Endoscopic transseptal transsphenoidal hypophysectomy with three-dimensional intraoperative localization technology.

Abstract

The transsphenoidal approach to the sella was introduced into the English literature by Cushing in 1910. This approach soon fell out of favor because of poor visualization and complications.1 The intracranial approach to sellar lesions was the standard until 1969, when Hardy reintroduced the transseptal, sublabial, and transsphenoidal approaches.2-5 The introduction of improved microscopic techniques and intra-operative fluoroscopy led to the acceptance of these new approaches.3-5 The essential components of the transsphenoidal operation have not changed in more than 80 years. Only in recent years with advances in nasal endoscopy have new surgical techniques been considered. Since 1985,6 with the popularization of nasal endoscopic surgery, surgeons have attempted to use the endoscope in surgery previously reserved for the microscope. Improved visualization and ease of use have led several surgeons to successfully utilize the endoscope in pituitary surgery.5, 7, 8 Frameless three-dimensional image localization technology is a state-of-the-art advancement in endoscopic surgery.9 This technology allows for the visualization of the surgeon's probe in all three orthogonal planes, in real time. We have reviewed our large series of endoscopic surgeries utilizing VTI's (Visualization Technology, Inc.) Instatrak three-dimensional localization system and have found it to be very accurate and helpful in the majority of cases.10 We have recently collaborated with the department of neurosurgery to utilize an endoscopic approach and apply intraoperative localization technology to pituitary surgery. We present our technique of a transseptal transsphenoidal hypophysectomy with three-dimensional intraoperative image localization technology. Six patients with pituitary tumors underwent surgery utilizing the transseptal transsphenoidal hypophysectomy with three-dimensional intraoperative image localization technology. All patients underwent preoperative magnetic resonance imaging, as well as endocrinologic and ophthalmologic evaluations. Five patients had surgery for pituitary macroadenomas and one patient had a craniopharyngioma. Each patient obtained a preoperative computed tomography (CT) scan with the VTI headset in place. This CT scan was then loaded onto the VTI console. On the day of surgery the VTI headset was repositioned on the patient's head and the probe (a ball on the suction tip) was registered and verified to the headset. Electromagnetic technology was used to track the position of the probe in relation to the sensor on the patient's headset. The procedure is performed under general anesthesia. Neurosurgical pledgets soaked in 4% cocaine are placed in the nasal cavity for 10 minutes. The anterior aspect of the septum is infiltrated with 2% lidocaine and 1:50,000 epinephrine. Through a left hemitransfixion incision the left mucoperichondrial flap is elevated. A left nasal floor tunnel is connected with the left septal flap. A right nasal floor tunnel is created and then the cartilaginous and bony septum are disarticulated from the maxillary crest and pushed toward the right side of the nose. This wide exposure then allows for the easy visualization of the sphenoid rostrum. Using a 0° endoscope mucosa is elevated off of the rostrum exposing the sphenoid ostia bilaterally. The Instatrak probe is placed at the rostrum to confirm the midline position prior to entering the sphenoid sinus. A sphenoidotomy is created with 2-mm chisels and enlarged to the ostium with backbiting Kerrison punches. With the aid of the sagittal images, the superior limit of the sphenoid face is easily identified to ensure a safe distance from the skull base (Fig. 1). A nasal Lempert speculum is introduced. A Hardy retractor is placed through a sublabial incision when a narrow nasal aperture precludes adequate exposure. The endoscopes together with the localizing probe provide excellent visualization of the sphenoid sinus and related structures. Landmarks such as the intrasinus septae, region of the optic nerves, carotid arteries, and skull base are easily identified by the imaging system (Fig. 1). Without the clear visualization of the attachment of the intrasinus septum, the carotid canal could be unroofed (Fig. 1). Probe (cross-hairs) in the sphenoid sinus. Note the proximity of the probe to the skull base (arrow) in the upper right quadrant; the region of the optic nerves (hollow arrow) in upper left quadrant; and the carotid (diagonal arrow) and the intrasinus septum (hollow arrow) in the lower left quadrant. Thirty-degree endoscopic view in the sphenoid sinus. Note the left optic nerve heading to the optic chiasm. The asterisk represents the opening into the sella turcica created by chipping the thin bony fragments from the anterior wall of the sella turcica. We use endoscopic techniques for the extirpation of the pituitary mass. Our neurosurgeon (d.w.a.) uses two hands during the tumor removal and operates off the monitor, while an assistant holds the endoscope. The neurosurgeon utilizes endoscopic technique to eggshell the anterior wall of the sella turcica with a penfield elevator. The pituitary tumor is removed with a blunt curette and pituitary forceps. Thirty-degree and 70° endoscopes are used to visualize the most antero-superior quadrant of the sella (Fig. 2). The localization probe augments the surgeon's orientation within the sella, and if necessary, above the diaphragma sellae (Fig. 3). At the conclusion of the tumor removal, the angled endoscopes are used to inspect the “hidden” regions of the sella for residual tumor and cerebrospinal fluid leak. The sphenoid sinus is