Peer-Review Reports

A Three-Dimensional Computer-Based Perspective of the Skull Base Matteo de Notaris1,4, Kenneth Palma1, Luis Serra5, Joaquim Ensen˜at1, Isam Alobid2, Jose´ Poblete1, Joan Berenguer Gonzalez3, Domenico Solari6, Enrique Ferrer1, Alberto Prats-Galino4

Key words 3D computer model - Dextroscope - Endoscopic endonasal surgery - Skull base - Surgical anatomy -

Abbreviations and Acronyms 3D: Three-dimensional CT: Computed tomography VRS: Virtual reality system From the Departments of 1 Neurosurgery, 2Otorhinolaryngology, Rhinology Unit, and 3Radiology, Neuroradiology Division, Hospital Clinic de Barcelona, and 4Laboratory of Surgical Neuroanatomy, Faculty of Medicine, Universitat de Barcelona, Barcelona, Spain; 5Center for Computational Imaging and Simulation Technologies in Biomedicine, Information and Communication Technologies Department, Universitat Pompeu Fabra, Barcelona, Spain; and 6 Department of Neurological Sciences, Division of Neurosurgery, Università degli Studi di Napoli Federico II, Naples, Italy To whom correspondence should be addressed: Matteo de Notaris, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2014) 82, 6S:S41-S48. http://dx.doi.org/10.1016/j.wneu.2014.07.024 Supplementary digital content available online. Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com

- OBJECTIVE:

To describe our designed protocol for the reconstruction of threedimensional (3D) models applied to various endoscopic endonasal approaches that allows performing a 3D virtual dissection of the desired approach and analyzing and quantifying critical surgical landmarks.

- METHODS:

All human cadaveric heads were dissected at the Laboratory of Surgical Neuroanatomy of the University of Barcelona. The dissection anatomic protocol was designed as follows: 1) virtual surgery simulation systems, 2) navigated cadaver dissection, and 3) postdissection analysis and quantification of data.

- RESULTS:

The virtual dissection of the selected approach, the preliminary exploration of each specimen, the real dissection laboratory experience, and the analysis of data retrieved during the dissection step provide a complete method to improve general knowledge of the main endoscopic endonasal approaches to the skull base, at the same time allowing the development of new surgical techniques.

- CONCLUSIONS:

The methodology for surgical training in the anatomic laboratory described in this article has proven to be very effective, producing a depiction of anatomic landmarks as well as 3D visual feedback that improves the study, design, and execution in various neurosurgical approaches. The Dextroscope as a virtual surgery simulation system can be used as a preoperative planning tool that can allow the neurosurgeon to perceive, practice reasoning, and manipulate 3D representations using the transsphenoidal perspective acquiring specifically visual information for endoscopic endonasal approaches to the skull base. The Dextroscope also can be used as an advanced tool for analytic purposes to perform different types of measurements between surgical landmarks before, during, and after dissection.

1878-8750/$ - see front matter ª 2014 Elsevier Inc. All rights reserved.

INTRODUCTION The skull baserepresents the floor of the cranial cavity on which the undersurface of the brain rests. Neurovascular structures that are crucial to the function and integrity of the entire central nervous system are contained and travel within the skull base. A precise working knowledge of the surgical anatomy is essential for effective treatment of diseases in such a complex region. In recent decades, as minimally invasive neurosurgery has progressed, there has been an increase in the depth and complexity of required anatomic knowledge for the delivery of adequate treatment. The acquisition of new surgical techniques and skills taught at hospital

residency training programs has also grown more complex in a parallel fashion. Despite restrictions for residents on maximum work hours per week, a limited number of instructors, and legal and ethical concerns, laboratory dissection, hands-on experience, and observerships still remain the “gold standard.” This situation is especially true in the field of endoscopic endonasal skull base surgery, in which factors such as the transition from microscopic techniques, visualization, and instruments to endoscopic ones restrict training to a few specialized centers and out-residency workshops all over the world. Surgical laboratories remain the optimal choice for acquiring both the skills and the anatomic knowledge through anatomic dissection (3). Such laboratories still

WORLD NEUROSURGERY 82 [6S]: S41-S48, DECEMBER 2014

represent the best strategy to make surgery as realistic as possible compared with a surgical scenario. Over the years, hands-on anatomic dissection has allowed surgeons to learn in a low-stress environment where mistakes are permissible and procedures can be repeated multiple times. However, such a process is not without limitations. Surgical laboratories are subject to shortage of cadaveric specimens, partially secondary to legislative restrictions in Europe, and appropriate surgical instrumentation is not as easily available as in the operating room. A potential new field in neurosurgical training arises as rapid advances in medical image data processing extend the use of computed tomography (CT) and magnetic resonance imaging beyond their purely diagnostic use. Computerized

www.WORLDNEUROSURGERY.org

S41

PEER-REVIEW REPORTS MATTEO DE NOTARIS ET AL.

