Modern transsphenoidal approaches began with the contributions of Jules Hardy, who built on the legacy of earlier transsphenoidal approaches from Cushing’s to Norman Dott to Gerard Guiot. Hardy introduced the use of the operating microscope and the concept of the microadenoma, reinvigorating and reintroducing transsphenoidal surgery for pituitary tumors and other lesions in and about the sella.
Since the original reports in 1969, the microscopic transsphenoidal approach has undergone progressive modifications. Based on the contribution of Huw Griffith and others, the sublabial approach has largely been supplanted by primary endonasal techniques, which have become increasingly direct and less invasive; frameless stereotaxy has been a welcome and useful adjunct, and the corridors accessible through the transsphenoidal approach have been extended. The introduction of the endoscope has added significantly to the versatility of the transsphenoidal surgeon.
In this chapter, we review the essential details of the microscopic transsphenoidal approaches with a focus on surgical technique, based on a personal series of over 5000 transsphenoidal procedures performed by the senior author (ERL).
Concepts of the Approach
Among the variations of the microscopic transsphenoidal approach are the “standard” sublabial submucosal transseptal approach, the transnasal septal displacement/septal pushover approach, and the direct transnasal sphenoidotomy.
Preoperative considerations in planning the microscopic transsphenoidal approach include anatomical and pathological characteristics, which may influence the choice of surgical corridor and more general considerations that can apply to all transsphenoidal approaches. Important anatomical factors include the age of the patient; the size of the nose and nostril; the presence of a nasal septal deviation, lateral septal spurs, or septal perforations; a prior history of septal surgery; the presence of associated sinus disease or infection; the size, type, and septation of the sphenoid sinus; and, in sublabial approaches, the presence of dentures, capped teeth, or root canal surgery involving the anterior incisors. Significant lateral parasellar cavernous sinus extension or aspects of suprasellar tumor extension may influence which nostril is used for initial exposure. In such circumstances, modifications of the transsphenoidal approach may be appropriate, such as a “cross-court” trajectory from one nostril to the opposite cavernous sinus.
Other preoperative considerations contributing to the safety of the procedure and its potential complications include being certain that adequate endovascular support is available to deal with potential carotid artery injury. It is important that the preoperative imaging studies are comprehensive and of excellent quality, and that any image guidance platforms to be used are precise, accurate, and fully functional.
Common transnasal microscopic transsphenoidal approaches include the transseptal submucosal technique, the septal displacement (“septal pushover”), and the direct sphenoidotomy. The location of the initial incision within the nose and the resultant width of the sellar exposure distinguish these approaches. The most traditional of the endonasal approaches is the transseptal submucosal approach. This begins with a hemitransfixion incision made just inside the alar ring. In the septal pushover approach, a vertical incision is made more posteriorly, at the junction of the osseous and cartilaginous septum. The direct sphenoidotomy begins more posteriorly, at the posterior septal articulation with the rostrum of the sphenoid. Potential advantages of a deeper nasal incision include the avoidance of septal trauma, complications such as symptomatic nasal septal perforation, submucosal hemorrhage, and postoperative discomfort. Potential disadvantages include a diminished width of exposure and an off-midline trajectory.
Patient positioning is critical in transsphenoidal surgery and is identical in the submucosal and endonasal variants. Although some surgeons prefer to operate with the head extended and the surgeon standing at the head of the operating table (a modified cleft palate position), we prefer to place the patient in a semirecumbent position with the surgeon facing the patient anteriorly. In either case, it is important to use a standard and reproducible method of positioning to maintain a midline approach, and to provide a comfortable and ergonomic position both for the patient and for the operating surgeon. In our practice, the patient is prepared so that the surgeon will be at the supine semirecumbent patient’s right, with the patient’s head angled and body turned so that the surgeon is operating without leaning over the patient. To accomplish this, we place the patient with the right shoulder at the upper right-hand corner of the table to facilitate placement of the head in a Mayfield headrest ( Figure 11-1 ). The operating table is positioned so that the thorax is elevated 25 to 30 degrees while gently flexing the knees, achieving a semirecumbent, lawn chair position, with the head above the heart to reduce venous pressure and to allow venous blood to drain from the sphenoid and sella. Significant venous air embolism has not occurred in our experience using this position. This attempt to reduce venous bleeding is particularly important in patients with Cushing’s disease and in surgery for microadenomas.
