techniques, nuances, advantages, limitations, and complications of operative approaches for CS involvement of malignant ventral skull base tumors.
Introduction
Sinonasal and ventral skull base malignancy refers to tumors arising from the nasal cavity, paranasal sinuses, orbit, salivary glands, and soft tissue and bone along the ventral skull base. This broad spectrum of malignant diseases includes, but is not limited to, nasopharyngeal squamous cell carcinoma, adenoid cystic carcinoma, lymphoma, chordoma, chondrosarcoma, hemangiopericytoma, malignant meningioma, osteosarcoma, rhabdomyosarcoma, adenocarcinoma, mucoepidermoid carcinoma, acinic cell carcinoma, undifferentiated carcinoma, clear cell carcinoma, liposarcoma, and esthesioneuroblastoma. These lesions spread to the cavernous sinus (CS) either directly (83.2%) via the superior orbital fissure, inferior orbital fissure, foramen rotundum, and foramen ovale or through metastasis (16.8%) via perineural extension or hematogenous or lymphatic spread. Malignant tumors of the nasal cavity and paranasal sinuses are rare, accounting for 0.2% to 0.5% of all cancer cases and only 3.0% of malignant tumors in the head and neck region. According to the Surveillance, Epidemiology, and End Results database of all reported sinonasal malignancies between 1973 and 2006, the cumulative incidence of sinonasal malignancy was 0.556 cases per 100,000 population per year; the most common sites of origin were the nasal cavity (43.9%) and the maxillary sinus (35.9%). CS involvement by sinonasal and ventral skull base malignancies carries dismal prognosis overall, because it precludes radical oncological resection in many instances. However, the overall and progression-free survival in a particular malignant ventral skull base tumor case with CS involvement depends on many factors besides the extent of resection, which include tumor burden, pathologic conditions, presence of metastasis, positive tumor margins, and age of patients. Aggressive surgical resection followed by adjuvant chemotherapy and radiotherapy offers the best possible chance of prolonging overall survival in most sinonasal and ventral skull base malignancies.
Since the seminal articles by Ketcham and colleagues in 1963 and 1966 detailing the role of surgery in intracranial involvement of head and neck malignancy, innovation of radical surgical resection procedures for the eradication of malignant ventral skull base tumors is arguably one of the most important advancements in the treatment of head and neck malignancy in the past half century. Besides the use of conventional open transcranial and transfacial approaches to achieve the surgical goal of oncologic resection, minimally invasive endoscopic approaches have been applied more recently to selected patients with reasonable outcomes. It was not uncommon in the past that malignant skull base tumors were considered inoperable. With advancement in skull base microneurosurgery, availability of better endoscopic devices, expertise in cerebral revascularization, more effective hemostatic agents, and improved ventral skull base reconstruction techniques, the collective opinion of the neurosurgical community has changed from this conventional thinking. More aggressive groups preferred piecemeal removal of tumor, including parts extending into the CS proper ; however, the oncological principle of en bloc resection was still not realized until the pioneering efforts of Sekhar and Moller, Saito and colleagues, and others. This review discusses the current surgical strategies, their indications, techniques, nuances, advantages, limitations, and complications of operative approaches for CS involvement of malignant ventral skull base tumors.
Surgical management
Since the pioneering work of Parkinson and others, more refined intradural and extradural transcranial surgical approaches have been described to access CS lesions; more recently, the extended endonasal/transmaxillary endoscopic approaches have been introduced in the realm of microneurosurgery. For a malignant tumor involving the CS, the approach needs to be individualized to optimize the treatment strategy. Firstly, confirming the diagnosis is of paramount importance in cases of inconclusive radiological impressions or suspicious-looking lesions. Use of minimally invasive approaches aided with stereotactic neuronavigation helps to achieve that goal safely. Once the diagnosis is confirmed, and if the patient is healthy and has no metastatic disease, aggressive surgical resection may be indicated, especially if the removal of the CS lesion may result in total tumor resection. Some investigators think that tumor resection with carotid preservation carries the lowest risk of cerebrovascular accidents and should generally be the treatment of choice. For patients with more advanced and recurrent malignant disease, whereby carotid preservation would prevent a meaningful resection, en bloc resection of the tumor and CS with cerebrovascular revascularization may be justified.
