Computer-generated microsurgical anatomy of the basilar artery bifurcation

Transcription

1 J Neurosurg 91: , 1999 Computer-generated microsurgical anatomy of the basilar artery bifurcation Technical note TORU KOYAMA, M.D., HIROSHI OKUDERA, M.D., HIROHIKO GIBO, M.D., AND SHIGEAKI KOBAYASHI, M.D. Department of Neurosurgery, Shinshu University School of Medicine, Matsumoto, Japan; and Showa- Inan General Hospital, Neurosurgical Service, Komagane, Japan R The authors goal was to develop a computer graphics model to represent the microsurgical anatomy of the basilar artery (BA) bifurcation and surrounding structures to simulate surgery of a BA bifurcation aneurysm performed via the transsylvian approach. The source of the input data was a variety of publications that showed detailed anatomy of the area. A computer graphics model of the area near the BA bifurcation including relevant structures, such as perforating branches or cranial nerves, was depicted in detail. A BA bifurcation aneurysm was added to the computer graphics model and it was rotated to simulate the transsylvian approach. After the internal carotid artery was displaced using a virtual retractor, the aneurysm was exposed, thus providing an understanding of the three-dimensional surgical orientation of the area. Designing a standard anatomical model on the basis of data culled from a variety of publications and adding morphological changes by using a virtual retractor to displace structures that obstruct the view along a critical path at the base of the brain are useful strategies of computer manipulation for surgical simulation in open microneurosurgery. This methodological tool would be useful in teaching surgical microanatomy and in introducing a new navigational system for virtual reality. Both concept and technical details are discussed. KEY WORDS anatomical study three-dimensional reconstruction basilar artery perforating artery virtual reality ECENT advances in three-dimensional computerized tomography (CT) and magnetic resonance (MR) imaging have markedly improved their quality, 7,8, 19,29 and high-resolution T 2 -reversed MR imaging performed using a high magnetic field system or MR cisternography can reveal fine microstructures while detecting various physiological and pathological conditions. 4,20 However, images obtained by those systems do not simulate microstructures that are located in a deep area with respect to superficial structures. Although figures of microstructures obtained during cadaver dissection are still the most important reference tools to show anatomical relationships and to facilitate surgical orientation, it is difficult to integrate such figures freely into simulations of surgical fields. In addition, brain atlas databases do not represent the dynamic morphology that is encountered in open microneurosurgery. 21 Making a computer graphics representation with a resolution as high as that of figures obtained during cadaver dissection and adding morphological changes to simulate surgical manipulation are necessary to show the surgical anatomical relationship of microstructures. In this report, we describe a standard computer graphics model of the basilar artery (BA) bifurcation and relevant structures. Using this model, simulation of the transsylvian approach to a BA bifurcation aneurysm is possible because the internal carotid artery (ICA) is displaced using a virtual retractor. 18 The concept and technical details are described and future possibilities in surgical simulation and virtual reality are discussed. Materials and Methods System Design We created an application program by using a spline. The operating system we used was Windows 98 and the programming language was Visual C++ (version 4.0; Microsoft Corp., Redmond, WA). We used a standard 400-MHz Pentium II personal computer (Gateway model GP6-400; Gateway 2000 Japan Inc., Yokohama, Japan). A high-resolution computer monitor was required for acquisition of full-color images. Perforating Branches in the Region Near the BA Bifurcation The numbers of perforating branches found in the region near the BA bifurcation were collected from previous studies 5,6,25 27,31 (Table 1). We reviewed studies of the perforating branches of the P 1 and P 2 segments of the posterior cerebral artery (PCA), the distal end of the BA, the superior cerebellar arteries (SCAs), and the posterior communicating artery (PCoA). The average number of perforating branches arising from the P 1 segment ranged from 3.6 to 4.3, that from the BA ranged from 3.6 to 8, that from the SCA ranged from 0.6 to 4, and that from the J. Neurosurg. / Volume 91 / July,

