Frameless image-guided stereotactic brain biopsy procedure: diagnostic yield, surgical morbidity, and comparison with the frame-based technique

Transcription

1 J Neurosurg 104: , 2006 Frameless image-guided stereotactic brain biopsy procedure: diagnostic yield, surgical morbidity, and comparison with the frame-based technique GRAEME F. WOODWORTH, B.S., MATTHEW J. MCGIRT, M.D., AMER SAMDANI, M.D., IRA GARONZIK, M.D., ALESSANDRO OLIVI, M.D., AND JON D. WEINGART, M.D. Department of Neurosurgery, The Johns Hopkins University School of Medicine, The Johns Hopkins Hospital, Baltimore, Maryland Object. The gold standard for stereotactic brain biopsy target localization has been frame-based stereotaxy. Recently, frameless stereotactic techniques have become increasingly utilized. Few authors have evaluated this procedure, analyzed preoperative predictors of diagnostic yield, or explored the differences in diagnostic yield and morbidity rate between the frameless and frame-based techniques. Methods. A consecutive series of 110 frameless and 160 frame-based image-guided stereotactic biopsy procedures was reviewed. Associated variables for both techniques were reviewed and compared. All stereotactic biopsy procedures were included in a risk factor analysis of nondiagnostic biopsy sampling. Frameless stereotaxy led to a diagnostic yield of 89%, with a total permanent morbidity rate of 6% and a mortality rate of 1%. Larger lesions were fivefold more likely to yield diagnostic tissues. Deep-seated lesions were 2.7-fold less likely to yield diagnostic tissues compared with cortical lesions. Frameless compared with frame-based stereotactic biopsy procedures showed no significant differences in diagnostic yield or transient or permanent morbidity. For cortical lesions, more than one needle trajectory was required more frequently to obtain diagnostic tissues with frame-based as opposed to frameless stereotaxy, although this factor was not associated with morbidity. Conclusions. With regard to diagnostic yield and complication rate, the frameless stereotactic biopsy procedure was found to be comparable to or better than the frame-based method. Smaller and deep-seated lesions together were risk factors for a nondiagnostic tissue yield. Frameless stereotaxy may represent a more efficient means of obtaining biopsy specimens of cortical lesions but is otherwise similar to the frame-based technique. KEY WORDS frameless stereotaxy image-guided surgery neuronavigation stereotactic biopsy S Abbreviation used in this paper: MR = magnetic resonance. TEREOTACTIC biopsy techniques have been widely utilized in the diagnosis of intracranial lesions for many decades. 4,14 These techniques provide a means of sampling tissues from small, isolated intracranial lesions for histological analysis. Preoperative diagnoses of intraaxial lesions based solely on clinical and radiological presentation may be incorrect in one third of patients. 4,16 Hence, histopathological analysis of tissue samples is critical in guiding therapeutic strategies and preventing unnecessary interventions. 11,14 Traditionally, framed-based methods have been the gold standard for the sampling of intracranial lesions. 2,3,15 Recently, frameless image-guided stereotaxy has been increasingly used. This technique requires the use of preoperative imaging to provide spatial reconstructions and neuronavigation in the operating room, while permitting precise targeting of the desired intracranial lesion without the use of the traditional stereotactic frame for three-dimensional positioning and targeting. Given the importance of sampling intracranial lesions, and the equivocal reports on the morbidity associated with stereotactic biopsy procedures, 2,9,11 continued evaluation of this new technique is critical. 1,7 Furthermore, analysis of preoperative predictors of the diagnostic yield or direct comparison of the differences in the diagnostic yield and morbidity rate between the frameless and frame-based methods has not been performed. In the present study we reviewed our 8-year experience with stereotactic biopsy procedures, including a detailed analysis of the frameless technique, an analysis of preoperative radiographic/neuroimaging predictors of nondiagnostic tissues, and a comparison of the differences in the diagnostic yield and complication rates between the frame-based and frameless techniques. Clinical Material and Methods All patients undergoing frame-based and frameless image-guided stereotactic brain biopsy procedures at The Johns Hopkins Hospital between 1995 and 2003 were identified. Patient demographics, clinical presentations, radiological/neuroimaging studies, surgical variables, pathological diagnoses obtained through stereotactic biopsy procedures, and clinical outcomes were retrospectively reviewed in all cases. Prior to the year 2000, all biopsy specimens were obtained via the frame-based technique. Following the adaptation of frameless stereotaxy at The Johns Hopkins Hospital in 2000, the majority of patients 233