The head is laterally flexed 20 to 30 degrees to the left—with the left ear toward the left shoulder. Once laterally flexed, the head is positioned so that the surgeon’s view into the nose aims toward the sella. A reliable way to achieve this in most patients is to ensure that the nasal dorsum is parallel to the floor (see Figure 11-1 ). The head can then either be supported by a Mayfield horseshoe headrest—having as its main advantage the ability to subtly manipulate the head position during surgery—or placed in a rigid three-point fixation, which we currently use when we employ frameless stereotactic image guidance during surgery. The operating table is tilted slightly to the right to allow the surgeon to operate comfortably, sitting or standing, with a straight on line of sight as opposed to leaning over the patient. Tilting may also help when positioning patients with cervical spondylosis in whom lateral neck flexion is difficult. The right arm is then tucked against the patient’s side.
Once the patient is positioned on the table and in the headrest, we adjust the operating room table so that the patient’s head is parallel to the walls of the room. In addition to making the midline of the head concordant with the midline of ambient space, this head positioning is especially useful for radiology technicians or other staff when aligning the fluoroscopic C-arm, if necessary, during surgery.
The assisting surgeon is positioned to the left of the operating surgeon, viewing through the observer objective of the microscope. The scrub nurse or technician is positioned across the patient from the operating surgeon so that instruments are easily exchanged while the principal surgeon remains in an optimal position and does not have to turn away from the microscope. The anesthesia team is based at the foot of the patient. The microscope enters from the left of the surgeon, and the image guidance system can be positioned on the left side of the patient across from the surgeons.
Skin preparation for surgery comprises preparing the face, nares, and umbilical region. A lumbar intrathecal drain is inserted if indicated. Intraoperative antibiotics are routinely administered along with cortisol support if preoperative testing reveals steroid deficiency. The anesthesia team is asked not to administer dexamethasone for prevention of postoperative nausea. We prepare the skin of the face with an aqueous antiseptic solution. To help minimize nasal mucosal bleeding during the approach, we apply topical vasoconstrictors and usually inject a local anesthetic solution (0.5% xylocaine with 1:200,000 epinephrine). We spray oxymetazoline (Afrin) into the nose before induction and then pack both nostrils with cotton pledgets soaked in 5% cocaine, and leave these in for 10 to 15 minutes. During this time, the remainder of the patient preparation and registration for image guidance is performed. We prepare the abdomen in the subumbilical region for a possible incision to obtain a small fat graft when necessary; the incision is usually made in the fold just below the umbilicus and is 2 cm in length. For patients harboring lesions with large suprasellar extensions, which may be driven down by injecting air or saline intrathecally, a lumbar drain may be placed at this time.
Ordinarily one dose of a broad-spectrum antibiotic is given, and if the patient is deficient in cortisol before surgery, steroid support in the form of hydrocortisone, usually 100 mg intravenous Solu-Cortef, is given at the beginning of surgery. Cortisol support is continued as indicated in the postoperative period, but antibiotics are given ordinarily as just one dose unless the nose is packed for prolonged periods of time. The most widely used intraoperative imaging device is the C-arm videofluoroscope. Most often, a lateral image confirms the appropriate trajectory to the sella turcica and is also used to confirm its superior and inferior confines. Knowing the superior and inferior limits of the sella turcica allows the surgeon to confirm adequate exposure and prevents inadvertent opening of the planum sphenoidale and cerebrospinal fluid leak. The advantage of using a standard fluoroscope is its simplicity and accuracy. Its disadvantages are the radiation exposure and its inability to depict soft tissue anatomy, including the tumor and neurovascular structures.
The proximity of sellar and parasellar masses to the optic apparatus and carotid arteries demands absolute surgical precision in the approach to this region. The increasing sophistication of image-guidance platforms allowing highly accurate and precise instrument tracking on a coregistered preoperative magnetic resonance image (MRI) scan has influenced our and others’ practices, and we now perform virtually every transsphenoidal operation with the aid of frameless stereotaxy. Several image-guidance platforms are available that require coregistration of the patient’s anatomy with a preoperative CT or MR image. The particular system we employ requires the fixation of a rigid fiducial bearing arm to the Mayfield headrest, which we angle off to the left side of the patient’s head. Its main utility is in the initial stages of the transsphenoidal approach to the sella, as such systems provide accurate information regarding operative trajectory and recognition of the midline, proximity to the sphenoid, sella, and cavernous carotid arteries. The sagittal view is especially helpful in tracking the trajectory to the sphenoid and sellar face, and the coronal and axial views are useful in verifying the midline and helping to prevent accidental entry into the cavernous sinus or carotid arteries.
