Preoperative Surgical Planning And Simulation Of Complex Cranial Base Tumors In Virtual Reality
Yi Zhi-Qiang,Li Liang,Mo Da-Peng,Zhang Jia-Yong,Zhang Yang,Bao Sheng-De
DOI: https://doi.org/10.1097/00029330-200806020-00019
IF: 6.133
2008-01-01
Chinese Medical Journal
Abstract:The extremely complex anatomic relationships among bone, tumor, blood vessels and cranial nerves remains a big challenge for cranial base tumor surgery. Therefore, a good understanding of the patient specific anatomy and a preoperative planning are helpful and crucial for the neurosurgeons. Three dimensional (3-D) visualization of various imaging techniques have been widely explored to enhance the comprehension of volumetric data for surgical planning.1 We used the Destroscope Virtual Reality (VR) System (Singapore, Volume Interaction Pte Ltd, software: RadioDexterTM 1.0) to optimize preoperative plan in the complex cranial base tumors. This system uses patient-specific, coregistered, fused radiology data sets that may be viewed stereoscopically and can be manipulated in a virtual reality environment. This article describes our experience with the Destroscope VR system in preoperative surgical planning and simulation for 5 patients with complex cranial base tumors and evaluates the clinical usefulness of this system. CLINICAL DATA Patient series Between July 2007 and November 2007, we used the Destroscope VR system to carry out preoperative surgical planning and simulation for 5 patients with complex skull base tumors. The clinical data for these patients are summarized in Table.Table: Clinical data of 5 patients with complex cranial base tumorsData acquisition and 3-D object rendering Magnetic resonance imaging (MRI) studies were acquired on a General Electric (GE) 1.5-T unit (FOV=24 cm, matrix=512 × 512, slice thickness=2 mm). The computerized tomography (CT) data were obtained as axial, contiguous 1.0mm slices (FOV=24 cm) on a GE scanner (16-D). The MRI, CT and CT angiography (CTA), etc were scanned with identical FOV for the sake of automatic coregistration in different image modalities. The scanning of the patients was scheduled 3 to 5 days before the surgery, and the imaging data were transferred by CD to the Dextroscope system for reconstruction into 3-D stereoscopic images. The user can freely reach into the 3-D image object as if it were in real space through electromagnetic sensors shaped in the handle and stylus held in the left and right hands respectively. The skull, blood vessels and tumor can be highlighted by adjusting resolution, contrast, transparency and brightness easily; hence, spatial relations were obtained. Preoperative surgical planning and simulation in VR environment Through cropping and drilling the skull and tumor, which can be restored and rehearsed, we can determine the optimal surgical approach, head position and the necessary bone work. In case 1, VR system revealed that the right internal carotid artery (ICA) bifurcation, proximal anterior carotid artery (ACA) and middle carotid artery (MCA) are encased in the tumor. Posterior cerebral artery (PCA) were pushed posteriorly and inferiorly. Good visualization and removal of the tumor was performed through a frontal-temporal-zygomatic approach without orbitotomy. The approach was rehearsed under a simulated intraoperative perspective and the simulation suggested that complete removal of the tumor could be achieved safely (Figure 1).Figure 1.: Supine position with the head left turn 15 degree. A frontal-temporal-zygomatic approach can provide good visualization and exposure for the tumor (green mass).In case 2, the VR system revealed middle cranial fossa (MCF), posterior cranial fossa (PCF), infratemporal fossa (ITF), pterygoid palatine fossa (PPF), parapharyngeal space (PS) and the cranio-cervical junction were involved by the tumor which were mainly fed by the internal maxillary artery (IMA) and ascending pharyngeal artery (APA). The left internal carotid artery (ICA) and vertebral artery (VA) are partially encased. The total resection of the tumor was impossible in any single approach and in one stage due to the massive extent of the tumor and its close relation to vital neurovascular structures. A transmandibular approach to resect the part located in PS to alleviate the dysphagia which most bothered the patient is feasible and safe (Figure 2).Figure 2.: As the main feeding artery for the tumor, the MA and APA are obviously enlarged. The ICA and VA are partially embedded in the extensive tumor (green mass).In case 3, VR system revealed that MCF, ITF, PPF and PS were all involved by the tumor which were mainly fed by the IMA and APA. The finding most worthy of mention is that the ICA was posterior to, not embedded in, the tumor. A combined frontal-temporal and transmandibular approach was rehearsed and selected to be the optimal. In case 4, it was found that the tumor basically occupied the PPF and eroded the great wing of the sphenoid with dura integrity. The degloving transmaxillary approach provided the shortest avenue and good cosmetic results. In case 5, the frontal sinus and left orbital cavity were involved with dura integrity. A transorbital approach was rehearsed to drill the lesion away completely through a 2 cm curved incision aside the inner canthus. Outcomes and follow-up In all the 5 cases, the intraoperative findings correlated well with the preoperative simulation. Total resection was achieved in cases 1, 4 and 5, subtotal in case 3, partial in case 2. No complications related to the operation occurred. In case 4, mild numbness of the right cheek with a possible cause of intraoperative stretching of the inferior orbital nerve occurred postoperatively and the patient began to recover in 3 days