three-dimensional (3D) models of cadaveric specimens obtained through CT and magnetic resonance imaging before dissection and the use of advanced virtual reality systems (VRS) coupled with neuronavigation provide reliable tools for the analysis of approaches from both a quantitative and a qualitative perspective. Quantitatively, these tools allow precise measurements of simple data, such as linear and angular distances, as well as calculation of more complex variables (i.e., the area exposed during a selected approach, the volume of the drilled bone while performing the surgical pathway, and the surgical freedom) (5, 9, 10). Qualitatively, 3D models serve as a highly effective means for the comprehension of anatomic structures and their relationships, whereas VRS enhance this process through dynamic manipulation. Neuronavigation and VRS have proven highly effective for the planning of surgery and better surgical outcomes (1, 2, 6, 20), and they have been effectively exploited for research and educational purposes (11, 12, 19). Our team has designed a protocol for the reconstruction of 3D models of dissected specimens for analytic and anatomic purposes. Its application and results have been described elsewhere (8). In this article, we discuss the addition of a VRS to our laboratory and describe our newly developed techniques of 3D visualization, predissection planning, and quantitative analysis applied to the main endoscopic endonasal approaches to the skull base. We also discuss this device and the experience we gained through the processing of medical data obtained preoperatively and postoperatively. MATERIALS AND METHODS During cadaveric dissections, all procedures (30 specimens) were performed emulating surgical approaches in the operating room to enhance the learning process and obtain as much useful data as possible. Preparation of Specimens Each specimen was fixed with an industrially manufactured embalming mixture composed of different percentages of phenol, ethanol, formaldehyde, and glycerin. The arterial system was injected with red latex. One cadaveric head was injected with a mixture of red latex with barium

S42

www.SCIENCEDIRECT.com

3D COMPUTER-BASED PERSPECTIVE OF SKULL BASE

sulfate at 0.015% to assess the feasibility of performing cadaveric CT angiography. Acquisition of Data Preoperative CT Scan. Before the dissections, CT scans were performed on each cadaver head. A multislice helical acquisition protocol (slice thickness 0.6 mm, gantry angle 0 degrees) was adopted; heads were precisely positioned in the scanner (SOMATOM Sensation 64; Siemens AG, Erlangen, Germany) to obtain a projection perpendicular to the palate. The images achieved were stored in a picture archiving and communication system that streamlined the “postproduction” management. Virtual 3D Model: Dextroscope. We added the Dextroscope (Volume Interactions Pte. Ltd., Singapore, Singapore), a dedicated VRS, to our laboratory (16). This device was used for specific surgical approaches using CT Digital Imaging and Communications in Medicine images of the specimens for which predissection planning was considered appropriate, mainly for the execution of complex endoscopic endonasal skull base approaches. With the Dextroscope, the user works with both hands inside a stereoscopic virtual workspace; this is achieved by reflecting a computer-generated 3D via a mirror into the user’s eye (Figure 1). Wearing liquid display shutter glasses synchronized with the time split display, the user reaches with both hands behind the mirror into the 3D data. Electromagnetic sensors in both hands convey the interaction and allow manipulation of the 3D data in real time. The user holds an ergonomically shaped handle to move the 3D data freely as if it were an object held in real space. The other hand holds a pen-shaped instrument that appears inside the virtual reality workspace as a computer-generated instrument that can be used to perform specific data manipulation. The Dextroscope software supports multimodality imaging, allowing the visualization (visual fusion) of several imaging modalities and complementary graphic objects sharing the same virtual space. To work on these imaging modalities, it provides a range of virtual tools that allow manual and semiautomatic segmentation; image coregistration; curved, linear, angular, and volumetric measurements; and removal of virtual tissue (which can be used to simulate bone drilling or soft tissue

Figure 1. The Dextroscope bimanual neurosurgical planning and simulation system.

resection). These tools are complemented by reporting tools such as screen captures and video making. Image-Guided Dissection. All dissections were performed in the Laboratory of Surgical Neuroanatomy, at the University of Barcelona, Barcelona, Spain. Various endoscopic endonasal transsphenoidal approaches to the midline skull base, to the cavernous sinus, and to the orbit were performed using rigid 0- and 45-degree endoscopes, 18 cm in length and 4 mm in diameter (KARL STORZ GmbH & Co., Tuttlingen, Germany), as the sole visualizing instruments. Each dissection was aided with the use of a neuronavigation system (StealthStation AxiEM; Medtronic). During this step, coordinates of relevant landmarks were retrieved to measure postoperatively the area exposed by the selected approach (Figure 2). Postoperative CT and 3D Model. After each dissection, a second, postoperative CT scan with the same characteristics as the

Figure 2. Laboratory of Surgical NeuroAnatomy of Barcelona (Spain): anatomical dissection step with the aid of the neuronavigation system.