Although frameless systems rely on preoperatively acquired image data, intraoperative MRI (iMRI) systems providing real-time images have been used in transsphenoidal surgery; their main benefit is in the immediate assessment of the extent of resection, particularly in cases with suprasellar extension.
A useful adjunct in the attempt to prevent cavernous carotid injury is a micro-Doppler probe, such as a 20-MHz surgical Doppler (Mizuho America, Beverly, Mass.). It is a helpful tool in identifying the carotid arteries before incising the sellar dura and during removal of lateral portions of sellar lesions. Instruments such as this may be increasingly applicable as centers adopt more direct transsphenoidal approaches (e.g., septal pushover, direct sphenoidotomy), which tend to provide a surgical trajectory just off the midline, with a more restricted exposure. Moreover, the development of extended transsphenoidal approaches, wherein wider sellar exposures are sought for an expanding range of cranial base lesions, may render the carotids more vulnerable.
The transsphenoidal corridor is accessed by either a transnasal or a sublabial approach. Trends in transsphenoidal surgery have included a shift towards endonasal approaches featuring increasingly direct approaches to the sphenoid. The microscopic transsphenoidal operation can be divided into several distinct phases:
The nasal phase, from initial sublabial or endonasal incision to entry into the sphenoid sinus;
The sphenoid phase, from entry into the sphenoid sinus to the sellar dura;
The sellar phase, from opening of the sellar dura to lesion resection to establishment of hemostasis and preparation for closure;
The reconstruction and closure phase
This initial part of the operation comprises the surgical maneuvers required to establish a corridor to the sphenoid sinus. The nasal cavity may be entered directly (transnasal) or through a sublabial approach. With either of these approaches, a submucosal transseptal approach may be employed to enter the sphenoid. We also discuss the more direct transnasal approaches to the sphenoid, the transnasal septal displacement (“septal pushover”), and direct sphenoidotomy.
Transnasal submucosal transseptal
This approach requires considerable submucosal dissection, the main advantage of which is broad septal mobilization, a wide surgical corridor, a potentially more sterile field, and strict maintenance of the midline. Sinonasal complications and postoperative discomfort associated with the submucosal transseptal approaches and the refinement of more direct approaches to the sphenoid have lessened the popularity of submucosal transseptal approaches. Nasal packing is usually required to prevent submucosal septal hematoma, and can result in facial pain and headache in as many as 35% of patients. These procedures can cause rhinological complaints, including alveolar numbness, anosmia, saddle nose deformity, and nasal septal perforations.
To diminish bleeding during the submucosal dissection, we inject 10 to 20 cc of 0.5% Xylocaine with 1:200,000 epinephrine along the inferior and lateral aspects of the nasal septum with an attempt to dissect the nasal mucosa away from the cartilaginous septum with the injection. The transseptal approach begins with a right-sided hemitransfixion incision in the right nostril with the columella retracted to the patient’s left, facilitating the dissection of the right anterior nasal mucosal tunnel away from the septum ( Figure 11-2 ). The inferior border of the cartilaginous septum is exposed with sharp dissection, and one side of the septum is exposed submucosally with a combination of sharp and blunt dissection, thereby creating the anterior tunnel. The dissection continues posteriorly, elevating the nasal mucosa away from the cartilaginous septum back to the junction with the bony septum. A vertical septal incision is then made at this junction, and bilateral posterior submucosal tunnels are created on either side of the perpendicular plate of the ethmoid. The articulation of the cartilaginous septum with the maxilla is then dissected free, and the inferior mucosal tunnel on the opposite side is raised so that the cartilaginous septum can be displaced laterally without creating inferior mucosal tears. A self-retaining nasal speculum can then be introduced to straddle the perpendicular plate of the ethmoid, exposing the face of the sphenoid sinus. If the nostril is too small to accommodate a standard speculum, an inferior extension of the hemitransfixion incision within the alar ring or a relaxing right alotomy may be performed.