Review of the Literature
Earliest Attempts
Transcranial Approaches
Although he did not report on it until 1906, Sir Victor Horsley performed the first
transcranial pituitary operation in 1889 but met with limited success using the approach
because of what was later determined to be forceful retraction of the frontal lobe.[1 ]
[2 ] In 1905, Fedor Krause of Berlin used a frontal transcranial approach to reach the
sella turcica in a living patient.[3 ] It was this initial work that provided the basis on which the majority of subsequent
variations on transcranial approaches were developed. One variation by McArthur involved
an extradural approach with resection of the supraorbital ridge and the orbital plate,
allowing dissection to extend posteriorly toward the optic chiasm.[4 ] Further modifications were made by several neurosurgical pioneers in the early part
of the 20th century, including Dandy, Heuer, Frazier, and Cushing.[3 ] Harvey Cushing advocated a transfrontal craniotomy with a direct right subfrontal
midline approach.[5 ] As a result of Cushing's commitment to perfecting intracranial approaches and his
powerful influence on American neurosurgery, the mainstream neurosurgical teaching
during the 1930s and 1940s continued to focus on a transcranial approach to the pituitary
gland.[3 ]
Transsphenoidal Approach
Based upon the work of Giordano, Hermann Schloffer of Austria reported the first successful
resection of a pituitary tumor via a transsphenoidal approach in 1907. With local
anesthesia provided by cocaine, Schloffer performed a three-stage procedure that appeared
to represent a modification of contemporaneous approaches to treat sphenoid sinusitis.[6 ] Although the first transsphenoidal operations by Schloffer, von Eiselsberg, and
Kocher required external rhinotomy incisions, techniques quickly developed that decreased
the invasiveness of this approach.[7 ] In 1910, Hirsch introduced the endonasal approach by reporting two cases.[8 ] Nearly simultaneously, Halstead pioneered the sublabial approach whereby he was
able to preserve the cartilaginous septum, thus obtaining more pleasing postoperative
aesthetic outcomes.[9 ] Although these approaches both required some degree of turbinectomy or ethmoidectomy,
they represent the earliest versions of the two most common transsphenoidal approaches
to the sella used today.
By 1914, Cushing described the successful use of the sublabial transseptal approach
and used the transsphenoidal approach between 1910 and 1925 to operate on 231 pituitary
tumors, with a mortality rate of only 5.6%.[10 ] Despite being recognized as less invasive and providing better visualization, Cushing
abruptly abandoned its use from 1929 to 1932 in favor of the transcranial route.[7 ] Cushing returned to the transcranial approach for sellar tumors largely due to poor
visualization and likely because he considered the extent of resection and intraoperative
complications to be more easily evaluated and treated from above.
By 1956, one of Cushing's pupils, Norman Dott of Edinburgh, who recognized the importance
of the transsphenoidal operation, had performed 80 consecutive transsphenoidal operations
with no deaths.[11 ] He is also credited with developing a lighted speculum retractor that improved illumination
of the surgical site. Dott then introduced his method to Gerard Guiot, who began to
perform the transsphenoidal approach in 1957 and subsequently accrued a series of
more than 1,000 cases of pituitary adenomas. These few pioneers, by preserving and
improving the transsphenoidal approach, paved the way for the modern era of neurosurgery.
Although leaving no visible cosmetic defects, these early transsphenoidal operations
could hardly be considered minimally invasive. Still, over the ensuing decades, advances
in technique and technology allowed future neurosurgeons to build upon the principles
set forth by Cushing, Halstead, Hirsch, Dott, and Guiot.
Technological Advances: Paving the Way for Modern Approaches to the Sella
Beginning in the 1950s, a series of technical and technological innovations would
set the stage for the transition toward 21st-century approaches to the sella. With
the increased use of antibiotics and the introduction of hydrocortisone replacement,
the mortality and morbidity associated with pituitary surgery continued to decrease.
Subsequently, two innovations contributed to a renewed interest in the transsphenoidal
approach.
Intraoperative Fluoroscopy
Soon after performing his first transnasal resection, Guiot introduced intraoperative
fluoroscopy, allowing the surgeon to visualize the depth and positioning of surgical
instruments in real time. This real-time visualization revolutionized the technical
aspects of pituitary surgery and can be considered the first step toward intraoperative
neuronavigation. Fluoroscopy allowed for safer, more extensive resection of sellar,
parasellar, and suprasellar lesions and was soon associated with improved surgical
outcomes.
Jules Hardy—The Operative Microscope and Selective Adenomectomy
Although the basic techniques of transsphenoidal surgery had been established, adequate
illumination of the operative field had always been a limitation.[2 ] Cushing had used a headlamp, Hirsch used the otolaryngologic mirror, and Dott tried
to improve illumination by attaching small lightbulbs near the tip of the retractor
blade.[12 ] Although otolaryngologists had used the operating microscope since the 1920s, its
applications to transsphenoidal surgery were pioneered by Jules Hardy in 1965, who
first used it for a total hypophysectomy for metastatic breast cancer. The microscope
improved illumination, added magnification, and provided stereoscopic visualization,
allowing Hardy to develop the technique of selective adenomectomy with pituitary gland
preservation.[3 ] These benefits quickly became widely recognized, and the microscope was soon adopted
as an essential component of the transsphenoidal approach.
Endoscope
Many argue that the major limitation of the microscope in the transsphenoidal approach
is the restricted visualization limited to a corridor confined within the nasal speculum.