2 G. F. Woodworth, et al. TABLE 1 Comparison of presenting characteristics in patients undergoing frameless as opposed to frame-based stereotactic brain biopsy procedures No. (%) Variable Frameless Frame-Based Total p Value no. of procedures mean age (yrs)* male sex 64 (58) 92 (58) 156 (58) 0.95 prebiopsy treatment resection 7 (6) 14 (9) 21 (8) 0.92 radiotherapy 6 (5) 14 (9) 20 (7) 0.31 chemotherapy 4 (4) 14 (9) 18 (7) 0.14 mean max lesion diameter (cm)* lesion appearance on 110 (100) 64 (40) 174 (64) 0.01 imaging (% MRI) enhancing 91 (83) 147 (92) 238 (88) 0.03 heterogeneous 97 (88) 121 (76) 218 (81) 0.01 deep-seated basal ganglia 8 (7) 14 (15) 22 (8) 0.66 brainstem 27 (25) 24 (15) 51 (19) 0.05 tissue type malignant glioma 54 (50) 74 (46) 128 (47) 0.64 low-grade glioma 14 (13) 15 (10) 29 (11) 0.38 lymphoma 12 (11) 22 (14) 34 (13) 0.46 abscess 6 (5) 5 (3) 11 (4) 0.10 metastatic tumor 3 (3) 8 (5) 11 (4) 0.86 necrosis 2 (2) 3 (2) 5 (2) 0.39 inflammatory mass 7 (6) 19 (11) 26 (10) 0.76 nondiagnostic results 12 (11) 14 (9) 26 (10) 0.56 * Values are presented as the means standard deviations (SDs). (85%) in this study underwent frameless image-guided stereotactic biopsy procedures. A head computerized tomography scan was obtained less than 48 hours after the biopsy procedure in 90% of the cases. End points reviewed for procedural comparison included diagnostic yield, number of biopsy specimens, number of biopsy trajectories, and biopsy procedure related neurological morbidity. Frameless Image-Guided Stereotactic Biopsy Procedure In all cases, a 1.5-tesla MR imaging unit (GE Medical Systems, Milwaukee, WI) was used to acquire preoperative images. Imaging studies were obtained the day of or the day before surgery. Twelve fiducial markers were appropriately positioned on the surface of each patient s head to enable operating room registration between the equipment and the patient. General anesthesia was induced, and the patient s head fixed in a three-point Mayfield clamp secured to the operating table. Intraoperative image guidance was achieved using a wand-based navigation system (Olivier device, ISG Viewing Wand System; ISG Technologies, Inc., Toronto, ON, Canada). Preoperative images were transferred to the ISG Hewlett Packard 700 workstation. The system calculated positioning accuracy by giving a root mean square error and used a matching algorithm to register successfully with the fiducial markers. With the registration process completed and its accuracy confirmed, the biopsy guide (SNN-Olivier FreeGuide; Philips Medical Systems, Best, The Netherlands) was then attached to the Mayfield clamp. This guide apparatus was secured to the calvaria with a metal pin to avoid further movement or play in the system. The guidance tube for the wand and biopsy needle allowed for targeting of the lesion in a three-dimensional orientation according to a straight trajectory, which was planned with the navigation system s software by defining an entry and target point. Areas corresponding to contrast enhancement on MR imaging were targeted for biopsy in both superficial and deep portions of the lesion. A trajectory was selected to pass through the largest dimension of the lesion when possible to allow for multiple biopsy sites via one needle pass. The target was selected and the distance to the site was measured and marked on the 1.7-mm Nashold biopsy needle (Radionics, Z Medical, Inc., Baltimore, MD). A 0.5-cm linear incision was made at the defined stereotactic site. A battery-operated drill was used to create a in opening into the skull, and the dura mater was opened using a sharp needle. The biopsy needle was then passed down to the area of the lesion at the previously determined distance. Approximately 8-mm-long and 1-mm-thick tissue specimens were obtained and sent to the pathology department for frozen sectioning. If the pathological reading was nondiagnostic, additional samples were taken from another enhancing region of the lesion. Because of the targeting capabilities of the frameless technique, the additional biopsy specimens were obtained without a second needle pass. Frame-Based Image-Guided Stereotactic Biopsy Procedure