2 T. Koyama, et al. TABLE 1 Number of perforating branches arising in the region near the BA bifurcation* Study/No. of Perforating Branches Saeki & Zeal & Gibo, et al. Sakata, et al., Segment Rhoton, 1977 Rhoton, , Comment P 2 segment thalamogeniculate artery medial posterior choroidal artery lat posterior choroidal artery 2 2 P 1 segment total in 1/3 of brains paramedian branch to the anterior half of PPS (to PThA) SPMA 1 IPMA 0.6 circumflex branch long circumflex branch 0.8 short circumflex branch 0.6 medial posterior choroidal artery BA total 8 (distal 10 mm) 3.6 (distal 7 mm) paramedian branch to PPS (Inoue, et al., 1992) IPMA 1.3 (35%) circumflex branch 4 2 in each side short circumflex branch 1.8 (50%) SCA total 4 (proximal 10 mm) 0.6 (proximal 7 mm) IPMA 0.2 (25%) short circumflex branch 0.3 (50%) PCoA * PPS = posterior perforated substance; PThA = posterior thalamoperforating artery. TABLE 2 Number and sites of termination of perforating branches used to design a computer graphics model in the region near the BA bifurcation No. of Per- Segment forating Branches Comment P 2 segment thalamogeniculate artery 2 arising from cortical branches of posterior half of P 2 medial posterior choroidal artery 1 arising from distal P 1 & anterior half of P 2 entering roof of 3rd ventricle lat posterior choroidal artery 2 arising from posterior half of P 2, terminating in choroidal plexus P 1 segment total 4 paramedian branch to anterior half of PPS 2 including PThA terminating in interpeduncular fossa SPMA 1 terminating in posterior half of PPS IPMA 0 circumflex branch long circumflex branch 1 encircling midbrain, providing branches as far posteriorly as colliculi short circumflex branch 0 medial posterior choroidal artery 0 BA total 8 paramedian branch IPMA 2 terminating in pontomesencephalic junction to PPS 2 circumflex branch short circumflex branch 4* encircling midbrain, providing branches to cerebral peduncle SCA total 2 IPMA 1 terminating in pontomesencephalic junction short circumflex branch 1 encircling midbrain, providing branches to cerebral peduncle PCoA 8 including infundibular and premammillary arteries * Two perforating branches in each side. 146 J. Neurosurg. / Volume 91 / July, 1999

3 Computer-generated anatomy of the BA bifurcation PCoA ranged from 7 to 7.8. The origin of the perforating branches was different in each study and the reported number and sites of termination also varied. The long circumflex branch arose predominantly from the P 1 segment (numbering 0.8 according to Sakata, et al. 26 ) and the short circumflex branch arose predominantly from the BA (numbering 4 according to Saeki and Rhoton 25 ). The superior paramedian mesencephalic artery (SPMA) arose predominantly from the P 1 segment (1 in the article by Sakata, et al.), and the inferior paramedian mesencephalic artery (IPMA) arose from the P 1 segment, the BA, and the SCAs (numbering 0.6, 1.3 and 0.2, respectively, according to Sakata, et al. 26 ). Computer Graphics Model in the Region Near the BA Bifurcation To create a standard computer graphics model of the BA bifurcation and surrounding structures, we used numbers and sites of termination suitable for generalization (Table 2). The site of termination of the perforating branches and its variations were confirmed by several references. 9,23 We designed a computer graphics model, in which the number of perforating branches arising from the P 1 segment, BA, SCA, and the PCA was 4, 8, 2, and 8, respectively. We designed 1 long circumflex branch arising from the P 1 segment, 4 short circumflex branches arising from the BA, 1 SPMA arising from the P 1 segment, 2 IPMAs arising from the BA, and 1 IPMA arising from the SCA. Using a variety of publications in which detailed anatomy of the area was shown, we integrated the anatomical relationships of microstructures with respect to the number and sites of termination of perforating branches that would be suitable for generalization. 5,6,9,10,23,25 27,31 We made axial and lateral view images that showed the contours of each microstructure of the area. To show the contours of these structures, the x, y, and z coordinates of each reference point were configured After we entered the input points, interpolated points were calculated using the application program and multiple perforating branches and surrounding structures were represented by full-color gradation of shading (Fig. 1). 14,18 Results We created a computer graphics model of the perforating branches and arterial trunks of the BA bifurcation and relevant structures to facilitate understanding of the anatomy of the area (Fig. 2). We selectively represented perforating branches arising from the P 1 segment of the PCA, BA, and SCA because surrounding structures such as the brainstem or third ventricular floor interfered with their visualization (Fig. 3). We added a BA bifurcation aneurysm to the computer graphics model and rotated and magnified the model to simulate surgery of a BA bifurcation aneurysm, performed via the transsylvian approach on the right side (Fig. 4). The right ICA hid the aneurysm and surrounding perforating branches and, thus, was displaced using a virtual retractor to simulate the surgical manipulation in open microsurgery (Fig. 5). 18 Using this system, it took approximately 90 seconds to obtain a computer graphics model of this area. J. Neurosurg. / Volume 91 / July, 1999 FIG. 1. Schematic representation showing the principle of fullcolor gradation of shading of microstructures. After the surface contours of a microstructure are configured, the inside of each polygon is dotted from the minimum X and Y coordinates (Xmin and Ymin, respectively) to maximum X and Y coordinates (Xmax and Ymax, respectively) with colors that correspond to the Z coordinate. When the number of interpolated points is increased, each polygon becomes small, corresponding to one dot, and full-color gradation of shading is achieved. Discussion Computerized Tomography and MR Images for Surgical Simulation Although recent advances in techniques have markedly improved the quality of CT and MR imaging, 3,7,8,19,29 and high-resolution T 2 -reversed MR imaging or MR cisternography can reveal fine microstructures, 4,20 images obtained using these techniques cannot simulate the surgical relationships of microstructures during open microsurgery. There are several difficulties in achieving surgical simulation. First, there is a resolution problem in three-dimensional CT and MR angiograms and representation of complex perforating branches is inadequate and at times impossible. Instead of using a voxel-rendering method, when contours of microstructures are configured and each reference point is connected by smooth curves passing through it, the natural contours of microstructures can be represented. Second, although a special frame or surface marker to integrate CT and MR images is used to represent microstructures such as perforating branches and cranial nerves, errors related to registration are inevitable for surgical simulation of microstructures. Third, during microsurgery, the original images would have to be appropriately altered to account for brain retraction, cerebrospinal fluid drainage, and other intraoperative events that change the orientation of microstructures to each other. 22 Although there are some recent reports concerning intraoperative brain shift, 2,24 CT or MR imaging reconstructions cannot simulate morphological changes of microstructures that are exposed during surgery. Previous surgical records concerning surgical approach and accessibility of clipping in BA bifurcation aneurysms are still the most reliable tools to determine the displacement and morphological changes in microsurgery of this area. 12,28,30 147