Frequently, en bloc resection will require CS exenteration (CSE), including sacrifice of the cavernous internal carotid artery (ICA), possible high-flow extracranial-to-intracranial bypass, and placement of a vascularized pedicled flap for ventral skull base reconstruction along with adjuvant chemotherapy and radiotherapy to effectively eradicate the microscopic tumor remnants. Often a combination of surgical approaches is necessary for optimal resection of aggressive ventral skull base malignancies with CS involvement, including frequent collaboration between the disciplines of otorhinolaryngology and neurosurgery. The surgical approach to the CS region includes the intradural and extradural transcranial and transfacial routes. When planning for a CS approach, preoperative planning is very important to prepare for any potential intraoperative events like ICA injury requiring vascular repair or cerebral revascularization or nerve damage requiring reanastomosis.
Minimally Invasive Transcavernous Biopsy of Cavernous Sinus Lesions
With the diverse radiological differentials for complex skull base lesions with CS involvement including infectious and inflammatory pathologies as well as benign and malignant tumors, whenever the radiological findings are inconclusive or suspicious for malignancy, minimally invasive tissue sampling methods should be attempted before proceeding with an aggressive surgical strategy because a precise identification of the cause of the lesion helps to optimize the treatment planning. The use of frameless stereotactic neuronavigation can be of great assistance in safely performing such blind procedures. Tissue sampling of radiologically inconclusive or suspicious CS lesions yields a positive diagnosis of malignant skull base tumor in as many as 37% to 40% of patients. Different surgical approaches can then be optimized to target the CS lesion based on the precise anatomic location of the lesion and surgeon’s preference.
Lateral orbitotomy approach
Altay and colleagues originally described the minimally invasive lateral orbitotomy extradural transcavernous approach in 2012. It is a practical, reliable, and low-risk minimally invasive technique for CS biopsy. The primary indication for this procedure is a lesion situated in the lateral compartment of the CS (primary position lateral to the carotid artery). Patients are positioned supine with the head stabilized on a Mayfield head holder with slight contralateral rotation (∼10°–15°). The LO approach to the CS involves a small 2-cm Y-shaped incision along the natural skin crease ( [CR] ). Next, the periosteum is dissected off the lateral orbital rim to expose the orbitozygomatic and fronto-orbital junctions. Subsequently, the temporalis muscle fibers and periorbita are carefully dissected from both sides of the lateral orbit wall. A 2-cm segment of the lateral orbital rim is cut using a fine-tip C1 drill bit to expose the inferior orbital fissure inferiorly, the orbitotemporal junction posteriorly, and up to the lower margin of sphenoid ridge superiorly ( Fig. 1 ). Before the bony rim is removed, thin-profile miniplates are screwed in and corresponding holes are tentatively made on either side of the bone removed to optimize the cosmesis of bony closure at the end. The remaining lateral orbital rim and lateral sphenoid wing are drilled to expose the temporal dura mater. Drilling the lateral orbital wall and deeper sphenoid wing under neuronavigation guidance helps in precise localization of the CS lesion.
The superior orbital fissure is centered in the operative trajectory, enabling easy detachment of 2 layers of the lateral CS, which is performed using an extradural Dolenc technique. Anterior clinoid process drilling can be added to augment the superior exposure of the CS and to decompress the optic nerve if desired. This drilling also aids in more posterior exposure as far back as the geniculate ganglion at a wider craniocaudal angle, besides proving an immediate vascular control of the clinoidal ICA. With further drilling superiorly along the lower aspect of sphenoid ridge and inferiorly along the remaining lateral orbital wall toward the inferior orbital fissure, the Meckel cave and the foramen rotundum are well within the operative reach. To access the more lateral aspect of the middle cranial fossa, including the mandibular division of the trigeminal nerve (V3), foramen ovale, foramen spinosum, greater superficial petrosal nerve (GSPN), and petrous ICA, an extended version of this approach that involves further drilling of the lateral aspect of greater sphenoid wing is required. The advantages of this approach are the small skin incision, minimal soft tissue dissection and blood loss, sparing of the temporalis muscle insertion, and, consequently, reduced risk of muscle atrophy. In addition, there is a shorter incision-to-target distance, with minimal brain retraction. These patients have a reduced hospital stay (usually 1 day). The primary limitation of this approach is the unfamiliar anatomy of the sphenocavernous region when viewed from below at a translateral orbital angle of view. Although infrequently encountered, potential complications include cerebrospinal fluid (CSF) leak, pseudomeningocele, orbital hematoma, and transient cranial neuropathy.
