Donald W. Kormos

Preliminary experience with a frameless, armless stereotactic localization system in brain tumor surgery is presented. The localizing wand emits ultrasonic pulses that are detected by a table-mounted array of microphones--with triangulation of the emitter positions. The wand tip and trajectory are determined by proprietary computer software. Real-time display of this information is presented in multiple, two-dimensional or three-dimensional displays. Forty-eight patients underwent 52 craniotomies for brain tumors. The wand was used to assist in placing a minimal craniotomy in 48 cases, to determine the tumor/brain interface in 27 cases, to localize subcortical tumors in 14 cases, and to correlate the physiological mapping with the surface anatomy in 5 cases. In 12 instances, the wand was used in conjunction with frame stereotaxy and found to be comparable or superior. Triplanar (coronal, sagittal, transverse) two-dimensional images provided sufficient information for the detection of tumor boundaries but proved difficult to use to access a subcortical lesion; two-dimensional or three-dimensional images along the localization axis were more helpful. Frameless stereotaxy with this sonic wand system proved to be a useful adjunct to open-tumor biopsy or resection.

Interactive frameless stereotaxy has been successfully applied to intracranial surgery. It has contributed to the improved localization of deep-seated brain lesions and has demonstrated a potential for reducing both operative time and morbidity. However, it has not been as effectively applied to spinal surgery. The authors describe the application of frameless stereotactic techniques to spinal surgery, specifically pedicle screw fixation of the lumbosacral spine. Preoperative axial computerized tomography (CT) images of the appropriate spinal segments are obtained and loaded onto a high-speed graphics supercomputer workstation. Intraoperatively, these images can be linked to the appropriate spinal anatomy by a sonic localization digitizer device that is interfaced with the computer workstation. This permits the surgeon to place a pointing device (sonic wand) on any exposed spinal bone landmark in the operative field and obtain multiplanar reconstructed CT images projected in near-real time on the workstation screen. The images can be manipulated to assist the surgeon in determining the proper entry point for a pedicle screw as well as defining the appropriate trajectory in the axial and sagittal planes. It can also define the correct screw length and diameter for each pedicle to be instrumented. The authors applied this device to the insertion of 150 screws into the lumbosacral spines of 30 patients. One hundred forty-nine screws were assessed to be satisfactorily placed by postoperative CT and plain film radiography. In this report the authors discuss their use of this device in the clinical setting and review their preliminary results of frameless stereotaxy applied to spinal surgery. On the basis of their findings, the authors conclude that frameless stereotactic technology can be successfully applied to spinal surgery.

A technique of "frameless" stereotaxy that allows real-time intraoperative neurosurgical localization is described. The system is composed of four components: a hand-held probe containing two ultrasonic emitters, a microphone array that is rigidly affixed to the operating table in proximity to the surgical field, hardware to control and detect timing of signal production and reception, and a color graphics computer workstation with software to calculate and present the location of the probe tip on reconstructed neuroimaging studies. Unlike previously reported mechanical or sonic navigational devices, this system is adaptable to a wide array of neurosurgical instruments, allows free movement of the operating table and conventional patient draping, and has accuracy in the hostile operating room environment that rivals that of frame stereotaxy. In the operating room environment, using four pulse pairs with the wand positioned optimally, reproducibility of a point in space is +/- 0.6 mm. The wand has a broad range of orientations that maintain error at or below 1.0 mm. The mean error when measuring distances within a 1000-cu cm cube is 1.1 +/- 1.0 mm (1.0% +/- 0.7%). The ability to localize a fourth point (a target) in space is typically within 1.5 mm (using computerized tomography scans with a 1-mm slice thickness) but is dependent on several variables. This technology provides a powerful yet flexible tool in the neurosurgical operating room.