4 T. Koyama, et al. FIG. 2. Computer-generated representation of perforating branches and surrounding structures in the region near the BA bifurcation, inferior view. Each microstructure is represented by full-color gradation of shading. The anterior, lateral, and posterior perforated substances are encircled by green lines. 1 = ICA; 2 = PCoA; 3 = PCA; 4 = SCA; 5 = BA; 6 = oculomotor nerve; 7 = optic chiasm; 8 = optic tract; 9 = stalk; 10 = mammillary body; 11 = anterior perforated substance; 12 = lateral perforated substance; 13 = posterior perforated substance; 14 = cerebral peduncle. Computer Graphics Model Representing Standard Anatomy Although the number and termination sites of perforating branches in the region near the BA bifurcation vary in each report and an idealized presentation of perforating vessels may not be a substitute for knowing precise details, a representation of the average number and common sites of termination of perforating branches is meaningful. Microsurgical anatomy in the area is complicated. It would be very interesting to develop a computer graphics model, on the basis of individual angiographic or other imaging studies, and to add details to understand the relative regional relationship for a specific case. This would establish a revolution in surgical navigation. Our system is not used for stereotactic surgery and the coordinates used to indicate the contours of microstructures are not precise. The computer graphics model is not made from anatomical data obtained from an individual case and, thus, it may not represent reality. However, adjustment of represented structures is possible and a representation of standard anatomy and the anatomical relationships of multiple microstructures facilitates understanding of the surgical anatomy of the BA bifurcation and relevant structures. Simulation of Morphological Changes Caused by Surgical Manipulation To design a computer graphics model to simulate morphological changes caused by surgical manipulation, a standard computer graphics model would have to allow for appropriate alterations. We reported the concept of a virtual retractor that can be used to simulate morphological changes during microsurgery. 18 A true representation of what happens when using retractor pressure would be more complex; however, this is a good beginning to simulate displacement of structures obstructing the view along a critical path at the base of the brain. Introduction 148 J. Neurosurg. / Volume 91 / July, 1999