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.024

PEER-REVIEW REPORTS MATTEO DE NOTARIS ET AL.

3D COMPUTER-BASED PERSPECTIVE OF SKULL BASE

preoperative CT scan was performed. The CT acquisitions obtained before and after dissection were coregistered and used for the development of a 3D model.

RESULTS Key Anatomic Landmarks for Extended Approaches: 3D Relationships Between Surgical Planning and Dissection Procedure To simulate the position of the head in the operating room in accordance with the type of approach, the head is extended or flexed for about 10e30 degrees. The head is more extended in the case of an anterior cranial base approach to provide a more anterior trajectory; the head is more flexed in the case of a sellar transclival approach because the trajectory of the route is inferior and directed downward. The extended approaches according to basic principles by Kassam et al. (13-15) require some preliminary steps aimed at increasing the working space and the maneuverability of the instruments inside the nose. In such manner, some essential surgical maneuvers are common to all extended approaches. Unilateral or bilateral middle turbinectomy, depending on the extension of the lesion, is usually performed. The external configuration and variant anatomy of the turbinates can be easily recognized preoperatively on the Dextroscope (Figure 3). The concha bullosa is a well-pneumatized distention of the middle nasal concha. This is the most common anatomic variation of the middle turbinate that is encountered during the approach. At this point, the posterior portion of the nasal septum is removed to allow a wider view of the contralateral side of the approach, without obstruction during insertion of the working instruments. The configuration of the nasal septum and the anatomic variations must be studied preoperatively on the VRS (Figure 3). Several types of nasal septal deviation may obstruct the surgical corridor and reduce the maneuverability and vision. If the nasal passage is too narrow to perform the procedure safely through an endonasal approach, a small septoplasty may be needed to provide more space and access within the nasal passage. When a septal flap is necessary to reconstruct the skull base defect, a wide antrostomy sometimes is performed. The

Figure 3. The anatomy of the maxillary sinus as seen through the Dextroscope system preoperatively in 2D (A) and in 3D (B, required red and blue 3D anaglyph glasses). IT: inferior turbinate; MT: middle turbinate; NS: nasal septum; EB: ethmoid bulla; UP: uncinate process; EC: ethmoid cell; O: orbit; OF: orbit floor; pwMS: posterior wall of the maxillary sinus.

morphology of the maxillary sinus also can be studied and measured on VRS preoperatively (Figure 3). Moving the endoscope downward and dissecting the mucosa of the inferior surface of the vomer, it is possible to expose the most important landmarks at the level of the inferior meatus: the eustachian tubes, vomer, pterygoid canals, inferior turbinates, and medial pterygoid plates (Figure 4). When the sphenoid sinus is completely unlocked, an ipsilateral or bilateral anterior and posterior ethmoidectomy can be performed depending on the type of extended approach. The ethmoid bulla, one of the most constant anterior ethmoidal air cells, lies just beyond the natural ostium of the maxillary sinus and forms the posterior border of the hiatus semilunaris. The lateral extent of the bulla is the lamina papyracea. Superiorly, the ethmoid bulla may extend to the ethmoid roof, whereas the suprabullar recess is defined by the space between the bulla

and the basal lamella of the middle turbinate. The posterior ethmoid consists of a variable number of air cells. The most lateral border of these air cells is closely related to the lamina papyracea, and the most superior border is closely related to the anterior skull base. The basal lamella of the middle turbinate separates the anterior ethmoidal cells from the posterior ethmoidal cells. The uncinate process is the next key structure to be identified. This L-shaped bone of the lateral nasal wall forms the anterior border of the hiatus semilunaris, or the infundibulum. The anterior and posterior ethmoidal cells can be analyzed, measured, and visualized with an excellent 3D perspective on the Dextroscope (Figure 5). The bone anatomy of the anterior, middle, and posterior cranial base is completely revealed to expose all the main landmarks to perform the extended skull base approaches (Figures 6 and 7). Subsequently, a wide anterior sphenoidotomy is

Figure 4. The most important landmarks at the level of the inferior meatus in 2D (A) and in 3D (B, required red and blue 3D anaglyph glasses). V: vomer; LCo: left choana; RCo: right choana; ET: eustachian tube; C: clivus; SS: sphenoid sinus; PC: pterygoid canal; V2: maxillary nerve foramen; LPP: lateral pterygoid plate; mPP: medial pterygoid plate.

WORLD NEUROSURGERY 82 [6S]: S41-S48, DECEMBER 2014

www.WORLDNEUROSURGERY.org

S43

PEER-REVIEW REPORTS MATTEO DE NOTARIS ET AL.