Percutaneous needle biopsy via the foramen ovale
Stechison and Bernstein reported the first attempt of a middle fossa biopsy through the foramen ovale in 1989. Lesions amenable to this approach are those involving the Meckel cave, the posterior part of the CS, and the upper part of the petroclival region. Patients are placed supine with the head under fluoroscopic control, and the procedure is performed under small-dose and short-lasting general anesthesia. After the local anesthetic is injected, the trajectory for percutaneous biopsy is made using the Hartel technique. The skin entry point is 3 cm lateral to the labial commissure. The tip of the biopsy needle is directed toward the foramen ovale, which corresponds to 3 cm anterior to the tragus on a horizontal line along the inferior border of the zygoma and pupilla. Depth is continuously assessed using fluoroscopy or stereotactic neuronavigation. Extreme care is required not to take a wrong trajectory and violate the internal jugular vein or the ICA at its entrance to the petrous canal and jugular foramen posterolaterally, the lateral segment of the ICA medially, or the optic nerve anteriorly along the orbital apex. Along the correct trajectory, the needle may encounter, successively, the parotid duct, maxillary artery, or auditory tube, with corresponding potential complications of hemosialorrhea, cheek hematoma, or middle ear hemoserous otitis, respectively.
The advantages of this approach include shorter operative time, short length of hospitalization, no brain retraction, and an acceptable safety profile. The primary limitations include the technically challenging nature of the approach, which requires adequate expertise, and the blind nature of tissue sampling with a significant number of nonproductive biopsies. In a series of 50 cases with percutaneous biopsy of CS lesions via the foramen ovale, Messerer and colleagues demonstrated that a productive biopsy was obtained in 86% (n = 43) of cases. Among those 43 cases, 28 patients underwent second (open) surgery and a second set of confirmatory histopathologic evaluation and resection. They demonstrated that percutaneous biopsy via foramen ovale had a sensitivity of 83% and specificity of 100%. Complications included 2 cases of facial cellulitis and cheek hematoma, with no permanent sequelae. Procedure-related complications have a low incidence but may include potential injury to the optic nerve, ICA, internal jugular vein, maxillary artery, and V3 resulting in possibly visual decline, ICA pseudoaneurysm, massive blood loss, cheek hematoma, and facial pain/numbness, respectively. This procedure is associated with some degree of facial hypesthesia, dysesthesia, or paresthesia in approximately two-thirds of cases ; however, most of the symptoms are short lasting, with associated permanent sequelae in only a fraction of cases.
Endoscopic transfacial approaches
Unlike the LO and percutaneous transforamen ovale approaches, which provide access to primarily the lateral CS compartment (lateral to carotid siphon), transfacial approaches (endonasal and transmaxillary endoscopic) provide access to the medial and lateral compartments of the CS. The extended endoscopic endonasal transsphenoidal trans-sellar and extended endoscopic endonasal transethmoidal transsphenoidal parasellar approaches are optimal for lesions situated primarily medial to the carotid siphon, whereas an extended endoscopic transmaxillary transpterygoid approach provides an access to the lateral CS lesions. Patients are placed supine with the head stabilized on Mayfield fixation in slight extension, lateral tilt, and ipsilateral rotation to provide a comfortable working trajectory. Use of stereotactic neuronavigation is essential in these extended approaches where surface landmarks may be erroneous at times, and the close proximity to vital neurovascular structures in the CS does not allow any margin of error in this narrow operative corridor. No incision is required for endonasal approaches while a sublabial incision is necessary for the transmaxillary approach. The endonasal approach requires a binostril, 2- or 3-handed technique to make the best use of the narrow working corridors. Lateralization of the middle turbinates, removal of distal bony nasal septum, and wide opening of the sphenoid sinus are common to both endonasal approaches to enhance exposure.