5 Computer-generated anatomy of the BA bifurcation FIG. 3. Computer-generated representations of perforating branches arising from the P 1 segment of the PCA, BA, and SCA. Upper: Superior view showing eight perforating branches arising from the bilateral P 1 segments (asterisks) colored light blue (branches to the anterior half of posterior perforated substance), yellowish green (SPMA), or pink (long circumflex branch). Center: Posterosuperior view, in which the arrows show the four paramedian branches arising from the BA and two IPMAs arising from each SCA. Lower: Posterior view showing four short circumflex branches arising from the BA (asterisks) and two short circumflex branches arising from each SCA (arrows). Numbered structures are defined in the legend to Fig. 2. J. Neurosurg. / Volume 91 / July,

6 T. Koyama, et al. FIG. 4. Computer-generated representations of perforating branches and surrounding structures in the region near the BA bifurcation. Upper: Superior view showing addition of a BA bifurcation aneurysm (asterisk); the right optic apparatus, brainstem, and third ventricular floor are not represented. Center: Computer graphics model that has been rotated and magnified to simulate the transsylvian approach; the left optic apparatus is not represented. Lower: Computer graphics model used to simulate the transsylvian approach; the left PCoA and its perforating branches are not represented. Rotation angles around the x axis are 0 (upper), 60 (center), and 80 (lower) and those of the Z axis are 0 (upper), 33 (center), and 45 (lower). 15 = posterior clinoid process; 16 = tent; 17 = right anterior choroidal artery. Other numbered structures are defined in the legend to Fig J. Neurosurg. / Volume 91 / July, 1999

7 Computer-generated anatomy of the BA bifurcation FIG. 5. Computer-generated representations used to simulate the transsylvian approach to a BA bifurcation aneurysm. Upper: A retractor (R) is placed on the right ICA. Center and Lower: The right ICA is displaced using a virtual retractor and microstructures such as perforating branches are exposed. The large asterisk shows the BA bifurcation aneurysm and the small asterisks show the perforating branches of the right PCoA. Arrows show the perforating branches from the P 1 segment and BA. Numbered structures are defined in the legend to Fig. 2. J. Neurosurg. / Volume 91 / July,