3D COMPUTER-BASED PERSPECTIVE OF SKULL BASE

Figure 5. The anterior/posterior ethmoidal cells, the maxillary and frontal sinus can be visualized with an excellent perspective on the Dextroscope in 2D (A) and in 3D (B, required red and blue 3D anaglyph glasses). MS: maxillary sinus; UP: uncinate process; EB: ethmoid bulla; SBR: suprabullar recess; PEC: posterior ethmoid cells; FS: frontal sinus.

sella turcica, can be encountered, including performed to identify the key landmarks conchal, presellar, sellar, and postsellar within the posterior wall of the sphenoid types. The presence or absence of intrasinus: the sellar floor, planum sphenoisphenoid septations is also a key point to dale, medial and lateral optocarotid the precise localization of the landmarks recesses, clival indentation, and, laterally, over the posterior wall of the sphenoid the bony prominences of the intrasinus. Single or multiple septations and the cavernous carotid arteries and optic nerves region of insertion—in the sellar floor, at (Figure 8). The sphenoid sinus embraces the carotid canal, or at the optic canal—are all the key anatomic landmarks that are important factors to determine typically used for orientation before going to the dissection during complex skull base laboratory or to the operating approaches (Figure 8A). Preroom (Figure 8). operative evaluation of the The virtual patient bone anatomy of the sphenoid sinus Video available at anatomy can be explored using using the Dextroscope is a very WORLDNEUROSURGERY.org the drill tool of the Dextroimportant step and can scope. This virtual tool is directly affect the surgical plan controlled by the dexterous hand of the (Video I). user, behaving like a handheld drill if Different anatomic variations of spheapplied to bone tissue (or like a resection noid sinus pneumatization, depending on tool if applied to soft tissue). The drill tool the position of the sinus in relation to the

Figure 6. The bone anatomy of the anterior, middle, and posterior cranial base is completely revealed to expose all the main anatomical landmarks in 2D (A) and in 3D (B, required red and blue 3D anaglyph glasses). V: vomer; SS: sphenoid sinus; PC: pterygoid canal; FR: foramen rotundum; LPP: lateral pterygoid plate; mPP: medial pterygoid plate; C: clivus; SOF: superior orbital fissure.

S44

www.SCIENCEDIRECT.com

changes the visibility of voxels of the CT image, from opaque to completely transparent, so that the user can remove bone and reveal the tissue behind. This “drilled” bone can also be changed back to its original opaque value using a restore tool. Using this tool, the user can explore the sphenoid sinus and perform an anterior sphenoidotomy to access the sphenoid sinus cavity. Once inside, the position of the sella and degree of pneumatization of the posterior wall of the sinus can be visualized using the same anteroinferior perspective of the endonasal approach (Figure 9B and C). The configuration of the suprasellar region can be also explored. While performing a transtuberculum-transplanum approach, it is very useful to calculate preoperatively the suprasellar notch, which coincides with a sort of indentation between the superior aspect of the sella turcica and the declining part of the planum sphenoidale (Figure 9A). This anatomic structure was described more recently by our group (7). Using the Dextroscope preoperatively, it is possible to calculate with precision this angle using the angle measuring tool of the system. The angle measuring tool allows the user to position 3 points manually and obtain the corresponding angle formed. Using the drill tool, it is possible to simulate a suprasellar osteotomy. Concerning the main vascular structures involved in endoscopic endonasal skull base approaches, the study of the internal carotid artery course and variations represents an essential objective for safe, successful surgery. The horizontal intrapetrous, lacerum, paraclival, clinoidal, ophthalmic, and communicating segments can be analyzed on the Dextroscope taking advantage of the transparency properties of the volumerendering technique available in the Dextroscope. This analysis is done by making the intensities corresponding to bone density almost transparent but still showing the shape of the bone surface as visible edges, while maintaining visible the intensities corresponding to the arteries. It is possible with this technique to reveal the whole carotid canal that encloses the internal carotid artery in the skull as well as the main intracranial arteries (Figure 10). DISCUSSION Endoscopic endonasal extended approaches require a different set of skills

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.024

PEER-REVIEW REPORTS MATTEO DE NOTARIS ET AL.

3D COMPUTER-BASED PERSPECTIVE OF SKULL BASE

Figure 7. The bone anatomy of the anterior, middle, and posterior cranial base is completely revealed to expose all the main anatomical landmarks in 2D (A) and in 3D (B, required red and blue 3D anaglyph glasses). V: vomer; awSS: anterior wall of the sphenoid sinus; PC: pterygoid canal; FR: foramen rotundum; C: clivus; SOF: superior orbital fissure; ASB: anterior skull base; FO: foramen ovale; JF: jugular foramen; D: dens; C2: second cervical vertebra; OC: occipital condyle.

