Endoscopic transnasal skull base surgery: pushing the boundaries
- First Online:
- Cite this article as:
- Zwagerman, N.T., Zenonos, G., Lieber, S. et al. J Neurooncol (2016) 130: 319. doi:10.1007/s11060-016-2274-y
- 828 Downloads
The endoscopic endonasal approach (EEA) has significantly evolved since its initial uses in pituitary and sinonasal surgery. The literature is filled with reports and case series demonstrating efficacy and advantages for the entire ventral skull base. With competence in ‘minimally invasive’ parasellar approaches, larger and more complex approaches were developed to utilize the endonasal corridor to create maximally invasive endoscopic skull base procedures. The challenges of these more complex endoscopic procedures include a long learning curve and navigating in a narrow corridor; reconstruction of defects presented new challenges and early experience revealed a significantly higher risk of cerebrospinal fluid leak. Despite these challenges, there are many benefits to the EEA including avoidance of brain and neurovascular retraction, improved visualization, a direct corridor onto many tumors and the two-surgeon approach. Most importantly, the EEA provides a midline corridor to directly access tumors, which displace critical neurovascular structures laterally, giving it an inherent advantage of minimizing any manipulation of these structures and thus decreasing their potential injury.
KeywordsEndoscopic endonasal approach Endoscopic endonasal surgery Skull base Transclival Transcribriform Transpterygoid
Like any new approach, there is a learning curve associated with EES (endoscopic endonasal surgery). Surgical experience is the key for development of competency, but the number of cases required is not clearly defined. Multiple groups have reported their learning curve experience. One initial study from 2004 evaluated a series of the first 45 cases and concluded that surgeons can safely transition from traditional open approaches to endoscopic approaches without suffering from the learning curve . This group however did not account for the complexity of the cases, amount of resection or experience of the surgeon. A more recent study from 2010 examined one team’s first 51 endoscopic cases and found that there was a significant decrease in the risk of DI and CSF leak in the last one-third (17) of cases . Both these studies could be interpreted to support a learning curve of at least 50 cases.
The acquisition of surgical skills for EES is multifaceted, but there is a series of recommended levels for a team to progress through to develop proficiency while respecting their learning curve. Level I consists of working as a team on basic endoscopic sinus surgery to build team skills and master sinus anatomy. After this, the team transitions to Level II where advanced sinus surgery is combined with cerebrospinal fluid (CSF) leaks and intrasellar pathology (adenoma, Rathke’s cleft cyst). These cases can be done without a team, but they provide the foundation and initial learning curve that allows for a smooth transition to more difficult cases. Level III includes sellar lesions that extend outside the sella including suprasellar adenomas, optic nerve decompression, orbital approaches and extradural skull base surgery. With success in this level, the team can advance to intradural skull base surgery for lesions such as craniopharyngioma and tuberculum meningiomas. Finally, Level V skills involve progression in the coronal plane to expose the carotid, vascular lesions, as well as lesions in cavernous sinus and Meckel’s cave . Figure 1 identifies the endonasal endoscopic approaches in the sagittal and coronal planes.
As with any new approach, experienced teams must both push the boundaries of the approach and evaluate their results to better define what these boundaries should be. These boundaries shift slowly as experience is gained, techniques are refined, anatomy defined, instrumentation improved, and complications addressed.
Tumors of the tuberculum sellae region present unique difficulties for EES. It is a compact anatomic region with fine but critical microvasculature and potential involvement of the Circle of Willis. The high flow leaks created require the use of a vascularized flap, the most common of which is the nasal septal flap [4, 5]. Several groups have published their experience and have demonstrated successful resection of these lesions [6, 7]. The endoscopic endonasal approach for craniopharyngioma has become relatively widely accepted as the preferred approach of all but purely ventricular tumors, with excellent vision outcomes and the potential for improved pituitary preservation [8, 9]. Meningiomas are more controversial but show improved visual outcomes, with the initial UPMC experience with 75 suprasellar meningioma patients consisted largely of patients (81 %) presenting with visual deficits. In 85 % of patients, vision normalized after surgery and deteriorated in only two patients, much lower than modern craniotomy studies . Gross total resection (Simpson Grade 1) was accomplished in 76 % of patients. Optic canal invasion was not a limiting factor for complete removal. However, tumor size (>3 cm), multilobular configuration and vascular encasement were significant factors preventing complete resection. One patient suffered from a vascular injury to the artery of Huebner with a resultant dominant lobe caudate infarct . CSF leak remained the drawback to this approach and a recent systematic review showed similar results, with improved visual outcomes compared to craniotomy in exchange for higher CSF leak rates . Several other groups have reported similar outcomes [12, 13, 14, 15]. Figure 2 identifies important structures and steps for suprasellar, transplanum and transcribriform approaches in a cadaveric specimen.