In contrast, the modern rod-lens endoscope provides a more panoramic view unobtainable
with the microscope. This limited “tunnel vision” microscopic view coupled with technical
advances and a growing experience in sinonasal endoscopy fueled the revolution in
endoscopic transsphenoidal surgery that began in the 1990s.[13 ]
Neuronavigation, Doppler Probe, and Electrophysiologic Monitoring
Prior to the development of modern neuroimaging techniques, tumor localization was
typically based on surface or internal landmarks. Currently, with the ability to apply
frameless image guided navigation, surgical planning and intraoperative maneuvering
can be more precise and thus reduce the risk of collateral damage to the normal brain,
cranial nerves, and cerebral vasculature. The application of image-guided navigation
in both transcranial and transsphenoidal surgery has allowed us to maximize surgical
resection while minimizing risk. The enhanced ability afforded by neuronavigation
to localize tumors has allowed surgeons to minimize the use of large craniotomy flaps
in favor of more precise “keyhole” approaches.[14 ] These minimally invasive approaches allow the surgeon to identify both the tumor
and key anatomical structures while minimizing the risk of injury and the risks and
discomforts of larger exposures and approaches. As a result, neuronavigation based
on computed tomography (CT) or magnetic resonance imaging (MRI) has become a standard
adjunct for sellar tumor resection.
Evolution of the Transsphenoidal Technique—Progress toward Minimally Invasive Neurosurgery
Fueled by advances in both technology and progressively more detailed understanding
of microsurgical anatomy, the end of the 20th century saw a relatively rapid evolution
in surgical techniques for access to the sellar region. Thanks to the increasing availability
of information through online publication of medical literature, this evolution was
led by a few pioneers but has spread rapidly to become widely accepted. In the late
1980s and 1990s, there was a transition for many neurosurgeons, including our group,
away from the traditional sublabial transsphenoidal approach toward the direct endonasal
transsphenoidal approach initially described by Griffith and Veerapen.[15 ]
[16 ] Although these techniques were performed primarily under microscopic visualization,
they represent advances along the stepwise progression leading to our current technique.
As neurosurgeons gained more experience with the rod-lens endoscope, increasing collaboration
with otolaryngologists resulted in the elimination of the nasal speculum and microscope
in transsphenoidal procedures. In 1997, Jho and Carrau published the first large series
(50 patients) with predominantly pituitary adenomas treated via a fully endoscopic
transsphenoidal approach.[17 ] Subsequently, during the late 20th and early 21st centuries, many neurosurgeons
began to transition initially from a traditional microscopic to an endoscope-assisted
and eventually to a fully endoscopic approach. Cappabianca et al further refined the
procedure by developing unique endoscopic instrumentation and identifying areas for
technical improvement.[18 ] In addition, by providing a panoramic view, the endoscope has been increasingly
utilized for lesions beyond the sella. Contributions made by Kassam, Carrau, Snyderman,
Gardner, Prevedello, and others facilitated a further reach of extended approaches
to the midline skull base, which were originally described using the microscope by
Weiss, Oldfield, and Laws.[13 ] These endoscopic approaches include the transcribriform and transplanum approaches
to the anterior cranial fossa, ethmoid-pterygoid-sphenoid or direct transsphenoidal
approach to the cavernous sinus, and the transclival approach for infrasellar skull
base and prepontine lesions.
Discussion
Advantages and Limitations of Pure Endoscopic Transsphenoidal Approach
The advantages of the endoscopic approach can best be appreciated in comparison with
the microscopic approach. Although both the endoscopic and microscopic techniques
have been used successfully on a variety of sellar and parasellar tumors, our evolving
experience suggests that the endoscope may have several distinct advantages in certain
situations. Perhaps most importantly, by providing a panoramic wide-angle view, the
endoscope provides superior visualization compared with the microscopic approach.
This attribute is of particular importance for tumors with extension beyond the sella,
including suprasellar, cavernous sinus, and anterior cranial fossa, as can often be
seen with nonfunctioning pituitary adenomas, craniopharyngiomas, and parasellar meningiomas.
The limited field of view provided by the microscope rarely allows for visualization
of the optic and carotid protuberances or the opticocarotid recesses. Although not
as important during resection of small, midline tumors, visualization of these key
landmarks can be critical to avoid catastrophic injury, particularly during resection
of larger and more expansive tumors. An additional advantage of the endoscope is the
ability to manipulate the line of sight. Once set in place, the operative microscope
provides a clear, three-dimensional view of the sellar region. The endoscope is mobile
and provides the ability to “see around corners,” particularly with using 30- and
45-degree angled lenses. This can be a particular advantage for close inspection of
the tumor/gland interfaces and for removing tumors from the optic chiasm region and
cavernous sinuses.
The advantages of the endoscope are not without cost. The endoscopic approach introduces
an entirely new system for neurosurgeons, who are usually more accustomed to the operating
microscope. The importance of surgeon familiarity with instruments and camera systems
cannot be overstated, and there is a significant learning curve when attempting endonasal
endoscopy. This fact is underscored by the reality that operator experience is associated
with better outcomes and lower complication rates.[19 ] Besides the new technology, the presence of the endoscope within the operative field
introduces a unique challenge, representing a physical obstacle that must be accommodated
with wide exposure and instrument adjustment. The endoscope itself limits maneuverability
of other surgical instruments, which is exacerbated by its limited zoom capacity.