The Leksell (Leksell Stereotactic System; Elekta Instruments, Atlanta, GA) or Cosman-Roberts-Wells stereotactic frame was used in all cases. Contrast-enhanced T 1 - weighted MR images or contrast-enhanced computerized tomography scans were used for the identification of three biopsy targets and to establish stereotactic coordinates. Those regions corresponding to an area of contrast enhancement were targeted for biopsy sampling. General anesthesia was induced in all patients, and in a few cases, a local anesthetic was applied with minimal sedation at the time of surgery. An approximately 0.5-cm linear incision was made at the defined stereotactic site. A battery-operated drill was used to create a in opening into the skull. Approximately 8-mm-long and 1-mm-thick tissue specimens were obtained using a 1.7-mm Nashold biopsy needle. If the pathological reading was nondiagnostic, additional samples were taken from another enhancing region of the lesion. In general, additional biopsies required additional needle passes. Statistical Analysis Risk Factors for Nondiagnostic Stereotactic Biopsy Procedures. All stereotactic biopsy procedures were included in a risk factor analysis. Preoperative MR imaging lesion characteristics were assessed for nondiagnostic biopsy specimens through univariate and multivariate logistic regression analyses (StatView, version 5; SAS Institute, Inc., Cary, NC). Statistical significance was defined as a probability value less than or equal to All recorded variables related to the nondiagnostic tissue were included in the multivariate analysis. Comparison of Frameless and Frame-Based Stereotaxy. For intercohort comparisons, continuous data were compared using a two-way analysis of variance for parametric data and the Mann Whitney U-test for nonparametric data. 234

3 Frameless stereotactic brain biopsy sampling Percentages were compared using chi-square and Fisher exact tests where appropriate. Statistical significance was defined as a probability value less than or equal to 0.01 due to the large number of variables in the intercohort comparisons. Results Frameless Stereotactic Biopsy Procedure One hundred ten frameless stereotactic biopsy procedures were performed at The Johns Hopkins Hospital between 1995 and The mean patient age was years and the study cohort consisted of 64 male patients (58%). Prior to the biopsy procedure, seven patients (6%) underwent surgical lesion removal, six (5%) received radiotherapy, and four (4%) received chemotherapy. Preoperative imaging revealed a mean maximal lesion diameter of cm, 91 enhancing lesions (83%), 97 heterogeneous or nonuniformly enhancing lesions (88%), and 35 deepseated lesions (32%; Table 1). Tumor tissue type varied as follows: 54 malignant gliomas (50%), 14 low-grade gliomas (13%), 12 lymphomas (11%), six abscesses (5%), three metastatic tumors (3%), two nontumor necroses (2%), seven inflammatory masses (6%), and 12 nondiagnostic masses (11%). The overall diagnostic yield was 89%, and 78% of the initial diagnoses based on the initial frozen tissue sections agreed with those based on the final tissue section (Table 2). Biopsy-associated morbidity occurred in 17 patients (15%) with permanent morbidity in six (6%). Death occurred due to intracranial hemorrhage in one patient (1%). Risk Factors for Nondiagnostic Stereotactic Biopsy Procedure In the analysis of all biopsy procedures performed during the study period, larger lesions ( 2 cm in maximal diameter) were fivefold more likely to be diagnostic (Table 3). Biopsy specimens of deeply located lesions (basal ganglia, thalamus, or brainstem) were 2.7-fold less likely to be diagnostic. Patient age; number of biopsy samples; and enhancement, heterogeneity, and mass-associated edema on preoperative MR images were not associated with the diagnostic yield on univariate analysis. In a multivariate analysis of all variables associated with nondiagnostic tissues, a deep-seated lesion location or small lesion size were not independent predictors of a nondiagnostic biopsy sample. Comparison of Frameless and Frame-Based Stereotaxy Two hundred seventy stereotactic brain biopsy procedures were performed at our institution during the study period. One hundred sixty of these (59%) were frame-based procedures, whereas 110 (41%) were frameless procedures combined with intraoperative imaging. The mean age of all patients undergoing the biopsy procedures was years. One hundred fifty-six patients (58%) were male, 21 (8%) had undergone previous lesion resection, 20 (7%) had received previous radiotherapy, and 18 (6.7%) had received previous