OBJECTIVE: We describe a shared-resource intraoperative magnetic resonance imaging (MRI) design that allocates time for both surgical procedures and routine diagnostic imaging. We investigated the safety and efficacy of this design as applied to the detection of residual glioma immediately after an optimal image-guided frameless stereotactic resection (IGFSR). METHODS: Based on the twin operating rooms (ORs) concept, we installed a commercially available Hitachi AIRIS II, 0.3-tesla, vertical field, open MRI unit in its own specially designed OR (designated the magnetic resonance OR) immediately adjacent to a conventional neurosurgical OR. Between May 1998 and October 1999, this facility was used for both routine diagnostic imaging (969 diagnostic scans) and surgical procedures (50 craniotomies for tumor resection, 27 transsphenoidal explorations, and 5 biopsies). Our study group, from which prospective data were collected, consisted of 40 of these patients who had glioma (World Health Organization Grades II-IV). These 40 patients first underwent optimal IGFSRs in the adjacent conventional OR, where resection continued until the surgeon believed that all of the accessible tumor had been removed. Patients were then transferred to the magnetic resonance OR to check the completeness of the resection. If accessible residual tumor was observed, then a biopsy and an additional resection were performed. To validate intraoperative MRI findings, early postoperative MRI using a 1.5-tesla magnet was performed. RESULTS: Intraoperative images that were suitable for interpretation were obtained for all 40 patients after optimal IGFSRs. In 19 patients (47%), intraoperative MRI studies confirmed that adequate resection had been achieved after IGFSR alone. Intraoperative MRI studies showed accessible residual tumors in the remaining 21 patients (53%), all of whom underwent additional resections. Early postoperative MRI studies were obtained in 39 patients, confirming that the desired final extent of resection had been achieved in all of these patients. One patient developed a superficial wound infection, and no hazardous equipment or instrumentation problems occurred. CONCLUSION: Use of an intraoperative MRI facility that permits both diagnostic imaging and surgical procedures is safe and may represent a more cost-effective approach than dedicated intraoperative units for some hospital centers. Although we clearly demonstrate an improvement in volumetric glioma resection as compared with IGFSR alone, further study is required to determine the impact of this approach on patient survival.

OBJECTIVE: Well-established surgical goals for pituitary macroadenomas include gross total resection for noninvasive tumors and debulking with optic chiasm decompression for invasive tumors. In this report, we examine the safety, reliability, and outcome of intraoperative magnetic resonance imaging (iMRI) used to assess the extent of resection, and thus the achievement of preoperative surgical goals, during transsphenoidal microneurosurgery. METHODS: Our magnetic resonance operating room contains a Hitachi AIRIS II 0.3-T, vertical-field open magnet (Hitachi Medical Systems America, Inc., Twinsburg, OH). A motorized scanner tabletop moves the patient between the imaging and operative positions. For transsphenoidal surgery, the patient is positioned directly on the scanner tabletop so that the surgical field is located between 1.2 and 1.6 m from the magnet isocenter. At this location, the magnetic field strength is low (<20 G), thus permitting the use of many conventional surgical instruments. Thirty consecutive patients with pituitary macroadenomas underwent tumor resection in our magnetic resonance operating room by use of a standard transsphenoidal approach. After initial resection, the patient was advanced into the scanner for imaging. If residual tumor was demonstrated and deemed surgically accessible, the patient underwent immediate re-exploration. RESULTS: iMRI was performed successfully in all 30 patients. In one patient, iMRI was used to clarify the significance of hemorrhage from the sellar region and resulted in immediate conversion of the procedure to a craniotomy. In the remaining 29 patients, initial iMRI demonstrated that the endpoint for extent of resection had been achieved in only 10 patients (34%) after an initial resection attempt, whereas 19 patients (66%) still had unacceptable residual tumor. All 19 of these latter patients underwent re-exploration. Ultimately, re-exploration resulted in the achievement of the planned endpoint for extent of resection in all of the 29 completed transsphenoidal explorations. Operative time was extended in all cases by at least 20 minutes. CONCLUSION: iMRI can be used to safely, reliably, and objectively assess the extent of resection of pituitary macroadenomas during the transsphenoidal approach. The surgeon is frequently surprised by the extent of residual tumor after an initial resection attempt and finds the intraoperative images useful for guiding further resection.