Because the scope must often be advanced deep into the operative field to achieve
the optimal view, collision and conflict (“sword-fighting”) with other instruments
is a frequent challenge that must be minimized. Because of the presence of multiple
simultaneous instruments, the nasal mucosa is at greater risk of being injured and
care must be taken to sufficiently lateralize the middle turbinates to minimize damage.
Finally, and perhaps most importantly, to perform two-handed microneurosurgery, the
endoscope must be held or driven by another surgeon who can provide an optimal view
of the surgical field while minimizing conflict with the other surgical instruments.
This cosurgeon is increasingly an otolaryngologist who is skilled in sinonasal endoscopy,
as originally described by Jho and Carrau in 1997.[17 ] Static instrument holders have been used as an alternative to the two-physician
strategy, though this comes with added equipment cost, cumbersome setup, and the need
for manual adjustment of view by the primary surgeon.
Another limitation of the endoscope is that it provides only a two-dimensional image
as compared with the three-dimensional microscopic visualization. Although this is
potentially a problem particularly for less experienced surgeons, the dynamic movement
of the endoscope within the sinonasal skull base space allows the surgeon to progressively
gain a three-dimensional anatomical understanding. Whether newer three-dimensional
endoscopes that are being progressively being improved will ultimately be proven superior
over current two-dimensional endoscopes in terms of tumor removal rates and complications
remains to be proven.[20 ]
Transsphenoidal Surgery in the 21st Century—Surgical Technique
Currently at our institution and many others around the world, endonasal endoscopic
surgery for sellar lesions utilizes a binostril technique with a neurosurgeon and
otolaryngologist working together throughout the majority of the procedure. The operation
is begun with a 4-mm 0-degree rod-lens endoscope with 30- and 45-degree endoscopes
available for use later in the procedure.
Patient Positioning
The patient's head is placed either in a horseshoe head-holder for standard transsphenoidal
cases or in a three-point Mayfield fixation for more extended endonasal procedures
that are anticipated to last over 6 hours. The head is tilted toward the left shoulder
and turned 20 to 30 degrees toward the right, the endotracheal tube is positioned
at the left side of the mouth, and the anesthesiologist and anesthesia equipment are
positioned on the patient's left side. This setup allows for both surgeons to have
a comfortable operative position on the patient's right. For sellar lesions, we position
the head in a neutral plane (0 degrees) relative to the floor; when removing lesions
primarily in the suprasellar region, 10 to 15 degrees of neck extension is applied;
for infrasellar and clival lesions, 10 to 15 degrees of neck flexion is used. Following
positioning of the head, the surgical navigation mask (Stryker Navigation, Stryker
Corp, Kalamazoo, MI) is placed on the patient's face, and the system is registered
to the preoperative MRI and/or CT angiogram. Finally two high-definition video monitors
displaying the endoscopic picture are positioned at almost 90-degree angles to one
another, allowing for each to be directed at one of the two operating surgeons. A
third monitor for neuronavigation is placed between the two high-definition monitors
([Fig. 1 ]).
Fig. 1 Artist representation of the room setup for endoscopic endonasal transsphenoidal
surgery. Two high-definition (HD) monitors are positioned at oblique angles to allow
for comfortable working position for both the otolaryngologist (ENT) and the neurosurgeon
(neuro).
Nasal Preparation and Approach to Sphenoid Sinus
The nasal cavity is prepared prior to beginning the approach by spraying oxymetazoline
in both nares. The face, perinasal area, and right lower abdominal area (for potential
fat graft) are then sterilely prepped and draped. The initial approach into the sphenoid
sinus is performed normally by the otolaryngologist alone with the 0-degree 4-mm endoscope.
Xylocaine 1% with 1:100,000 epinephrine is first injected into the middle turbinates
and posterior nasal septum bilaterally. Both inferior turbinates are first in-fractured
then out-fractured. Similarly, the middle turbinates are out-fractured, exposing the
sphenoid ostia. Next, bilateral nasoseptal “rescue flaps” are elevated.[21 ] A curved microtip unipolar electrocautery is used to incise the mucoperiosteum beginning
immediately inferior to the sphenoid ostium and extending to a point ∼2 cm anteriorly,
along the inferior vomer and posterior nasal septum. The two “rescue flaps” are then
pushed downward toward the nasopharynx to minimize obstruction into the sphenoid sinus
and are kept in place during the procedure with oxymetazoline-soaked cottonoids, preserving
the posterior nasoseptal artery on both sides. The septal olfactory strip is preserved
bilaterally and elevated away from the perpendicular plate of the ethmoid bone ([Fig. 2 ]). This technique allows for preservation of the olfactory fibers and opens additional
space in which the endoscope can be positioned during the remainder of the procedure.
Next, a posterior septectomy is performed with a back-biter to connect the right and
left nasal cavities. A wide sphenoidotomy is then performed that extends lateral to
the sphenoid ostia and generally to the floor of the sphenoid inferiorly and to the
roof of the sphenoid/ethmoid junction, superiorly. An attempt is made to remove the
vomer in one piece to preserve the bone for sellar floor reconstruction if necessary.
Posterior ethmoid air cells are also opened and removed as necessary to facilitate
better superior exposure. This is often done using the 30-degree angled-lens endoscope.
Fig. 2 Artist representation of the nasal septal anatomy, vasculature, and location of olfactory
fibers. Proper placement of the rescue flap incision allows elevation and preservations
of the septal olfactory strip. Abbreviations: PS, posterior septum; SOS, septal olfactory
strip.