Video 1 of the Endoscopic endonasal approach for a craniopharyngioma with suprasellar extension and stalk preservation.
A natural extension for the endoscopic endonasal approach is anteriorly along the skull base to the planum and cribriform for resection of planum/cribriform. Previous studies have demonstrated the safety of this approach, however, again with a potentially greater risk of CSF leak [16, 17]. A series of 50 olfactory grove meningiomas resected via EEA alone indicated complete tumor resection (Simpson Grade 1) in 67 % of patients in whom that was the goal, but residual tumor was seen lateral over the orbit or anterior along the frontal sinus/falx with vascular involvement and size (>4 cm) again being limitations. Of note, the overall CSF leak rate for these patients was 30 % . These results raise question about the approach, but the benefits of this approach include absent brain retraction and direct access to and early ligation of tumor attachment and blood supply as well as potentially improved visualization of tumor margin.
The risks of brain retraction are demonstrated in surrogate markers like radiographic frontal lobe changes and seizure risk. Recently, a matched pair comparison was performed between patients who underwent open, bifrontal approach or endoscopic endonasal approach (EEA) and indicated that there was a significant decrease in postoperative frontal lobe changes like MRI FLAIR signal change and porencephalic cave with EEA compared to the bifrontal approach . Another study has shown a lower incidence of seizures (0.84 %) following endonasal resection of 827 tumors compared with craniotomy despite the lack of routine prophylactic anticonvulsant use .
Finally, endoscopic endonasal surgery for esthesioneuroblastoma appears to be associated with outcomes that compare very favorably to open approaches, with negative margins achieved in up to 100 % of cases in some series .
Endoscopic Endonasal Surgery (EES) has become a powerful tool for treating pathology of the clivus, the petroclival region, as well as intradural posterior fossa lesions immediately adjacent to the clivus. Traditionally, the clival and paraclival regions have been difficult to approach, especially for pathology that has significant extension in the sagittal plain, and/or had significant bilateral extension. For such tumors, often a combination of open approaches was required, as evidenced by the traditional division of the clivus into thirds, each requiring a separate approach .
By comparison, EEAs are very versatile in the sagittal plain, while the ventral surgical corridor provides access to lesions that extend bilaterally across the midline. Thus, even larger tumors that span the entire clival region can be accessed through a single endonasal corridor.
Upper transclival approach
An endoscopic superior transclival approach provides midline access to the interpeduncular cistern, basilar apex, mammillary bodies, and the floor of the third ventricle. The upper clivus, or “sellar clivus” is formed by the posterior clinoid processes and the dorsum sella, which have to be accessed and removed during this approach. Access to the interpeduncular cistern and midbrain behind the upper clivus usually requires posterior clinoidectomy, which can be unilateral or bilateral depending on the pathology. Removal of the posterior clinoids most of the times requires transposition of the pituitary gland, which can be done in one of three ways: (a) extradurally, (b) interdurally, or (c) intradurally:
The extradural pituitary transposition involves upward mobilization of the gland with both its meningeal and periosteal dura [23, 24, 25]. This is associated with the least risk of injury to the pituitary gland, but is also the least versatile of the three in terms of the exposure that can be achieved. As such, this approach may lead to inadvertent tearing of the cavernous sinus and inadequate access to the posterior clinoid.
The interdural pituitary transposition, is essentially a transcavernous approach between the meningeal (medial) and periosteal (lateral) layers of the cavernous sinus . This approach allows preservation of the venous outflow of the pituitary since the meningeal layer of the gland is preserved, while providing excellent mobilization of the gland and allowing access to paramedian lesions within or behind the parasellar space, such as para/retro/supra- sellar extension of chordomas or chondrosarcomas, This approach requires mobilization or sacrifice of the ipsilateral inferior hypophyseal artery and for this reason bilateral interdural transpositions theoretically carry a higher risk for pituitary dysfunction. In our experience, however, sacrifice of both inferior hypophyseal arteries does not necessarily cause posterior pituitary dysfunction, suggesting alternative vascular supply by the superior hypophyseal arterial system .
The intradural transposition of the pituitary provides the greatest degree of mobilization. However, the necessary detachment of the gland from its meningeal layer of dura also carries the highest risk for pituitary dysfunction . For this reason we usually avoid this approach; however, it could be used for tumors involving the posterior surface of the stalk and infundibulum, which are already associated with a high risk for pituitary dysfunction. Such tumors include retroinfundibular craniopharyngiomas, pituicytomas, or infundibular astrocytomas.