Apart from the steps these approaches have in common, for the trans-sellar approach, the sellar floor is removed, the pituitary gland is mobilized if needed, and biopsy of the CS lesion is done under direct vision using an angled endoscope, whereas for a parasellar endonasal approach, the superior turbinate and anterior and posterior ethmoidal air cells are opened. The parasellar endonasal approach aims to enter the CS more anteriorly than the trans-sellar approach, with the steps of removing the cavernous carotid bony wall, gently mobilizing the ICA, and obtaining the biopsy of the CS lesion. For a transmaxillary approach, the anterior and posterior walls of the maxillary sinus are opened to access the pterygopalatine fossa. Next, the pterygoid process is removed along with the ethmoid and sphenoid to expose the lateral aspect of CS, which allows for tissue sampling. The primary advantages of these approaches are the lack of incision and ability to achieve tissue sampling under direct vision; however, certain limitations to these approaches include protracted sinonasal symptoms after surgery, a risk of CSF leak, a higher risk of meningitis because of potential contamination in crossing the nasal cavity or paranasal sinuses, and risk of cavernous ICA injury. For a transmaxillary approach, damage to the neurovascular structures along paramedian skull base, including maxillary, mandibular, and vidian neurovascular bundles, is a possible procedure-related complication.
Aggressive Transcranial Approaches to the Cavernous Sinus
In most young, healthy patients who have been diagnosed with malignant ventral skull base tumors with CS involvement and limited metastatic disease, aggressive surgical exploration of the CS and resection of the diseased tissue remains an option as an initial modality of therapy. It differs from radical en bloc resection in principle in that it allows only piecemeal removal of tumor rather than aiming at oncological resection of malignant disease. This approach has been tailored to achieve maximal safe resection of the tumor with minimum possible risk of iatrogenic neurovascular injury, aiming at a functional preservation for better quality of life. Although it may benefit patients from a functional standpoint, it may compromise on the best possible overall and progression-free survival for a given ventral skull base malignancy. Patients with multiple metastases and poor overall health status with multiple comorbidities are relative contraindications for extensive skull base surgery (both aggressive and radical en bloc resections), whereby high intraoperative blood loss is expected along with the requirement of prolonged general anesthesia.
It is critical to understand the surgical anatomy of the 4 CS and 4 middle fossa triangles to grasp the concept of transcranial transcavernous surgical approaches. These surgically relevant triangles provide invaluable working operative corridors and help with accessing different CS tumors with middle and posterior fossa extensions. For performing a transcranial route to the CS, a frontotemporal craniotomy with or without orbitozygomatic or transzygomatic osteotomy is required. There are 3 primary safe operative trajectories to enter the CS transcranially, vis-á-vis anteromedially through the roof of CS via the Hakuba (oculomotor) triangle using the Hakuba/Dolenc approach; anterolaterally through the lateral wall of the CS via various CS and middle fossa triangles using a modified Dolenc approach; and posterolaterally through the posterolateral wall of the CS using the Kawase approach. A suitable approach (or a combination of these approaches) to the CS is chosen depending on the tumor size, location, epicenter, extent of the tumor, nature of pathology, and the preference of the operating surgeon.
There are certain common operative steps in all 3 transcranial approaches. The patient is placed supine with the head turned 30° to 60° to the contralateral side (depending on the epicenter of the lesion and suitable operative corridor) and slightly extended. Intraoperative electrophysiologic monitoring for motor evoked potential, somatosensory evoked potential, and electroencephalogram is set up. A standard curvilinear incision is made from just anterior to the tragus to the midline along the hairline, and a single myocutaneous flap is elevated to expose the frontotemporal bone, the root of the zygoma, and the Key burr-hole point. After a frontotemporal craniotomy is performed, the sphenoid ridge and lateral sphenoid wing are drilled flush to the anterior and middle cranial fossa. Once the meningo-orbital band is identified, it is coagulated and divided. To achieve adequate access in all 3 transcranial transcavernous approaches, expansion of the operative corridor can be achieved by augmenting bone removal in the form of orbitozygomatic osteotomy and transzygomatic osteotomy. Specific further primary operative steps are highlighted later in their respective sections. Closure is done in a standard fashion at the end of the procedure, and the skull base defect is reconstructed and plugged using a fat graft, free fascia lata, pedicled nasoseptal flap from below, pedicled pericranial flap, temporalis muscle graft, or a combination of these depending on the size of the dural defect and the surgeon’s preference.