and symptoms disappeared within 2 months after operation. In case 2, the chin pulled to the left slightly with the possible reason being a 5th division of facial nerve injury and it has been improving in the 3 months of the follow-up period. In case 3, the dura was spitted and cerebrospinal fluid (CSF) leakage occurred intraoperatively. We sutured the dura closely and reinforced it using lateral pterygoid muscle, buccal pad and gelatin sponge mingled with biogel. The patient was discharged 14 days after operation uneventfully. However, the patient was rehospitalized 1 month postoperatively due to central nervous system (CNS) infection and CSF leakage. The CNS infection was cured by the intravenous and intrathecal administration of sensitive antibiotic combined with consistent CSF external drainage through a lumbar puncture and ventricle puncture for 55 days. Then, a ventriculoperitoneal shunt was successfully placed to treat the hydrocephalus secondary to the CNS infection. Now, in the 7th month after the first operation, the modified Rankin Scale (mRS) of the patient is 3. DISCUSSION Several groups have evaluated the usefulness of a variety of neurosurgical planning programs with 3-D imaging data.2-6 However, common to all these planning systems is the fact that interaction with the increasingly sophisticated 3-D data sets is still achieved in a rather 2-D fashion by moving a screen-bound cursor with a mouse. The Destroscope system enables the surgeon to interact intuitively and in a user friendly manner with complex 3-D graphics in a natural, easy, and direct manner resembling the way one interacts with real objects. Moreover, the user can segment the scalp, skull, blood vessel, lesion, ventricle, etc. and paint them in different colors, which help to reveal the relationship among them. On the individualized 3-D object, one can simulate the operative environment: zooming, measuring (length, width, height, volume, area, angle, etc.), cropping, and drilling (for simulating a drilling or a suctioning tool) in different head positions, surgical approaches and exposure extensions. The Destroscope system has been reported to apply in intracranial tumor operations and in aneurysm clipping simulations.7-11 The Destroscope system also can simulate the angioscopy technique which can be used to observe intralumen structures, such as the atheromatous plaque, angiostenosis and the neck of an aneurysm which will undoubtedly facilitate cerebral vascular disease treatment. The diffusion tensor imaging (DTI) technique can reveal the relationship between the lesions and white matter tracts which is crucial to identify and reserve the vital tracts. Overall, the system substantially contributed to surgical planning primarily by (1) providing a faster and better understanding of the intracranial 3-D anatomic and abnormal spatial relationships, (2) optimizing the craniotomy and the required cranial base exposure through a rehearsal of bone removal, and (3) simulating intraoperative views.7 The visualization and manipulation capabilities of the Destroscope system were the most useful in the area of the cranial base, in which the computer-generated 3-D surgical area of interest usually consists of multitechnique imaging information, depicting soft tissue (MRI), blood vessels (MRA, CTA and MRV), and bone (CT). However, even the most detailed information can not replace the neurosurgeon’s judgment and management in the surgical procedure. In case 3, we attributed the main cause of the CNS infection to the questionable repair for the split dura, which should be fixed from intracranial, not inferiorly to the cranial base. Because the dura fix from extracranial is vulnerable to the buccal cavity movement. The communication between the oral cavity and the intracranial cavity then resulted in severe infection. The biggest lesson learned from the failed case is that we can not overemphasize the importance of skull base reconstruction. Clearly, this Destroscope system is very useful in the training of residents or neurosurgeons unfamiliar with a particular surgical approach. Furthermore, surgeons using this device may be able to preoperatively evaluate various surgical approaches and choose the best one. Likewise, it is an excellent tool in surgical education.12 Though the Destroscope system has demonstrated great potentials, it also has limitations. First, this system provides only a rough approximation of the actual anatomy and pathological features. Many structures, such as the cranial nerves and small blood vessels, are beyond the current resolution of MRI or CT technology and therefore cannot be segmented and displayed as 3-D structures. Second, it does not provide any new information on the properties and intraoperative movement of tissue. Likewise, it does not simulate the feel or texture that a surgeon senses during an operation. However, sensory feedback is very important to a neurosurgeon for a true virtual reality experience. Third, an effective stereoscopic image cannot be rendered in print and snapshots of the planning procedure that could be used for intraoperative guidance can appear only in 2-D format (like the images in this article), thereby losing much of the 3-D information. Furthermore, if the system can be linked directly to the CT and MRI scanners, the data transfer and loading work would be more convenient. Along with the development of the high speed computer and image identification technique, the registration of multimodality images and segmentation of the target structures would be highly automated. This application of the Destroscope system integrated with the intraoperative navigation will provide surgeons more extensive and practical potential in the surgery.