Sellar Exposure
Once the sphenoid sinus has been entered, any bony septations limiting access to the
sella are carefully removed using a rongeur or high-speed drill with a 4-mm course
diamond bit. Special attention is paid to lateral septations as they often lead directly
to the petrous and cavernous carotid arteries. Aggressive removal or torquing of these
septations can result in carotid artery laceration. The mucosa over the sella is then
removed but the remaining sphenoid sinus mucosa is left intact to preserve as much
normal sinus architecture and functional tissue as possible. The sellar face is then
opened to expose the sellar dura. Depending on the size of the tumor and the degree
of surrounding invasion, bone is typically removed laterally from cavernous sinus
to cavernous sinus, superiorly to the tuberculum sella, and inferiorly to the sellar
floor. When possible, the floor is preserved to facilitate fat graft placement. The
opening is typically started with the drill and then completed with a 2-mm Kerrison
rongeur. For large macroadenomas, there is often extensive bony erosion or thinning
and dural invasion, which may extend out to and over the cavernous carotid arteries.
If this is found to be the case, it should be assumed that the bone over the carotid
arteries may also be eroded by tumor and extreme caution should be exercised before
definitive localization of both arteries has occurred.
Dural Opening
Prior to sellar dural opening, the location of the cavernous carotid arteries is precisely
determined with a micro-Doppler probe (10-MHz ES-100X MiniDop with NRP-10H bayonet
probe; Koven, St. Louis, Missouri, United States) and surgical navigation ([Fig. 3 ]).[22 ] The locations of the adenoma and pituitary gland should be anticipated based on
the preoperative MRI. Ideally, the dura is opened without entering into the gland
or transgressing the adenoma pseudocapsule. In most instances of macroadenomas, the
pituitary gland will be compressed laterally and/or superiorly, but occasionally,
a portion of the gland may be draped anteriorly over the tumor. A wide dural opening
is performed in u -shaped fashion with a standard straight microblade (Mizuho America Inc., Union City,
CA). Next, angled microdissectors are used to carefully separate and elevate the dura
from the underlying tumor and pituitary gland. The dural opening can then be extended
more laterally as needed with a right-angle microhook blade or curved microscissors.
Laterally, the opening will normally extend to within 1 to 2 mm of the medial wall
of the cavernous sinus. If cavernous sinus bleeding is encountered, this low-pressure
venous bleeding is generally easily controlled with Surgifoam (Ethicon Inc., Johnson
& Johnson Co., Piscataway, New Jersey, United States) or Gelfoam (Pfizer Inc., New
York, New York, United States) and gentle direct pressure. In patients with microadenomas
and a low-lying diaphragm, care must be taken to not extend the dural opening too
far superiorly as this may often result in an early and anterior cerebrospinal fluid
(CSF) leak.
Fig. 3 Artist representation of micro-Doppler usage for localization of the cavernous carotid
arteries.
Tumor Removal
Complete tumor resection with preservation or improvement in pituitary gland function
is the goal for all patients undergoing endonasal resection of a pituitary adenoma.
Oldfield and Vortmeyer were the first to describe the technique of adenoma removal
utilizing the tumor pseudocapsule in Cushing disease as a means of achieving complete
tumor resection.[23 ] This technique is particularly useful for microadenomas but can also be applied
in macroadenomas. For microadenomas, in which the tumor is behind a small rim of anterior
gland, an incision can be made in the gland at its thinnest point to reach the pseudocapsule,
which is a thin rim of compressed normal gland. A plane is then established between
the adenoma and normal gland using micro-dissectors, irrigation, and gentle traction.
The tumor is carefully separated from the compressed normal gland and gently removed
with the surrounding pseudocapsule intact.
For larger macroadenomas with suprasellar extension, it is often best to first inferiorly
and centrally debulk the tumor with ring curettes and suction. This initial decompression
allows the more superior portion of the tumor with its pseudocapsule intact to be
separated from compressed normal gland and the diaphragma sella and is often removed
in one remaining large rind. With this technique, one avoids pulling the tumor down
from the diaphragm sella without direct visualization, thus decreasing the likelihood
of CSF leak.
Once all the visualized tumor has been removed with the 0-degree endoscope, the 30-
or 45-degree angled lenses are utilized to obtain a clear view of regions not in direct
line of sight.[24 ] This angled view is especially helpful for tumors with extensive suprasellar or
cavernous sinus extension. The 45- and 90-degree up-angled ringed curettes may be
used along with angled suctions to probe the folds of the diaphragma to dislodge residual
tumor. In addition, a Valsalva maneuver or bilateral jugular vein compression can
be administered to encourage downward descent of any suprasellar tumor that remains
attached to the diaphragma. Ultimately, full inversion of the diaphragma into the
enlarged sella should be seen if complete tumor removal has been accomplished. In
cases of large macroadenomas, the redundant and collapsed diaphragma sella often falls
fully into the sella and obscures visualization of the sellar recesses. In such cases,
it is extremely helpful to elevate this tissue with a spatula dissector and/or a cottonoid
to facilitate an unobstructed view of these hidden regions to avoid missing residual
adenoma. In all tumors with suspected cavernous sinus invasion, inspection of the
medial cavernous sinus wall is performed with a 30- or 45-degree angled-lens endoscope.