The primary advantages of the microscopic transcranial approach over the endoscopic transfacial approach include the wide operative corridor between various cranial nerves located in the lateral wall of the CS, better vascular control in case of inadvertent ICA injury, and a lower risk of CSF leak and associated complications. There is no limitation in the size and extent of the intracranial disease process beyond the CS involvement. Limitations include large scalp incisions with higher wound-related postoperative morbidity, a higher risk of iatrogenic cranial neuropathy, and possible brain retraction–related complications. In addition, extracranial sinus disease is less amenable to resection via a transcranial approach. Extensive skull base malignancies with CS involvement often warrant combined otorhinolaryngology and neurosurgery surgical procedures, with simultaneous transcranial and transfacial approaches for the best possible outcome ( [CR] ). Overall, the possible complications for transcranial approaches include wound complications, infection, CSF leak, pseudomeningocele, cranial neuropathy, vascular injury, retraction hematoma/contusions, seizures, and neurologic deficits.
Anteromedial transcavernous approach
For entry into the CS roof, a combined intradural and extradural approach is required including drilling of the anterior clinoid process either intradurally or extradurally. Once the anterior clinoid process is drilled, it exposes the clinoidal ICA, carotid-oculomotor membrane, optic strut, superior orbital fissure, and optic canal. Adequate irrigation must be used while drilling the anterior clinoid process to prevent optic nerve damage from thermal injury. Additionally, a thin rim of bone along the optic nerve and ICA should be left after the drilling; it is safer to remove this thin shell of bone with microdissectors to avoid any inadvertent damage to the optic nerve and clinoidal ICA while drilling. Careful preoperative assessment of the bony anatomy of the skull base should be used to rule out physiologic variations like an osseous bridge between the anterior clinoid process and middle clinoid process, forming a caroticoclinoid foramen. Care should be taken to seal off any air spaces, which might open inadvertently while drilling.
Besides anterior clinoid process drilling, adequate CS roof exposure requires wide splitting of the sylvian fissure. In addition, the temporal lobe should be freed from arachnoid adhesions along its medial and basal surface to allow gentle retraction to access the roof of CS. After the distal dural ring is opened, the supraclinoidal and clinoidal ICA can be mobilized to increase the access to the CS roof. Vertical entry into the CS roof via the oculomotor (Hakuba) triangle is then achieved by carefully dividing the dural membrane along the anterosuperior aspect of the oculomotor nerve entry into CS roof, until the nerve becomes incorporated into the lateral wall of the CS. The next step is to open the carotid-oculomotor membrane in the clinoidal (Dolenc) triangle medial to the oculomotor nerve to widen the intracavernous operative corridor. Any venous bleeding can be controlled using absorbable hemostatic agents.
Anterolateral transcavernous approach
Umansky and Nathan described the 2-layer composition of the lateral wall of the CS, which allows safe peeling of the outer layer of dura mater away from the inner layer along the middle cranial fossa to provide adequate extradural access to the lateral wall of the CS and its contents through the various CS and middle fossa triangles. Dolenc also pioneered this technique of the interdural transcavernous approach, followed by piecemeal removal of tumor via the narrow operative corridors between the nerves in the lateral wall of the CS. The dissection starts at the greater sphenoid wing and proceeds posteromedially toward the superior orbital fissure, where the intracranial periosteum is continuous with the periorbita. A shallow cut is made, which allows gentle separation of the dura mater from the lateral wall of the CS and middle cranial fossa floor. This outer (meningeal) layer is peeled away from the inner (endosteal) layer to expose cranial nerves III, IV, V1, V2, V3 and gasserian ganglion, along with the CS and middle fossa triangles. The appropriate operative corridor is chosen to access the tumor filling the CS ( Fig. 2 ).