chemotherapy. The lesion location was frontal in 79 cases (29%), parietal in 44 (16%), and temporal in 39 (14%). The lesion involved the brainstem and/or thalamus in 51 cases (19%), the corpus collosum in 22 (8%), the basal TABLE 2 Comparison of clinical, neuroimaging, and operative variables between frame-based and frameless stereotactic biopsy procedures No. (%) Variable Frameless Frame-Based Total p Value no. of procedures procedure comparison diagnostic yield 98 (89) 146 (91) 244 (90) 0.78 diagnostic agree- 86 (78) 112 (70) 198 (73) 0.16 ment btwn initial & final tissue sections no. of specimens* no. of trajectories* no. w/ 1 trajectory 1 (1) 15 (9) 16 (6) complications total morbidity 17 (15) 20 (13) 37 (14) 0.54 transient deficit 11 (10) 15 (10) 26 (10) 0.42 permanent deficit 6 (6) 5 (3) 11 (4) 0.36 asymptomatic bi- 6 (5) 11 (7) 17 (6) 0.81 opsy site hemorrhage symptomatic biopsy 4 (4) 4 (3) 8 (3) 0.72 site hemorrhage death 1 (1) 2 (1) 3 (1) 0.79 * Values are presented as the means SDs. ganglia in 22 (8%), the cerebellum in four (1.5%), and the pineal gland in four (1.5%). Lesions accessed using frameless stereotaxy were more frequently heterogeneous (nonuniformly enhancing on imaging). Magnetic resonance imaging was used more frequently with the frameless than with the frame-based technique (100% compared with 40%, respectively; p 0.01). The lesions in patients who had undergone frameless biopsy procedures tended toward less frequent enhancement on MR imaging and more frequently involved brainstem structures, although these differences did not reach statistical significance according to our criteria (Table 1). Otherwise, lesion characteristics were similar between the two procedure groups. The diagnostic yield was similar between the frameless (89%) and frame-based procedures (91%) (Table 2). On average, four tissue samples were taken from three biopsy sites for both surgical techniques. For cortical lesions, a greater number of trajectories were used in the frame-based compared with those in the frameless procedures ( compared with , respectively; p = 0.003; Table 3). A second needle trajectory was needed to obtain diagnostic tissue in 13 frame-based cases (11%) compared with only one frameless case (1%; p 0.01). For deep-seated lesions, the number of required trajectories was similar between the frame-based and frameless groups ( compared with 1 0, respectively; p = 0.160). For such lesions, a second needle trajectory was rarely used in either biopsy procedure group (2% compared with 0%, p = 0.49). Neurological morbidity did not differ between framebased and frameless biopsy procedures, occurring in 37 cases (14%) overall (Table 2). Deficits were transient in 26 patients (10%) and permanent in 11 (4%). Biopsy site hemorrhage occurred in 12 frame-based (8%) and 13 frameless 235

4 G. F. Woodworth, et al. TABLE 3 Univariate association between preoperative MR imaging characteristics and nondiagnostic stereotactic biopsy samples Odds Ratio Nondiagnostic Diagnostic for Nondiag- Variable Sample Sample nostic Sample p Value no. of samples age (yrs)* no. of biopsy samples* no. of deep-seated lesions (%) 10 (53) 73 (29) max lesion diameter (cm)* lesion diameter 2 cm (%) no. of enhancing lesions (%) 16 (84) 227 (91) no. of heterogeneous lesions on MRI (%) 14 (74) 204 (81) samples w/ lesion-associated edema (%) 10 (53) 180 (72) * Values are presented as the means SDs. biopsy procedures (9%). Overall, biopsy site hemorrhage was symptomatic in only eight cases (3%). Three patients (1%) died within 14 days of the biopsy procedure: two of tumor-associated edema that was present prior to the biopsy procedure and one after suffering a large intracranial hemorrhage after the stereotactic biopsy procedure. Discussion In our review of 110 frameless stereotactic biopsy procedures performed during an 8-year period at The Johns Hopkins Hospital, the diagnostic yield was 89%, with a total permanent morbidity rate of 6% and a mortality rate of 1%. These results compare well with data in a recent literature review of stereotactic brain biopsy procedures performed by Hall, 9 who showed an overall diagnostic yield of 91%, a morbidity rate of 3.5%, and a mortality rate of 0.7%. In the analysis of risk factors for nondiagnostic biopsy, we found that a smaller lesion size and deep-seated location together were associated with nondiagnostic tissue sampling. In a direct comparison of the 110 frameless procedures and the 160 frame-based procedures, we found that the diagnostic yields and complication rates were similar. Collectively, these results confirm the notion that the stereotactic brain biopsy procedure is a safe and effective method sufficient to guide treatment in the majority of intracranial lesions. 