If present, tumor within the medial cavernous sinus may be safely removed using angled
ring curettes and gentle suction. Venous bleeding from the cavernous sinus is once
again controlled with Surgifoam or Gelfoam. In contrast to tumor within the medial
cavernous sinus compartment, tumor that has extended along or lateral to the internal
carotid artery (ICA) is difficult to access safely, and removal is associated with
a higher risk of neurovascular injury. Monitoring and direct stimulation of cranial
nerves III and VI is helpful for this lateral, posterior cavernous sinus dissection.
Skull Base Reconstruction and Cerebrospinal Fluid Leak Repair
Once tumor removal is complete, the sellar resection area is irrigated with full-strength
hydrogen peroxide for approximately 1 minute and hemostasis is achieved with Surgifoam.
If there is significant diaphragmatic defect, irrigation with hydrogen peroxide should
not be performed. The type of skull base reconstruction performed depends primarily
on whether or not a CSF leak is present. Although the grading system for categorizing
CSF leaks we described in 2007 is still quite useful, the technical details of repair
have evolved in the endoscopic era.[25 ] All repairs involved the use of collagen sponge (Helistat Integra, Hudson NH) as
part of the reconstruction. In patients found to have no evidence of a CSF leak following
a Valsalva maneuver (grade 0), a single layer of collagen sponge placed over the exposed
dura is utilized as the only repair material. This is sealed in place using fibrin
glue. Exceptions include very large “dead space” defects or translucent diaphragm
sellae that may have potential for hemorrhage or leakage. In such cases, the dead
space is filled with an abdominal fat graft. In cases where a small amount of CSF
is detected following a Valsalva maneuver but no obvious diaphragmatic defect is visualized
(grade 1), a layer of collagen sponge is initially placed under the dural edges. An
intrasellar extradural buttress consisting of either the patient's previously harvested
bony septum or a Medpore polyethylene plate (Stryker, Kalamazoo, Michigan, United
States) is then placed over the collagen sponge followed by a second outer layer of
collagen. The repair is bolstered with tissue glue (DuraSeal, Integra US, Hudson NH;
or Tisseel, Baxter Healthcare Corp., Deerfield, IL). Medium-sized CSF leaks (grade
2) or grade 1 leaks with a large amount of intrasellar dead space require the placement
of an abdominal fat graft. After harvesting from the lower abdomen, the fat graft
is initially placed within the intrasellar space, taking care not to re-create too
much mass effect on the suprasellar neurovascular structures. Again, this is followed
by an intradural layer of collagen sponge and in some instances an intrasellar extradural
rigid buttress. A second layer of fat and collagen is typically placed over this construct
and bolstered with tissue glue. To assess the adequacy of the repair, prior to placing
tissue glue, the anesthesiologist is asked to perform a Valsalva maneuver to raise
the patient's intracranial pressure. In some instances of grade 1 and 2 leaks in which
a buttress is needed but no lateral bone edges are available to safely wedge the buttress
in place, a temporary buttress with a Merocel sponge (Medtronic Inc, Minneapolis,
MN) is placed in one or both nostrils for up to 5 days (continuous antibiotics while
the packs are in place are necessary to prevent toxic shock syndrome). Large defects
(grade 3) are typically seen only in extended suprasellar approaches for tumors such
as craniopharyngiomas or tuberculum meningiomas. The repair for such grade 3 leaks
consists of virtually the same construct utilized to treat a grade 2 leak with the
addition of a nasoseptal flap that is held in place with bilateral Merocel nasal packs,
left in place for 5 days.[26 ] Any patient with an intraoperative CSF leak is placed on acetazolamide for 48 to
72 hours after surgery to decrease CSF production. CSF diversion with lumbar drainage
is rarely employed.
Closure
Following the sellar reconstruction, blood is suctioned from the sphenoid sinus, nasal
cavity, and nasopharynx, and good nasal mucosal hemostasis is obtained to minimize
the amount of blood that is swallowed in the immediate postoperative period. If a
nasoseptal flap is necessary to complete our repair, one of the two rescue flaps may
be extended in the standard fashion. If not, then the rescue flaps are carefully elevated
back along the remaining portion of the vomer and inferior nasal septum to their original
location. The middle turbinates are then repositioned anatomically. Nasal packing
is typically not utilized unless a buttress is needed as described above to help hold
the repair in position. To minimize chances of a postoperative CSF leak, nasal epistaxis,
or intrasellar bleeding, excessive coughing or “bucking” should be avoided during
extubation and blood pressure carefully monitored and controlled in the early postoperative
period. The pharyngeal and gastric contents are typically aspirated of blood products
to reduce postoperative nausea/vomiting. Throat packs are not routinely used as they
are abrasive to the oropharynx, frequently resulting in patient discomfort.
Indications and Case Illustrations
The endoscopic endonasal transsphenoidal approach and its superior, inferior, and
lateral extensions is ideal for a wide variety of sellar and parasellar lesions and
is now the most common route for access to the pituitary gland. Lesions accessible
through this route include pituitary adenomas, Rathke cleft cysts, clival chordomas,
craniopharyngiomas, and sellar arachnoid cysts as well as tuberculum sellae, cavernous
sinus, and petroclival meningiomas. The following cases are illustrative examples
of lesions that are best suited for this technique.