Middle transclival approach
A middle transclival approach provides access to the ventral pons and prepontine cistern, the basilar trunk and anterior inferior cerebellar artery, as well as the cisternal segment of the abducens nerve. The sphenoidal clivus is limited laterally by the paraclival ICAs, and the petroclival fissure. Laterally, the middle transclival exposure is limited by the interdural segment of cranial nerve VI.
Video 2 indicates a transclival approach for a chordoma invading the intradural space, compressing the pons and midbrain.
Lower transclival approach
The lower transclival approach through the lower segment of the clivus, which lies below the roof of the choana, exposes the premedullary cistern and ventral medullary surface, the vertebral arteries, vertebrobasilar junction and posterior inferior cerebellar arteries, as well as cranial nerves IX–XII.
The lateral inferior clival segment is divided into two compartments by the hypoglossal canal : (a) the tubercular or superior compartment that corresponds to ventral aspect of the jugular tubercle, and drilling of which provides access to the cisternal segment of cranial nerves IX–XI in their course towards the jugular foramen , and (b) the inferior or condylar compartment, which corresponds to the ventral occipital condyle. Drilling of the medial condyle provides access to the vertebral artery as it enters the dura of the posterior fossa. In our experience, drilling of up to 75 % of the condyle does not lead to guaranteed craniocervical instability if the remainder of the condyle is structurally intact and not invaded by tumor .
This approach may be taken to the craniocervical junction to address pathology below the foramen magnum to the level of the odontoid. The lower limit of an endonasal lower transclival approach is determined by the hard palate. Because of this anatomical limitation, chordomas involving the lower clivus and craniocervical junction are extremely challenging, and may be associated with higher recurrence rates . The nasopalatine line can be used to estimate the lowest extent of resection; however, in our experience, soft tissue may limit our ability to access this point and in the absence of curved drills or instruments, the area accessed may be slightly above this line . Many groups have reported their experience with endoscopic endonasal odontoidectomy [32, 33, 34, 35]. Although still controversial subject, mounting evidence suggests that the endonasal route may be superior to the transoral route, resulting in a statistically significant lower rate of tracheostomy . Figure 3 demonstrates the transclival approaches including pertinent structures and landmarks.
Coronal plane boundaries
Extending beyond the sagittal plane involves an in depth understanding of the relationships between the maxillary sinus and skull base. It also requires a team with experience with simpler approaches and an anatomic understanding of and technical facility to manage the ICA, which is the structure that must be crossed in order to expand laterally into the coronal plane. Figure 4 depicts the transpterygoid approach and the expansion into the coronal plane.
Cavernous sinus/Meckel’s cave
Cavernous sinus extension has traditionally been considered a contraindication for access via a transsphenoidal approach. However, the endoscope and coronal plane extensions allow for wide visualization and access well beyond the cavernous sinus. Instrumentation has slowly improved to allow greater control and ability to perform microsurgical dissection in these regions as well.
Approaches depend largely on the access the operator is trying to attain. The cavernous sinus has been divided in many different ways, but we have found that the most useful model involves dividing the cavernous sinus into four compartments based on tumor extension and relationship to the cavernous ICA (superior, posterior, inferior and lateral). Figure 5 describes the important landmarks and boundaries for the cavernous sinus compartments. The transsellar approach to the cavernous sinus accesses the medial compartments (superior and posterior) while the transpterygoid approach provides wider access to the lateral compartments (inferior and lateral) with greater ICA control.
The superior compartment is located superior to the horizontal cavernous ICA, behind the anterior genu and is the most common location for invasion by pituitary adenomas. Posterior to the short, posterior vertical segment of the cavernous ICA is the posterior compartment, which extends inferiorly to the superior aspect of the abducens nerve just posterior to the carotid. Also in this area is the meningohypophoseal trunk, which gives off the inferior hypophoseal artery, less important to preserve than to not injure.
The inferior compartment is located inferior to the horizontal cavernous ICA and anterior to the short vertical segment. Important structures in this region include the sympathetic plexus that travels along the carotid to with branches that join the abducens nerve, traveling in parallel and lateral to the horizontal cavernous ICA as it enters the lateral compartment.
Finally, the lateral compartment of the cavernous sinus comprises the space lateral to the parasellar ICA. Ensheathed in the lateral wall of the cavernous sinus are cranial nerves III, IV, and the first division of the trigeminal nerve. The abducens nerve is free floating in this compartment and its location may be unpredictable if tumor has altered the anatomy. Also, the inferolateral trunk is a branch of the inferior aspect of the horizontal cavernous carotid and supplies the lateral wall and its contents. Many groups have published their results, proving that the endoscopic endonasal approach is safe and effective in dealing with pathology in the cavernous sinus [36, 37, 38, 39].