compared with the skills required when practicing traditional open surgery, and these new skills demand a high level of cognitive and technical training to reach proficiency. We have been working on a methodology for endonasal surgical training based in the anatomic laboratory consisting of 3 main pillars: 1) virtual surgery simulation systems, 2) navigated cadaver dissection, and 3) postdissection analysis and quantification of data. Virtual Surgery Simulation Systems The virtual model can be manipulated independently and in parallel with the real dissection process using specimens from which the model originates. This manipulation is done with the use of virtual surgery simulation systems. These systems can be classified broadly into surgical skill training systems and surgical decision-support or decision-making systems, with a combination of both into 1 system being the ultimate

goal. Surgical skill training systems aim at acquiring manual surgical skills, monitoring the performance of the trainee to provide quantifiable guidance. This training does not require patient-specific information and can be accomplished on a series of exemplary cases. Surgical decision-support systems aim at supporting surgical decisions based on patient-specific data to decide whether to operate or not and, if so, to choose strategies and to anticipate problems. The actual skills involved in the treatment are not discussed here. Ideally, a system could incorporate both skill training and decision support, and this should be a worthwhile goal. Surgical Skill Training Systems. The use of virtual reality in skill training for surgical simulation is gaining acceptance as a tool to reduce the learning curve and improve the credentialing process for neurosurgeons based on objective measures of

Figure 8. Anterior sphenoidotomy using the dextroscope is performed to identify the key landmarks within the posterior wall of the sphenoid sinus in 2D (A) and in 3D (B, required red and blue 3D anaglyph glasses). V: vomer; LRSS: lateral recess of the sphenoid sinus; O: orbit; SS: sphenoid sinus; S: spenoidal septation; ASB: anterior skull base.

WORLD NEUROSURGERY 82 [6S]: S41-S48, DECEMBER 2014

operator performance. This acceptance is manifested in the fact that there are already several products entering this market from companies such as Simbionix USA Corporation (Cleveland, Ohio, USA), Mentice AB (Gothenburg, Sweden), and Surgical Science Sweden AB (Gothenburg, Sweden). As a consequence, many centers worldwide are beginning to adopt them. These new devices represent a paradigm shift for new surgical procedures. This approach has great potential to allow physicians to acquire meaningful new procedural skills, while tracking and objectively quantifying their learning curve, without endangering patient safety in the process. The procedure initially can be performed in the simulation skills laboratory on virtual patients before moving to the dissection laboratory and then to the operating room, reducing the risk for real patients. To date, numerous techniques have been developed for the assessment of manual dexterity and hand-eye coordination with the combined use of virtual and mixed reality simulators, with virtual temporal bone dissections having taken the lead as a result of the fact that the available technology to simulate bone drilling (using force feedback devices) has found a close match in the real bone drilling procedure (17, 21, 22). This technology has resulted in the availability of commercial products such as VOXELMAN Tempo (Voxel-Man, Hamburg, Germany), which claims to incorporate patient-specific data in simulations (18). There has been less research and development with regard to endoscopic sinus surgery. Voxel-Man more recently announced the release of their VOXELMAN Sinus product, which according to them is “based on the same mechanical setup as VOXEL-MAN Tempo” [http:// www.voxel-man.com] and “[t]he input devices serve as cutting instrument and endoscope, respectively.” Also, researchers at the National Research Council of Canada, who developed NeuroTouch, an open neurosurgery training system, announced a new module for endonasal surgery, but this module is not yet available for review (http://www.neurotouch.ca/). Surgical Decision-Support Systems. Surgical decision-support systems are required to support patient-specific data and function

www.WORLDNEUROSURGERY.org

S45

PEER-REVIEW REPORTS MATTEO DE NOTARIS ET AL.

3D COMPUTER-BASED PERSPECTIVE OF SKULL BASE

Figure 9. Sphenoid sinus. The comparison between the dissected cadaver head (A) and preoperatively obtained CT-based 3D reconstructions using the dextroscope in 2D (B) a in 3D (C, required red and blue 3D anaglyph glasses). SF: sellar floor; ICAs: sellar segment of the internal carotid artery;

within the workflow of the clinical environment. These systems need to be able to process patient data fast enough to assist in a decision before the surgery, usually 1e2 days before. Most of the systems in the surgical training category require sophisticated processing of the data to generate the 3D models with which the

ICAc: clival segment of the internal carotid artery; SSN: suprasellar notch; OP: optic protuberance; LOCR: lateral optocarotid recess; PS: planum sphenoidale; PC: pterygoid canal; FR: foramen rotundum.

user will interact. Given that the level of detail required for a realistic simulation is unavailable in most diagnostic imaging data, these training systems are forced to embellish the 3D models with additional 3D structures. Our work with cadaveric data demanded that the virtual surgery simulation system

Figure 10. Using the dextroscope is possible to reveal the whole carotid canal that encloses the internal carotid artery in the skull as well as the main intracranial arteries in 2D (A and C) and in 3D (B and D, required red and blue 3D anaglyph glasses). VA: vertebral artery; SF: sellar floor; ICAP: petrous segment of the internal carotid artery; ICAL: lacerum segment of the internal carotid artery; ICAC: clival segment of the internal carotid artery; ICAS: sellar segment of the internal carotid artery; ICACE: cervical segment of the internal carotid artery; BA: basilar artery; C1: first cervical vertebra; C2: second cervical vertebra; NS: nasal septum; FS: frontal sinus; C: clivus; SS: sphenoid sinus; A1: pre-communicating segment of the anterior cerebral artery; A2: post-communicating segment of the anterior cerebral artery; M1: pre-bifurcation segment of the middle cerebral artery.