Case 1—Nonfunctioning Macroadenoma
The patient was a 39-year-old man with headaches and rapid visual loss in the left
temporal field. Ophthalmologic examination demonstrated a bitemporal hemianopsia,
which was more pronounced on the left. The patient was also found to have a left afferent
pupillary defect as well as decreased visual acuity (20/100) in the left eye. The
remainder of his neurologic exam was unremarkable. An MRI of the brain demonstrated
a 25 × 29 × 19-mm sellar mass with suprasellar extension, causing severe chiasmal
compression ([Fig. 4 ]). Pituitary hormonal testing revealed low T4 (3.5 μg/dL), inappropriately normal
thyroid-stimulating hormone (1.95 μIU/mL), low free (23.5 pg/mL) and total (72ng/dL)
testosterone, normal adrenocorticotropic hormone (58 pg/mL) and cortisol (10.6 μg/dL),
elevated IGF-1 (388 ng/mL) and prolactin (1,253 ng/mL). The patient did not exhibit
any of the usual stigmata of acromegaly or hyperprolactinemia. Due to the acute nature
of the patient's visual loss and the concern for a combined growth hormone/prolactin
cosecreting tumor, surgical intervention was performed. The patient underwent an endoscopic
endonasal transsphenoidal tumor resection with gross total removal. Pathology was
consistent with a typical pituitary adenoma with immunohistochemical staining positively
for both growth hormone and prolactin and an elevated Ki -67 proliferative index of 5 to 7%.
Fig. 4 Case 1. (A, B) Preoperative T1-weighted images with gadolinium demonstrating a hypoenhancing
sellar and suprasellar mass with significant optic chiasmal compression and right
cavernous sinus invasion. (C, D) Three-month postoperative images demonstrate no evidence
of residual or recurrence.
The sellar floor was repaired in the standard fashion using collagen sponge, abdominal
fat graft, nasal septal bone, and fibrin glue. Although an intraoperative CSF leak
was not observed, the decision to reinforce the sellar repair with an abdominal fat
graft was made due to the large nature of the tumor and to prevent excessive diaphragmatic
and optic chiasmal herniation. The patient initially tolerated the procedure well
without complications. His postoperative day 2 prolactin level decreased to 53.7 ng/mL.
He noticed an immediate improvement in his visual function postoperatively and was
discharged home on postoperative day 2. At 3-month follow-up his visual acuity is
20/20 OU and his visual fields are full to confrontation. His follow-up 3-month postoperative
MRI demonstrated no obvious residual tumor with a small questionable nodule in the
right posterior cavernous sinus region ([Fig. 4 ]).
Case 2—Tuberculum Sella Meningioma
A 65-year-old woman presented with transient blurry vision. On further questioning,
she reported gradual visual deterioration over several years. Visual acuity testing
revealed 20/25 OU. Formal visual field testing revealed bilateral optic nerve dysfunction
but without respecting the vertical plane; ophthalmoscopy confirmed bilateral optic
atrophy. Sellar MRI demonstrated a 25 × 24 × 23-mm contrast-enhancing suprasellar
mass arising from the tuberculum sellae causing severe elevation of the optic apparatus,
worse on the right, consistent with a tuberculum sellae meningioma that partially
encased the left A2 segment. She had an extended endoscopic transsphenoidal, transplanum
tumor removal. At surgery, the tumor was fibrous and significantly adherent to the
right optic nerve and chiasm. Due to the adherent nature of the tumor, a small, 1-
to 2-mm fragment of tumor was left along the inferior aspect of the right optic nerve
and along the left A2 segment. A grade 3 CSF leak was repaired with collagen sponge,
abdominal fat graft, nasoseptal flap, and fibrin glue. Two Merocel nasal packs were
left in place as a temporary buttress to hold the repair. She tolerated the procedure
well without complications and was discharged home on postoperative day 3. She had
transient worsening of her right temporal visual field, which improved by discharge.
She had transient diabetes insipidus, requiring one dose of desmopressin (DDAVP),
which then resolved and cortisol levels remained normal after surgery. Pathology confirmed
a World Health Organization grade 1 meningothelial meningioma. Her MRI on the first
postoperative day demonstrated no obvious evidence of residual tumor, despite the
known small residual at the time of surgery. She will have another MRI at 3 months
postsurgery ([Fig. 5 ]).
Fig. 5 Case 2. (A, B) Preoperative T1-weighted images with gadolinium demonstrating a homogeneously
enhancing lesion arising from the tuberculum sellae causing significant elevation
of the optic chiasm. (C, D) Initial postoperative images demonstrate a gross total
resection with no evidence of residual.
Lessons Learned—Endoscopic Technique and Technology Applied to Intracranial Surgery:
The Supraorbital Eyebrow Craniotomy
As stated by Wilson more than 4 decades ago, “The ideal exposure is one which is large
enough to do the job well, while preserving the integrity of as much normal tissue
as possible.” (p. 106)[27 ] The supraorbital eyebrow craniotomy is a well-described, minimally invasive keyhole
technique through a small anterolateral craniotomy that provides access to a wide
range of anterior and midline skull base pathologies including the anterior fossa
floor, the parasellar region, proximal sylvian fissure, circle of Willis, basal frontal
lobe, and ventral brainstem.[14 ]
[28 ] As we described several years ago, this approach is now routinely used as an alternative
or complementary approach to the endonasal endoscopic approach for parasellar tumors.[29 ]
Determining Which Approach for Which Parasellar Tumors
The most commonly encountered surgical lesions amenable to the supraorbital approach
are tuberculum sellae, planum and anterior clinoid meningiomas, some olfactory groove
meningiomas, craniopharyngiomas, and intra-axial tumors of the orbitofrontal region,
frontal pole, and medial temporal lobe.[14 ]
[28 ]
[29 ]
[30 ]
[31 ] We have employed the supraorbital approach most commonly for select tuberculum sella
meningiomas that are over 3 cm in maximal diameter and have vascular encasement or
extend well lateral to the supraclinoid carotid arteries or optic nerves. The supraorbital
route is used for craniopharyngiomas that are not predominantly in the retrochiasmatic
space and instead extend lateral to the supraclinoid carotid arteries or into the
anterior cranial fossa, as well as some recurrent craniopharyngiomas.