6,11,13 Our results suggest that frame-based stereotaxy offers no benefit for lesion targeting or risk reduction compared with frameless techniques. In fact, to achieve a clear pathological diagnosis, additional needle passes were required in 11% of the frame-based cortical biopsy samplings compared with only 1% of the frameless procedures (p 0.01). A second trajectory was needed in these cases only when several passes via the first trajectory did not yield diagnostic or pathological tissue. The imaging feedback may have enabled the surgeon more confidently to take a greater number of samples prior to changing the trajectory, owing to the ability to see the lesion dimensions and locations while performing the biopsy procedure. Previous authors have reported that increased numbers of stereotactic needle biopsy passes are associated with increased biopsy-related morbidity. 5,9 Such was not the case in the present study, however, in which these additional needle passes were not associated with morbidity. The fewer trajectories required in the frameless biopsy sampling of cortical lesions leads one to infer a greater diagnostic yield per needle trajectory with this system. A comparison of the two cohorts showed that there were more heterogeneous (nonuniformly enhancing) lesions and that there tended to be fewer enhancing lesions in the frameless group, making this set of lesions more difficult to assess accurately by needle biopsy methods. The frameless technique was combined with MR imaging more frequently and tended to involve a greater percentage of brainstem biopsy procedures. Given these differences in imaging modalities and lesion characteristics and types, there are inherent limitations in this comparison and an increased risk of a Type II error. Even with these discrepancies, there was no difference in the biopsy procedure associated morbidity in patients with deep-seated lesions when comparing the two techniques. The higher-resolution imaging modality combined with the intraoperative neuronavigation of the frameless technique facilitated safe biopsy sampling of these deep, eloquent tissues. This result suggests that both techniques can successfully and safely be used to sample tissue from the basal ganglia, pons, or brainstem. Some believe that the more rigid frame used in the frame-based system conveys greater precision, particularly in targeting deep-seated lesions in critical areas. 5,10,12,16 Regardless of the frame type, the use of intraoperative MR imaging guided stereotaxy is as effective as the traditional frame-based technique when combined with preoperative lesion targeting. We also determined whether preoperative lesion characteristics on MR images could indicate lesions at an increased risk of yielding nondiagnostic tissue. We found that stereotactic biopsy samples of larger lesions were fivefold more likely to be diagnostic, whereas samples of deeper lesions within the basal ganglia, thalamus, or brainstem were 2.7-fold less likely to be diagnostic. Results of multivariate analysis did not indicate small-sized lesions or deep-seated locations to be independent predictors of nondiagnostic biopsy. Hence, a small lesion size and a deep-seated location should be considered together as risk factors for nondiagnostic biopsy procedures. Frameless stereotaxy allowed us to collect multiple biopsy samples of larger lesions along a single needle track over a greater distance across the lesion. Thus we gathered tissue samples across the spectrum of lesion tissue types, increasing the likelihood of obtaining 236

5 Frameless stereotactic brain biopsy sampling specimens representative of the lesion without multiple needle passes. Field, et al., 8 found that the diagnostic yield was lower in younger patients undergoing stereotactic brain biopsy procedures. This finding was attributed to the more diverse spectrum of tumor types and grades seen in the pediatric population. We did not observe an association between patient age and diagnostic yield in our series, which may have been related to the diversity of lesions across our patients or simply due to the fact that there were only 14 patients (5%) in the group of 270 who were younger than 25 years of age. Field, et al., also found lower diagnostic yields in cases in which greater numbers of biopsy samples were obtained. This increased number of samples was likely a result of the lesion itself and offered no preoperative predictive value. Although