S46

www.SCIENCEDIRECT.com

was able to work with real cadaveric data, and the choice of device had to be in the category of decision-support systems. We started using visualization platforms such as Amira (Frankfurt Am Main Area, Germany), and we have reported on this work elsewhere (8). The basic platform of this software offers powerful processing tools and good visualization but has limitations when it comes to user interaction with the 3D models, given that it relies solely on the mouse and keyboard. This software is usually intended for biomedical engineers and offers a complex user interface that is unsuitable for a surgeon. An expert physician who is familiar with anatomy and with the surgical knowledge of each approach is required to process the cadaveric data to generate useful 3D models. The design of the Dextroscope overcomes some of these problems, providing rapid rendering and ergonomic user interface with surgically oriented tools that allow easy quantification of data and manipulation of the model. In the Dextroscope, 3D models of specimens can be visualized before their dissection. These 3D models represent a reliable tool for the detailed understanding of the anatomy of endoscopic endonasal approaches. With just a display of a two-dimensional perspective projection of a complex 3D structure during endoscopic procedures, the full comprehension of the spatial relationships between landmarks is difficult to appreciate. Using computer-generated 3D models, especially the Dextroscope, which offers a stereoscopic view of its renderings, for presurgical planning serves this purpose. These models also

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.024

PEER-REVIEW REPORTS MATTEO DE NOTARIS ET AL.

provide the physician with a sense of confidence in the subsequent steps of the dissection, and their use has been reported to produce a sense of déjà vu. The Dextroscope, as a decision-support tool, can visualize structures along the path of a virtual endoscope from any angle. However, the main limitations include the following: 1) It is unable to give views that accurately match views that would be obtained through the lenses of a real endoscope. 2) It still lacks some key dedicated tools that would greatly simplify its usage. 3) It cannot change virtual lenses (it is fixed at 40 degrees, the field of view of 1 person looking at the computer screen). 4) It cannot move the viewing point as an actual endoscope (it moves the 3D model, the virtual head, as though one were holding the head in one’s hand). 5) It cannot feel the constraints of movement of an endoscope along the nostrils (it freely moves in and out of the path). 6) It provides only a linear surgical path trajectory tool. The Dextroscope is intended as a decision-making tool, with extended applicability as an advanced anatomy and surgery training system. It is not intended for training the manual skills of the surgeon. It addresses the issue of what would be the best corridor to approach a lesion on a specific patient (i.e., what size craniotomy and where, what structures will be encountered on the path). It does not address manual skills of how to train the surgeon on the technique to treat the lesion (4). When combined with image-guided cadaveric dissection, the Dextroscope provides a power complement to the study of anatomy and pathology as well as the exploration and validation of new surgical approaches. The training capabilities of the Dextroscope are limited to rigid object manipulation, and it does not support any tissue deformation. It does not provide support for force-feedback (haptics) to the user to enable the mechanical feeling of objects. The user interactions are limited to manipulation of rigid objects without sensing mechanical properties (how soft is certain tissue, how hard is certain bone). Finally, the software does not incorporate any training scenario so that the performance of the surgeon can be measured and analyzed. Image-Guided Cadaver Dissection Cadaveric dissection represents an irreplaceable experience (3). The skull base

3D COMPUTER-BASED PERSPECTIVE OF SKULL BASE

surgeon requires specific training to achieve competency in his or her field of practice. For instance, during endoscopic surgery, the surgeon’s direct view is often restricted, requiring a higher degree of manual dexterity. The complexity of the instrument controls, restricted vision and mobility, and difficult hand-eye coordination are major obstacles in performing such procedures. Basic skills such as craniotomies and more advanced drilling techniques as well as endoscope handling should be acquired and improved in the anatomic laboratory using cadavers to obtain the skills closest to the real surgical procedures in the operating room. We routinely combine cadaveric dissection with the use of a navigation system. This combined approach allows us to link the dissected specimen (a real physical object) and the virtual model to obtain both photographic and 3D renderings of the data that serve to validate our work.

perceive, practice reasoning, and manipulate 3D representations of the skull base using the transsphenoidal perspective and acquiring specific visual information for endoscopic endonasal approaches. The main advantages of this system for such approaches are as follows:

Postdissection Analysis and Quantification of Data The postdissection step provides the actual quantification of the approach as executed during the dissection in the laboratory as described elsewhere (8). In our laboratory, we bring together the information obtained from the image-guided cadaveric dissection with the information obtained from the virtual surgical simulation systems to analyze the surgical performance, providing a valuable representation of the 3D visual feedback of each approach. Data analysis is a fundamental step to discover critical landmarks in complex surgical environments, to improve general knowledge of each approach, and to “predict” the surgical anatomy. It also provides the opportunity to compare different neurosurgical approaches in terms of effectiveness and maneuverability to reach the surgical target.