packed with absorbable hemostat (Surgicel Medical, Inc., Arlington, TX) and thrombin-soaked microfibrillar collagen (Avitene, Alcon, San Juan, PR). If a cerebrospinal fluid leak is suspected, autologous fat is harvested and placed in the sella and sphenoid sinus, after the mucosa is removed from the sphenoid sinus. None of our six cases required this packing technique. The intact sphenoid rostrum or posterior piece of vomer is replaced. The mucoperichondrial flap and the bony and cartilaginous septum are repositioned and secured to the maxillary crest. The left anterior hemitransfixion incision is closed with interrupted sutures. Two running mattress stitches are placed through and through the septum to prevent the development of a septal hematoma. Each patient receives a single bolus of dexamethasone, 10 mg, in the operating room. Antibiotic-coated nasal packs are placed bilaterally. All patients are admitted for 48 hours to monitor for diabetes insipidus. Intravenous antibiotics are administered while the nasal packs are in place. After 48 hours the nasal packs are removed and the patient is discharged on oral antibiotics for 8 additional days. Patients are seen in follow-up 2 weeks after surgery. There were no cerebrospinal fluid leaks or sinonasal complications in our series. Both the endoscopes and localizing systems serve as adjuncts to good surgical technique in the complete removal of pituitary tumors while enhancing safety. It has been demonstrated by several authors that endoscopes afford the surgeon greater visualization, light amplification, and magnification in pituitary surgery.5, 7, 8 We find that with practice, the endoscopes are very easy to use and far less cumbersome than the microscope. The localizing system replaces traditional cross-table radiographs or intra-operative fluoroscopy. Both of these modalities are unable to demonstrate the lateral dimensions of the operative field to help prevent injury to lateral vital structures. The real-time virtual motion “within the patient's scans” helps the surgeon to operate with three-dimensional visualization without having to stop and review the CT scan. The curved probe and angled endoscopes allow improved visualization around corners. Localization accuracy is a function of the initial data acquisition, proper positioning of the headset, probe registration, a metallic environment, and the surgeon's visual acuity. These factors are magnified as the probe is moved further from the sensor on the headset. In our VTI endoscopic surgery series the overall accuracy was ±0.7 mm in the cranial-caudal axis, and ±0.45 mm in all other axes.10 Occasionally during the case the surgeon will be warned if an excess of metal in the nose is affecting the electromagnetic sensor. This occurs when using a Hardy retractor made of 200 series stainless steel. We have found that when using 300 series stainless steel there is no metal artifact. When the angled endoscopes were used in conjunction with VTI, the neurosurgeons were able to safely remove additional tumor above the diaphragma sellae that might not have been visualized with the traditional microscope approach. Of the six cases, residual tumor was noted in two cases when the angled endoscopes and VTI were utilized. The VTI system would be particularly helpful in revision pituitary tumor cases where landmarks are distorted or absent. VTI's Instatrak system offers several advantages over other localization systems. Only one CT scan is necessary. There is no need for the placement of fiducial markers on the patient's face. The operating room equipment has been simplified to include just the probe and headset. The surgeon's mobility is maximized by a frame-less headset and armless probe. Three-dimensional localization is used in conjunction with angled endoscopes to navigate safely in the sella (cross-hairs) away from adjacent critical structures. This will be significantly improved with magnetic resonance imaging and soft tissue—formatted computed tomography scans. At present the VTI Instatrak system can only localize based on a preoperatively obtained CT scan. While the CT scans are good for bony anatomy, future modifications will allow the incorporation of MRI images into the system. In future cases we are considering utilizing a soft tissue formatted CT scan of the sinuses and parasellar region. The MRI images and soft tissue formatted CT will help to localize the extent of the pituitary tumor relative to adjacent fixed critical structures (i.e., optic nerves, carotid arteries, clivus, and skull base) prior to resecting the tumor. It should be cautioned that any alteration of he preoperative anatomy during surgery will not be demonstrated by the image localization technology. That is, once bony landmarks have been moved or soft tissues have been altered or removed, the remaining soft and hard tissues may shift, thus displaying a false localization. Intraoperative imaging would allow for real-time localization and would account for surgical alteration of the patient's anatomy.11 There is no need for the surgeon to leave the operative field to re-examine the CT scans because the endoscopic view and the three-dimensional image localization images are in front of the surgeon and they track with the surgeon's instruments in real time. Three-dimensional localization can enhance resident education by displaying the position of the probe in relation to the vital structures surrounding the sella. Modern imaging technologies cannot supplant the surgeon's basic anatomic knowledge but can allow surgeons to visualize in three dimensions so they can more safely perform endoscopic pituitary surgery. The authors wish to thank Robert T. Sataloff, MD, DMA, for his critical review of the manuscript.