The vidian nerve serves as a key landmark for localizing the petrous and paraclival segments of the carotid when other landmarks are distorted by pathology . The classic example of pathology involving this area is juvenile nasoangiofibroma. Our series of 34 patients, including 14 patients with intracranial extension, who underwent EEA supplemented with a Caldwell-Luc endoscopic maxillary antrostomy showed that this area can be widely accessed. The endonasal (with or without anterior transmaxillary as needed) approach provides a ventral working corridor for resection with a chief advantage of direct tumor access both medial and lateral to the carotid.
Expanding lateral and inferior to the cavernous sinus, several authors have published reports on approaches to Meckel’s cave lesions [41, 42, 43]. Kassam et al. published their experience with Meckel’s cave lesions and found that the most common lesions were adenoid cystic carcinoma, meningioma, and schwannomas. They reported a 12.5 % rate of new facial numbness in at least one new distribution but was only permanent in 5 % of patients. Whereas 30 % of patients indicated improvement of preoperative deficits . In addition, 17 trigeminal schwannomas were resected at our institution, with the vast majority (ten) being V3 tumors while only five involved the middle and posterior fossa. Furthermore, some had to be combined with either a retromastoid approach, anterior maxillary antrostomy or other ‘open’ approach for removal. A gross total resection was achieved in 82 % of those in whom it was the goal (n = 11). Others were intended for either decompression, biopsy or subtotal resection with preservation of function.
Coronal expansion of the lower transclival approach to access tumors extending lateral to the petroclival junction can be achieved with a transpterygoid infravidian approach followed by a sublacerum or infrapetrous approach [45, 46].
The petrous apex itself can be accessed relatively easily, but only if the surgical team is comfortable skeletonizing the ICA so that it can be lateralized. Otherwise, only lesions that expand the petrous apex such as cholesterol granulomas can be easily accessed transsphenoidally. EEA in this case proves to be highly effectively with little or no morbidity . Another option is to utilize a contralateral transmaxillary approach to reach behind the paraclival ICA to avoid skeletonization or retraction.
Orbital apex/superior orbital fissure/orbit
This approach, via combining maxillary and ethmoid sinus exposure, can be used to decompress the optic canal, the orbit, or to remove lesions from within the periorbita [47, 48, 49]. Using this approach, we have been able to document 180° decompression of the optic nerve at the canal and are only limited by the optic nerve as well as the ophthalmic artery . For intraorbital pathology, lesions that are located medial to the optic nerve posterior to the globe can be accessed using the EEA . Risks of surgery in this region are greater for intraconal lesions compared with extraconal or extraorbital compressive lesions. Figure 6 details the endoscopic endonasal approach to the orbit highlighting pertinent anatomy and landmarks.
EEA aneurysm clipping
Several case reports have established that aneurysm clipping is possible from an endonasal approach but limitations exist [51, 52, 53]. There are very limited indications, but the nasal corridor is an ideal anatomic approach for medially pointing aneurysms such as superior hypophyseal or medially pointing ophthalmic anterior circulation aneurysms . Reports of posterior fossa circulation aneurysms exist but this only will work under conditions of a low riding basilar apex and favorable anatomy . The endonasal approach limits the angle of clip placement and generally only one or two trajectories may be used. In addition, a common problem faced with aneurysm clipping is covering the clips or delayed erosion of the clip through the flap/graft. This has been mitigated by robust packing of fat around the clips for added protection. Video 3 depicts an endoscopic endonasal approach for clipping of bilateral carotid aneurysms.
Any endonasal vascular surgery should be approached with extreme caution and only by longstanding teams with significant experience.
The field of skull base surgery continues to evolve as treatments and approaches improve. The combination of multi-corridor approaches has the potential to be extremely useful to approach extensive pathologies for more complete resections while limiting morbidity. The combination of endonasal and supraorbital approaches has the potential to attain complete resections of large olfactory groove meningiomas that extend well over the orbit yet into the sinus. Endonasal and posterolateral approaches may provide for improved resection of midline tumors with lateral extension. These approaches are complementary to each other and their roles will be further defined to provide maximal advantage. Furthermore, with improvements in technology, robotic-assisted approaches to the skull base may become useful and have already been described in a case report . Improvements in suturing techniques and closure devices will help to decrease CSF leaks and lead to better patient outcomes. Finally, as this field becomes more commonplace, surgical tools will improve and will further advance the approach to improve efficiency and effectiveness.