 The ability to analyze, taking advantage of the transparency properties of the volume-rendering technique, the main vascular structures involved in endoscopic endonasal skull base approaches, in particular, the horizontal intrapetrous, lacerum, paraclival, clinoidal, ophthalmic, and communicating segments of the internal carotid artery

CONCLUSIONS We believe that our working methodology for surgical training in the anatomic laboratory has proven to be very effective, producing a depiction of anatomic landmarks as well as 3D visual feedback that improves the study, design, and execution in various neurosurgical approaches. The Dextroscope as a virtual surgery simulation system can serve as a preoperative planning tool to allow the neurosurgeon to

WORLD NEUROSURGERY 82 [6S]: S41-S48, DECEMBER 2014

 The ability, using a standard CT scan, to recognize and analyze preoperatively anatomic structures and landmarks, in particular: B

During nasal step: the nasal turbinates, bulla ethmoidalis, uncinate process, choana, eustachian tubes, vomer, and nasal septum

B

During sphenoidal step: the parasellar and paraclival segments of the internal carotid arteries, optic protuberances, medial and lateral optocarotid recesses, sellar floor, clival indentation, suprasellar notch, and intrasphenoidal septations

 The ability to quantify volumetric data by means of its measurement tools at the preoperative, intraoperative, and postoperative stage of the dissection process. The main disadvantage of this method is that it is time-consuming for daily surgical practice; however, it is highly recommended for complex skull base approaches. Another important disadvantage is that training capabilities are limited to rigid object manipulation, and it does not support any tissue deformation. It does not provide support for force-feedback (haptics) to the user to enable the mechanical feeling of objects. Our system offers the ability to study in detail the surgical anatomy in a realistic surgical training rehearsal. The real spatial perception of the transsphenoidal route is difficult to acquire with only traditional one-dimensional or two-dimensional images, so the use of a 3D imaging system in the analysis of surgical anatomy could be

www.WORLDNEUROSURGERY.org

S47

PEER-REVIEW REPORTS MATTEO DE NOTARIS ET AL.

very useful. In selected complex cases such as skull base lesions or vascular malformations, this system could be adopted as a complementary advanced preoperative planning system, providing the ability to tailor the approach, addressing properly any troublesome step that could be encountered during the surgical maneuvers. Finally, the development of 3D models and VRS training methods for skull base approaches has many implications for the future, including the ability to measure a physician’s technical skills and clinical judgment regularly. The availability of simulation systems allows neurosurgeons to practice very difficult procedures before performing them in vivo. The implementation of these new systems by many centers around the world could change the way new, high-risk procedures are introduced and may prove the old adage that repeated “practice makes perfect.” REFERENCES 1. Apuzzo ML: The Richard C. Schneider Lecture. New dimensions of neurosurgery in the realm of high technology: possibilities, practicalities, realities. Neurosurgery 38:625-637 [discussion 637639], 1996. 2. Apuzzo ML: Virtual neurosurgery: forceps, scissors, and suction meet the microprocessor, rocket science, and nuclear physics. Neurosurgery 64: 785, 2009. 3. Cappabianca P, Magro F: The lesson of anatomy. Surg Neurol 71:597-598 [discussion 598-599], 2009. 4. Caversaccio M, Eichenberger A, Hausler R: Virtual simulator as a training tool for endonasal surgery. Am J Rhinol 17:283-290, 2003. 5. Chang SW, Wu A, Gore P, Beres E, Porter RW, Preul MC, Spetzler RF, Bambakidis NC: Quantitative comparison of Kawase’s approach versus the retrosigmoid approach: implications for tumors involving both middle and posterior fossae. Neurosurgery 64 (3 Suppl):ons44-ons51 [discussion ons51-52], 2009. 6. Cusimano MD: Virtual reality surgery: neurosurgery and the contemporary landscape a threedimensional interactive virtual dissection model

