Introduction
The availability of powerful computer workstations the size of a personal computer, along with new technology for tracking the position of a probe in 3-dimensional space, spurred certain far-seeing neurosurgical investigators to develop turnkey systems for navigation in neurosurgery. Dr. Richard Bucholz, a neurosurgeon at St. Louis University, saw the surgical potential in infrared digitizer tools. Together with a young engineer named Kurt Smith, he developed the system known as the StealthStation®. As with other frameless stereotactic (FS) devices, while the digitizing technology may be similar, the image rendering, user interface, means of maintaining stereotac-tic accuracy, essential hardware, and surgical tools are unique to this particular system.
System Description and Use
The StealthStation® system was designed by Surgical Navigation Technologies (Louisville, CO), now a division of Medtronic, Inc. (Minneapolis, MN). It has been described by Kurt Smith [1] and was independently evaluated by Germano [2]. System components include: a Unix-based Silicon
Graphics workstation, an infrared digitizer with active, infrared light emitting diodes (IRLED) or with passive reflector spheres, and infrared cameras (Polaris, Inc.); a dynamic reference frame that maintains registration accuracy even with movement of the patient or cameras; and a monitor with the StealthStation proprietary user interface. Software packages are available for cranial navigation, otolaryngological use (nasal sinus surgery, in essence), spinal surgery (for instrument insertion), functional stereotactic surgery, and certain orthopedic applications for limb surgery. These options make the StealthStation more attractive to hospitals as the system is more likely to be used and not gather dust—an important consideration for a device whose retail price is above $300,000.
Use of the StealthStation begins with preoperative imaging. The scan used for registration may be obtained with adhesive fiducial markers applied to the scalp in a manner that encompasses the volume of interest. It is possible to attempt registration with anatomical landmarks alone but at the New Jersey Medical School we have not found this to be sufficiently reliable. (SNT is developing a tool called the Fazer™ that will acquire a map of the patient’s face using laser scanning, and provide registration; this may avoid the need for fiducials in the future.) The choice of MRI or CT will depend on the needs of the surgeon—when bony landmarks are the main concern then CT may be sufficient or ideal, but for most intracranial applications MRI will be the modality of choice. A scan may be obtained well in advance of surgery without fiducials and fused in the operating room (OR) to a different scan acquired with fiducials the day before operation. In any event, scans for the StealthStation must be obtained with zero gantry tilt, a slice thickness of 3 mm or less, no slice overlap or skip, and a field of view large enough to encompass the volume of interest.
Scan data is transferred to the OR workstation by an Ethernet connection or, if needed, by such removable media as optical disks ("sneakernet"). Image fusion of CT and MRI scans is easily done using the automated fusion program; functional data may be added in the same fashion [3]. Landmarks for registration (scalp markers or anatomical points) are numbered on a 3-dimensional image reconstruction. After the patient’s head is secured in the head clamp, the dynamic reference frame (DRF) is attached to the clamp (see Fig. 1). Registration is then done using at least four fiducials. The program then provides an estimate of the accuracy of the mathematical match between the scan and physical space. This number is not a true estimate of registration accuracy, which must be verified using anatomical points before and during surgery. At this point, navigation may be used to plan an incision and approach for craniotomy or stereotactic biopsy.
The DRF is removed and the patient prepped and draped. A sterile DRF is placed (or the previous one sterilized and replaced), and maintenance of registration accuracy is confirmed.
Figure 1 Patient positioned for transsphenoidal surgery with the StealthStation. Note placement of the dynamic reference frame, cameras, and monitor.
During surgery, navigation is performed as desired. Draping the keyboard with a clear plastic sheet allows the surgeon to operate the computer if necessary. The operating microscope may be registered as a navigational tool, but often it is more convenient to use a wand with IRLEDs. Hardware and software tools for needle biopsy and functional targeting are available. Details are beyond the scope of this topic but are based on the concepts and techniques of registration de scribed above. Figures 2 and 3 show images from the StealthStation in use for transsphenoidal surgery. Images in Figure 2 were created with the Cranial program™, while Figure 3 shows the use of the Landmarx ENT™ package, which demonstrates the versatility of the StealthStation.
DISCUSSION
At the New Jersey Medical School we have used the StealthStation for over six years. At first, use of the StealthStation seemed like a cumbersome addition to the OR and added at least two hours to any procedure. With further experience the system became a completely routine part of surgery and its use is transparent to OR and radiology staff. All of the neurosurgeons in our group are comfortable with its use. Concerns have been raised that residents will come to rely on this technology to the abandonment of fundamental surgical principles; the obvious answer to this issue is that such fears have not prevented the incorporation of other advances such as CT and MRI scans into neurosurgical education.
The versatility of the StealthStation is one of its most appealing features. In this it is not unique but not all commercially available systems share this feature. The choice of using active IRLED probes or the passive probes with reflector spheres gives the surgeon yet another degree of freedom. Cost is comparable to other navigational devices.
Limitations of the StealthStation are to some extent generic for IRLED-based surgical navigation (SN) systems. A line of sight between the reference arc and the probes and the IR cameras is required. Downloading data may not go as smoothly as hoped in all cases. Shifting of the patient’s scalp by positioning changes, or scalp pressure from the probe, may lead to registration inaccuracies. The biopsy arm is bulky and unwieldy, although a new device—the Vertek biopsy guide— will be released by the vendor and many of the ergonomic problems in the old arm should be solved. And of course, as with all SN systems that rely on preoperative datasets, brain shift will often render the intracranial registration inaccurate as surgery proceeds.
In May 2000 we began our experience with intraoperative imaging when we were the first North American site to install a PoleStar N-10 in-traoperative MR (iMRI) MRI unit. This system, with its integrated navigational tool, provides the advantages of navigation updated by intraoperative images that eliminate concerns related to brain and lesion shift [4]. The promise of combining the best features of the StealthStation with the capabilities of the PoleStar N-10 is a new and exciting development, as surely SN in the near future will require the incorporation of new images. However, the utility of a stand-alone surgical navigation device without iMRI or intraoperative CT will remain for patients undergoing a wide variety of surgery.
Figure 2 Screen images from transsphenoidal surgery for pituitary macroadenoma. (A) CT navigation shows probe at border of sella (OR view seen in lower right). (B) Probe points to basilar artery, just in back of eroded sella turcica.
Figure 3 Endoscopic transsphenoidal navigation for microprolactinoma using the noninvasive Landmarx frame. (A) Setup and surgery. (B) Monitor view with probe in tumor bed on reformatted MRI.
CONCLUSION
Surgical navigation with the StealthStation has become a routine part of many operating rooms. Little time is added to the procedure and, as with other advances in neurosurgical technology, a great deal of guesswork is eliminated in intracranial neurosurgery. The versatility of the StealthStation is a great advantage and in a relatively short time it can become a workhorse tool for any neurosurgeon. Even with the eventual incorporation of intra-operative imaging as an OR routine, the StealthStation will still have its place in the management of patients in whom intraoperative imaging is superfluous and as a system for updated navigation with newly acquired images.