A major benefit of the supraorbital approach over the endonasal endoscopic approach
is the simplified closure, which is in stark contrast to the extended endonasal approach
that typically requires a nasoseptal flap. Although the benefits of the supraorbital
approach are numerous, there are also several key limitations and cautions. These
include narrow viewing angles and the necessity for near coaxial control of the microinstruments
used through narrow anatomic windows. As with transsphenoidal approaches, one of the
most important developments to aid in the evolution of transcranial approaches is
the endoscope. The most important disadvantage of a small, less-invasive keyhole approach
is the loss of intraoperative light and sight, causing significantly reduced optical
control during surgery. For the purpose of bringing light into the surgical field
with adequate magnification, the optical properties of modern surgical microscopes
can be effectively supplemented with true high-definition endoscopes. Advances in
fiberoptic lighting, smaller lenses, and digital camera technology have recently allowed
the endoscope to be used for intracranial neurosurgery. Along with these technological
advances, neurosurgeons, by building endoscopic skills through transsphenoidal approaches,
have increased their experience and versatility with these new techniques. In a series
of 450 supraorbital craniotomies, endoscope-assisted microsurgical techniques were
used in 135 cases, giving the advantage of higher light intensity, a clear depiction
of details in close-up positions, and an extended viewing angle.[32 ] With the use of angled endoscopes, “blind” corners of the surgical field can be
safely controlled without additional extension of the craniotomy. These areas would
have been otherwise inaccessible through the keyhole approach using microscopic visualization.
Although the supraorbital approach offers access to most of the anterior fossa floor,
parasellar area, and medial aspect of the middle fossa, there are four specific areas
that can be difficult to reach and adequately visualize ([Fig. 6 ]). These include the anterior aspect of the olfactory groove, the sellar floor, the
region under the ipsilateral optic nerve, and the anterior aspect of the middle fossa
under the sphenoid ridge. The midline depression of the olfactory groove along the
anterior skull base typically lies below the line of sight provided by the operating
microscope. Likewise, a surgical trajectory along the orbital roof will not provide
a direct line of sight into the sella itself. Similarly, the area directly under the
ipsilateral optic nerve cannot be visualized with microscope without undue optic nerve
manipulation. However, reaching these three relative blind spots is possible with
the use of an angled endoscope and angled instrumentation. The fourth anatomic limit
of the supraorbital approach is the lesser wing of the sphenoid. Lesions with significant
extension below the lesser sphenoid wing into the far anterior aspect of the middle
cranial fossa may not be accessible from a supraorbital craniotomy. In such cases,
a traditional pterional or minipterional craniotomy may be required to achieve access
inferior and anterior to the sphenoid ridge.[14 ]
Fig. 6 Artist representation of the surgical access provided by the supraorbital craniotomy.
Surgical Technique—Supraorbital “Eyebrow” Craniotomy
The operation is performed under general anesthesia with placement of a Foley catheter
and arterial line in most cases. The patient's head is fixed in a Mayfield three-pin
head holder, angled ∼30 degrees away from the operative side and extended such that
malar eminence is prominent, similar to traditional pterional positioning. The Stryker
navigation system is attached to the patient's head and is registered to the surgical
navigation MRI. The ipsilateral upper quadrant of the abdomen is marked and prepped
in case an abdominal fat graft should become necessary. The eyebrow incision is marked
and prepped and draped in usual sterile fashion. The eyebrow incision is made, starting
from just medial to the supraorbital notch and extending laterally in the midst of
the eyebrow to its termination with ∼5-mm extension in that trajectory. Sharp dissection
is taken down to the pericranium, and a subgaleal plane is created in the supraorbital
region. The dissection extends laterally to the superior part of the temporalis fascia
and just lateral to the superior temporal line. The supraorbital nerve is identified
medially and preserved. Multiple fishhooks are placed on the superior aspect of the
incision to expose the supraorbital area. A pericranial cuff is then cut, extending
from immediately lateral to the supraorbital nerve, coursing superiorly in the arc
of the planned craniotomy, over the superior temporal line and into the temporalis
fascia and muscle. This pericranial cuff is elevated with subperiosteal dissection
and retracted over the brow with a stitch.