this study is inherently limited by being retrospective, we attempted to restrict the bias associated with surgeon experience or technique by including only those cases treated by the two surgeons who performed the majority of the stereotactic biopsy procedures at the institution. Pathologist interpretation bias was limited by the fact that the majority of neuropathological studies at our hospital are reviewed by a single senior neuropathologist. Finally, as previously mentioned, the intercohort differences in lesion characteristics, location, and imaging modality used may have limited this comparison and any differences seen between the two techniques. This study was conducted during a greater portion of the learning period for the frameless biopsy technique; this factor could not be controlled and is a possible confounding factor. Conclusions Results obtained using the frameless stereotactic biopsy procedure were found to be comparable to or better than those of published studies on stereotactic biopsy sampling with regard to diagnostic yield and complication rate. Smaller and deep-seated lesions together represented risk factors for nondiagnostic tissue yield. In comparing frameless and frame-based stereotactic biopsy procedures, diagnostic yield and overall morbidity were found to be similar. Frameless as opposed to frame-based stereotactic biopsy procedures may more efficiently allow a pathological diagnosis in cortical lesions and can successfully lead to tissue diagnosis in brainstem lesions without increased morbidity. Regardless of the frame type used, intraoperative MR imaging guided stereotaxy is as effective as the traditional frame-based technique with preoperative lesion targeting. References 1. Alberti O, Dorward NL, Kitchen ND, Thomas DG: Neuronavigation impact on operating time. Stereotact Funct Neurosurg 68:44 48, Apuzzo ML, Chandrasoma PT, Cohen D, Zee CS, Zelman V: Computed imaging stereotaxy: experience and perspective related to 500 procedures applied to brain masses. Neurosurgery 20: , Apuzzo ML, Chandrasoma PT, Zelman V, Giannotta SL, Weiss MH: Computed tomographic guidance stereotaxis in the management of lesions of the third ventricular region. Neurosurgery 15: , Boethius J, Collins VP, Edner G, Lewander R, Zajicek J: Stereotactic biopsies and computer tomography in gliomas. Acta Neurochir (Wien) 40: , Callovini GM, Sherkat S, Rinaldi A, Santucci N, Gazzeri G: Stereotactic-guided microsurgical removal of lesions without cortical appearance planned by three-dimensional CT reconstruction: limits and advantages of the frame-based technique. Minim Invasive Neurosurg 41: , Chandrasoma PT, Smith MM, Apuzzo ML: Stereotactic biopsy in the diagnosis of brain masses: comparison of results of biopsy and resected surgical specimen. Neurosurgery 24: , Dorward NL, Paleologos TS, Alberti O, Thomas DG: The advantages of frameless stereotactic biopsy over frame-based biopsy. Br J Neurosurg 16: , Field M, Witham TF, Flickinger JC, Kondziolka D, Lunsford LD: Comprehensive assessment of hemorrhage risks and outcomes after stereotactic brain biopsy. J Neurosurg 94: , Hall WA: The safety and efficacy of stereotactic biopsy for intracranial lesions. Cancer 82: , Heilbrun MP: Computed tomography-guided stereotactic systems. Clin Neurosurg 31: , Jackson RJ, Fuller GN, Abi-Said D, Lang FF, Gokaslan ZL, Shi WM, et al: Limitations of stereotactic biopsy in the initial management of gliomas. Neuro-oncol 3: , Kelly PJ, Earnest F IV, Kall BA, Goerss SJ, Scheithauer B: Surgical options for patients with deep-seated brain tumors: computer-assisted stereotactic biopsy. Mayo Clin Proc 60: , McGirt MJ, Bulsara KR, Cummings TJ, New KC, Little KM, Friedman HS, et al: Prognostic value of magnetic resonance imaging-guided stereotactic biopsy in the evaluation of recurrent malignant astrocytoma compared with a lesion due to radiation effect. J Neurosurg 98:14 20, Ostertag CB, Mennel HD, Kiessling M: Stereotactic biopsy of brain tumors. Surg Neurol 14: , Sawin PD, Hitchon PW, Follett KA, Torner JC: Computed imaging-assisted stereotactic brain biopsy: a risk analysis of 225 consecutive cases. Surg Neurol 49: , Wild AM, Xuereb JH, Marks PV, Gleave JR: Computerized tomographic stereotaxy in the management of 200 consecutive intracranial mass lesions. Analysis of indications, benefits and outcome. Br J Neurosurg 4: , 1990 Manuscript received January 5, Accepted in final form August 29, Address reprint requests to: Graeme F. Woodworth, M.D., Department of Neurosurgery, The Johns Hopkins Hospital, 600 North Wolfe Street, Harvey 8-161, Baltimore, Maryland