8 T. Koyama, et al. of various functions may be necessary to simulate complex morphological changes during microsurgery. Future Possibilities in Surgical Navigation for Open Microneurosurgery We are moving into an area of practical use of virtual reality in the operating room and computer-assisted interactive planning will be used for neurosurgical procedures. 1,11 Creating a standard model of microstructures, preparing multiple computer graphics models to represent individual differences, and adding morphological changes to represent surgical manipulation would be a good strategy to achieve surgical navigation. Our method would be a valuable contribution to the computer manipulation of anatomical data and its morphological changes for microneurosurgery. At the present time, this system can only be utilized as an educational tool for both training and practice. However, after multiple computer-generated models representing various anatomical variations and surgical manipulations are prepared, multiple image based data of the brain can be used not only in surgical simulations, but also in virtual reality for open microneurosurgery. References 1. Apuzzo MLJ: The Richard C. Schneider lecture. New dimensions of neurosurgery in the realm of high technology: possibilities, practicalities, realities. Neurosurgery 38: , Dorward NL, Alberti O, Velani B, et al: Postimaging brain distortion: magnitude, correlates, and impact on neuronavigation. J Neurosurg 88: , Filler AG, Kliot M, Howe FA, et al: Application of magnetic resonance neurography in the evaluation of patients with peripheral nerve pathology. J Neurosurg 85: , Fujii Y, Nakayama N, Nakada T, et al: High-resolution T 2 -reversed magnetic resonance imaging on a high magnetic field system. Technical note. J Neurosurg 89: , Gibo H, Hokama M, Wakui K, et al: [Microsurgical anatomy of the basilar artery and the proximal posterior cerebral artery (P- 1)], in Yamaura A (ed) [Surgical Anatomy for Microneurosurgery IV.] Tokyo: SciMed Publications, 1991, pp (Jpn) 6. Gibo H, Lenkey C, Rhoton AL Jr: Microsurgical anatomy of the supraclinoid portion of the internal carotid artery. J Neurosurg 55: , Harbaugh RE, Schlusselberg DS, Jeffery R, et al: Three-dimensional computed tomographic angiography in the preoperative evaluation of cerebrovascular lesions. Neurosurgery 36: , Imakita S, Onishi Y, Hashimoto T, et al: Subtraction CT angiography with controlled-orbit helical scanning for detection of intracranial aneurysms. AJNR 19: , Inoue T, Matsuno H, Nagata S, et al: [Microsurgical anatomy of the posterior perforated substance and posterior perforating arteries], in Gibo H (ed): [Surgical Anatomy for Microneurosurgery V.] Tokyo: SciMed Publications, 1992, pp (Jpn) 10. Inoue T, Rhoton AL Jr, Theele D, et al: Surgical approaches to the cavernous sinus: a microsurgical study. Neurosurgery 26: , Kikinis R, Gleason PL, Moriarty TM, et al: Computer-assisted interactive three-dimensional planning for neurosurgical procedures. Neurosurgery 38: , Kobayashi S, Sugita K, Nakagawa F: An approach to a basilar aneurysm above the bifurcation of the internal carotid artery. Case report. J Neurosurg 59: , Koyama T, Gibo H, Kyoshima K, et al: Computer-generated microsurgical anatomy of the paraclinoid area. Skull Base Surg 8:71 76, Koyama T, Gibo H, Okudera H: Computer-generated microsurgical anatomy of the supraclinoid portion of the internal carotid artery. J Clin Neurosci (In press, 1999) 15. Koyama T, Gibo H, Okudera H, et al: Clinical implications of computer assisted surgical design in a carotid cave aneurysm. J Clin Neurosci 3: , Koyama T, Okudera H, Kobayashi S: Computer-assisted geometric design of cerebral aneurysms for surgical simulation. Neurosurgery 36: , Koyama T, Okudera H, Kobayashi S: Computer-assisted surgical design of a basilar aneurysm in open microneurosurgery. J Comput Aided Surg 1:78 82, Koyama T, Okudera H, Kobayashi S: Computer-generated surgical simulation of morphological changes in microstructures: concepts of virtual retractor. Technical note. J Neurosurg 90: , Maeder PP, Meuli RA, de Tribolet N: Three-dimensional volume rendering for magnetic resonance angiography in the screening and preoperative workup of intracranial aneurysms. J Neurosurg 85: , Mamata Y, Muro I, Matsumae M, et al: Magnetic resonance cisternography for visualization of intracisternal fine structures. J Neurosurg 88: , Nowinski WL, Fang A, Nguyen BT, et al: Multiple brain atlas database and atlas-based neuroimaging system. Comput Aided Surg 2:42 66, Okudera H, Kobayashi S, Nagashima H, et al: Consideration of intraoperative brain shift for frameless stereotaxy, in Tamaki N, Ehara K (eds): Computer-Assisted Neurosurgery. Tokyo: Springer-Verlag, 1997, pp Pedroza A, Dujovny M, Ausman JI, et al: Microvascular anatomy of the interpeduncular fossa. J Neurosurg 64: , Roberts DW, Hartov A, Kennedy FE, et al: Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery 43: , Saeki N, Rhoton AL Jr: Microsurgical anatomy of the upper basilar artery and the posterior circle of Willis. J Neurosurg 46: , Sakata S, Fujii K, Fukui M, et al: [Microsurgical anatomy of perforating branches in the region near the basilar bifurcation], in Gibo H (ed): [Surgical Anatomy for Microneurosurgery V.] Tokyo: SciMed Publications, 1992, pp (Jpn) 27. Sakata S, Rhoton AL Jr: [Surgical anatomy of the posterior cerebral artery and its clinical application], in Yoshimoto S (ed): [Surgical Anatomy for Microneurosurgery VIII.] Tokyo: SciMed Publications, 1995, pp (Jpn) 28. Sugita K, Kobayashi S, Shintani A, et al: Microneurosurgery for aneurysms of the basilar artery. J Neurosurg 51: , Tampieri D, Leblanc R, Oleszek J, et al: Three-dimensional computed tomographic angiography of cerebral aneurysms. Neurosurgery 36: , Yamaura A, Ise H, Makino H: Treatment of aneurysms arising from the terminal portion of the basilar artery with special reference to the radiometric study and accessibility of trans-sylvian approach. Neurol Med Chir 22: , Zeal AA, Rhoton AL Jr: Microsurgical anatomy of the posterior cerebral artery. J Neurosurg 48: , 1978 Manuscript received November 24, Accepted in final form March 12, Address reprint requests to: Toru Koyama, M.D., Department of Neurosurgery, Shinshu University School of Medicine, Asahi, Matsumoto , Japan. 152 J. Neurosurg. / Volume 91 / July, 1999