A single bur hole is placed with a “matchstick” bit just inferior and lateral to the
superior temporal line and behind the frontozygomatic process. The underlying dura
is then elevated from the overlying bone. The craniotomy extends from the bur hole,
medially along the supraorbital ridge. Although effort should be made to avoid entering
the frontal sinus, this should not be at the expense of limiting the extent of the
craniotomy, as maximizing the inferior extent of the craniotomy is critical both for
visualization and for maneuvering of surgical instruments. The standard supraorbital
craniotomy is 2 to 2.5 cm wide and 1.5 to 2 cm tall ([Fig. 7 ]). Again, proper positioning of the craniotomy as low as possible on the anterior
fossa floor is important to obtain an optimal surgical trajectory. If the frontal
sinus is entered, temporary occlusion accomplished with an iodine-soaked Gelfoam,
and definitive repair is completed with an abdominal fat graft during closure. To
facilitate exposure, the inner surface of the calvarium is drilled inferiorly along
the floor of the frontal fossa including any protuberances of the orbital roof. This
maneuver is essential to optimize the flat surgical trajectory along the frontal floor,
as even small boney ridges may significantly impair the line of sight to the sellar
region. The dura is then opened in a u -shaped fashion and reflected inferiorly over the orbital rim. The dural flap is kept
moist and under tension throughout the case to prevent shrinking and allow for a watertight
closure. Immediately after dural opening, a small arachnoid incision into the opticocarotid
cistern is made to allow for CSF egress as the microscope is brought into position.
The remainder of the procedure is completed with a combination of microscopic and
endoscopic visualization. Brain retractors are used infrequently.
Fig. 7 Intraoperative photograph of a left supraorbital craniotomy demonstrating bur hole
and craniotomy placement.
Under microscopic visualization, the subfrontal corridor is dissected. The frontal
lobe is protected with a strip of Telfa (Integra US, Hudson, NH), and the ipsilateral
optic nerve and supraclinoid carotid artery are identified. The arachnoid of the opticocarotid
and optico-oculomotor cisterns is opened widely to allow further CSF egress. The release
of CSF is essential to achieve adequate brain relaxation and avoid frontal lobe retraction.
Additional dissection of arachnoid at the base of the frontal lobe and within the
proximal sylvian fissure will free the frontal lobe from the basal cisterns and temporal
lobe and allow it to fall away with gravity.
Following these maneuvers, a field of view including the ipsilateral proximal sylvian
fissure, the ipsilateral third nerve, both optic nerves, the optic chiasm, the lamina
terminalis, both supraclinoid carotid arteries, both A1 segments, the anterior communicating
artery, both A2 segments, and the pituitary stalk is possible. Additional dissection
through the opticocarotid or optico-oculomotor windows will expose the ventral brainstem,
basilar artery, and other posterior circulation vessels and perforators. Standard
microsurgical dissection and tumor removal then proceeds with care taken to preserve
arachnoid membranes and key neurovascular structures such as the superior hypophyseal
arteries. Tumors densely adherent to the optic apparatus or circle of Willis vessels,
which is common with craniopharyngiomas, are often best left behind to avoid new neurologic
deficits.
With the use of a 30- or 45-degree angled endoscope, a view over the tuberculum sellae
into the pituitary fossa, over sphenoid ridge into the middle fossa, over the dorsum
sellae into the prepontine cistern, and over the orbital roof to the cribriform recess
is possible.[14 ] When necessary, the supraorbital approach can be used to reach as far posteriorly
and inferiorly to the ventral brainstem and superior one-third of the clivus.
The dural closure involves reapproximation and primary closure (when possible) of
dural edges, followed by placement of a large piece of collagen sponge. Fat grafts
are utilized to seal large defects of the frontal sinus. These are reinforced and
secured with Tisseel. The bone flap is reapproximated with a lateral bur hole cover
and a medial straight plate with 4-mm titanium screws. The bone flap is placed to
eliminate any gap superiorly, and the gap inferiorly is filled with collagen matrix.
The incision is irrigated with antibiotic-containing saline. The pericranial flap
is repositioned with multiple 3–0 Vicryl stitches (Johnson & Johnson, New Brunswick,
NJ). The scalp is closed with interrupted inverted 3–0 Vicryl sutures followed by
a running subcuticular 5–0 Monocryl stitches (Johnson & Johnson, New Brunswick, NJ)
Antibiotic ointment, sterile dressing, and gentle head wrap compressive dressing is
applied to help avoid a hematoma. Care is taken to avoid excess pressure with the
head wrap to prevent pressure urticaria or skin necrosis.
Supraorbital Craniotomy—Case Illustration
Case 3—Clinoidal Meningioma
A 67-year-old woman suffered a ground-level fall and was found to have an incidental
right paraclinoidal meningioma on CT. An MRI further characterized the lesion as a
homogeneously enhancing 26 × 22 × 22-mm mass arising from the right anterior clinoid
and medial sphenoid wing with an associated arachnoid cyst. Moderate surrounding vasogenic
edema and 6 mm of right-to-left subfalcine herniation were noted. The imaging characteristics
were consistent with an anterior clinoidal meningioma. Her neurologic examination
was nonfocal. Due to the tumor extending lateral to the optic nerves and supraclinoid
carotid artery along the sphenoid wing, a right supraorbital craniotomy was selected
for resection. The 30-degree angled endoscope was invaluable in visualizing and resecting
tumor along the lateral wall of the cavernous sinus and extending into the sylvian
fissure. At the time of surgery, it was felt that gross total resection was achieved.
The patient tolerated the procedure well and there were no complications. Postoperative
follow-up MRI demonstrates a small amount of persistent enhancement along the clinoidal
dura and medial sphenoid wing. Two years after surgery, this has been stable and has
not demonstrated any evidence of tumor progression or recurrence ([Fig. 8 ]).
Fig. 8 Case 3. (A, B) Preoperative T1-weighted images with gadolinium demonstrating a homogenously
enhancing lesion arising from the right clinoid and medial sphenoid wing with an associated
arachnoid cyst. (C, D) Two-year postoperative images demonstrate no evidence of residual
or recurrence.