S48

www.SCIENCEDIRECT.com

3D COMPUTER-BASED PERSPECTIVE OF SKULL BASE

to simulate transpetrous surgical avenues. Neurosurgery 53:1010-1011 [author reply 10111012], 2003. 7. de Notaris M, Solari D, Cavallo LM, D’Enza AI, Enseñat J, Berenguer J, Ferrer E, Prats-Galino A, Cappabianca P: The “suprasellar notch,” or the tuberculum sellae as seen from below: definition, features, and clinical implications from an endoscopic endonasal perspective. J Neurosurg 116: 622-629, 2012. 8. de Notaris M, Solari D, Cavallo LM, Enseñat J, Alobid I, Soria G, Gonzalez JB, Ferrer E, PratsGalino A: The use of a three-dimensional novel computer-based model for analysis of the endonasal endoscopic approach to the midline skull base. World Neurosurg 75:106-113 [discussion 136-140], 2011. 9. de Notaris M, Topczewski T, de Angelis M, Enseñat J, Alobid I, Gondolbleu AM, Soria G, Gonzalez JB, Ferrer E, Prats-Galino A: Anatomic skull base education using advanced neuroimaging techniques. World Neurosurg 79 (2 Suppl):S16.e9-S16.e13, 2013. 10. Figueiredo EG, Deshmukh P, Zabramski JM, Preul MC, Crawford NR, Siwanuwatn R, Spetzler RF: Quantitative anatomic study of three surgical approaches to the anterior communicating artery complex. Neurosurgery 56 (2 Suppl): 397-405 [discussion 397-405], 2005. 11. Figueiredo EG, Deshmukh P, Nakaji P, Crusius MU, Teixeira MJ, Spetzler RF, Preul MC: Anterior selective amygdalohippocampectomy: technical description and microsurgical anatomy. Neurosurgery 66 (3 Suppl Operative):45-53, 2010. 12. Gu SX, Yang DL, Cui DM, Xu QW, Che XM, Wu JS, Li WS: Anatomical studies on the temporal bridging veins with Dextroscope and its application in tumor surgery across the middle and posterior fossa. Clin Neurol Neurosurg 113: 889-894, 2011.

rostrocaudal axis. Part II: posterior clinoids to the foramen magnum. Neurosurg Focus 19:E4, 2005. 16. Kockro RA, Serra L, Tseng-Tsai Y, Chan C, YihYian S, Gim-Guan C, Lee E, Hoe LY, Hern N, Nowinski WL: Planning and simulation of neurosurgery in a virtual reality environment. Neurosurgery 46:118-135 [discussion 135-137], 2000. 17. Pflesser B, Petersik A, Tiede U, Hohne KH, Leuwer R: Volume cutting for virtual petrous bone surgery. Comput Aided Surg 7:74-83, 2002. 18. Pflesser B, Petersik A, Pommert A, Riemer M, Schubert R, Tiede U, Höhne KH, Schumacher U, Richter E: Exploring the visible human’s inner organs with the VOXEL-MAN 3D navigator. Stud Health Technol Inform 81:379-385, 2001. 19. Safavi-Abbasi S, de Oliveira JG, Deshmukh P, Reis CV, Brasiliense LB, Crawford NR, FeizErfan I, Spetzler RF, Preul MC: The craniocaudal extension of posterolateral approaches and their combination: a quantitative anatomic and clinical analysis. Neurosurgery 66 (3 Suppl Operative): 54-64, 2010. 20. Stadie AT, Kockro RA, Reisch R, Tropine A, Boor S, Stoeter P, Perneczky A: Virtual reality system for planning minimally invasive neurosurgery. Technical note. J Neurosurg 108:382-394, 2008. 21. Wiet GJ, Bryan J, Dodson E, Sessanna D, Stredney D, Schmalbrock P, Welling B: Virtual temporal bone dissection simulation. Stud Health Technol Inform 70:378-384, 2000. 22. Wiet GJ, Stredney D, Kerwin T, Hittle B, Fernandez SA, Abdel-Rasoul M, Welling DB: Virtual temporal bone dissection system: OSU virtual temporal bone system: development and testing. Laryngoscope 122 (Suppl 1):S1-S12, 2012.

13. Kassam AB, Gardner P, Snyderman C, Mintz A, Carrau R: Expanded endonasal approach: fully endoscopic, completely transnasal approach to the middle third of the clivus, petrous bone, middle cranial fossa, and infratemporal fossa. Neurosurg Focus 19:E6, 2005.

Conflict of interest statement: This research was partially supported by the Maratò TV3 Grant Project (411/U/2011— title: Quantitative analysis and computer aided simulation of minimally invasive approaches for intracranial vascular lesions.).

14. Kassam A, Snyderman CH, Mintz A, Gardner P, Carrau RL: Expanded endonasal approach: the rostrocaudal axis. Part I: crista galli to the sella turcica. Neurosurg Focus 19:E3, 2005.

Citation: World Neurosurg. (2014) 82, 6S:S41-S48. http://dx.doi.org/10.1016/j.wneu.2014.07.024

15. Kassam A, Snyderman CH, Mintz A, Gardner P, Carrau RL: Expanded endonasal approach: the

1878-8750/$ - see front matter ª 2014 Elsevier Inc. All rights reserved.

Received 18 September 2013; accepted 25 July 2014

Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com

WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.024

A three-dimensional computer-based perspective of the skull base.

To describe our designed protocol for the reconstruction of three-dimensional (3D) models applied to various endoscopic endonasal approaches that allo...
2MB Sizes 4 Downloads 7 Views