Stereoscopic virtual realistic surgical simulation in intracranial aneurysms
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.96399
Source of Support: None, Conflict of Interest: None
Background: Three-dimensional (3D)-computed tomographic angiography (CTA) has been widely used for surgical simulation of intracranial aneurysms. Stereo imaging technology is progressing rapidly in recent years and stereo imaging may make more realistic surgical simulation possible. Therefore, we aimed at the establishment of a technique for stereoscopic viewing of minute volume rendering images while pursuing a low cost. Materials and Methods: Between January 2009 and June 2011, 54 patients with ruptured intracranial aneurysms were enrolled in this study. CTA data was transferred to the workstation equipped with image-processing software, and multilayer fusion images were processed by neurosurgeons. Image data for stereoscopic viewing of multilayer fusion image from arbitrary directions were collected form rotational trajectories around an aneurysm and were output to MPEG file. Stereoscopic viewing using MPEG data was achieved by the freeware named Stereo Movie Maker. Stereo viewing method using QuickTime VR format was also tried. Results: Multilayer fusion image created from CTA data displayed clearly the anatomical information about not only the aneurysm but also the surrounding structures, such as parent artery, venous system, brain tissue, skull bone, and scalp. The quality of the resulting multilayer fusion image was suitable for surgical simulation with virtual reality. Virtual realistic surgical simulation became possible by the combination of minute multilayer fusion image and stereoscopic viewing by our method. Conclusions: Our method for stereo viewing of multilayer fusion images resulted in an improvement in the capability of diagnostic imaging and the image-guided support for neurosurgical procedures in intracranial aneurysm.
Keywords: 3D-CTA, intracranial aneurysm, stereo imaging, surgical simulation
Prognosis of aneurysmal subarachnoid hemorrhage (aSAH) remains poor with a mortality rate of 25-50%, and 10-20% of the affected patients remain severely disabled.  The factor that most correlates with prognosis is the severity of illness at the time of first rupture of aneurysm, and aneurysm rerupture can greatly aggravate the condition of the patient. In Japan, surgical treatment for aSAH, craniotomy, or endovascular treatment is usually performed "early," wherein a surgical procedure for a ruptured aneurysm is performed within 72 h of the onset of SAH in order to prevent rerupture. , Moreover, for patients with Hunt and Kosnik grades I-IV, urgent surgical procedure is usually performed within a few hours of the onset of aneurysmal rupture. Preoperative accurate diagnostic imaging and optimal surgical planning are keys to the success of the surgical procedure. Although there is sufficient preparation time in a scheduled surgery, this may not be the case in emergency surgeries. Therefore, it is necessary to set up the optimal surgical plan promptly in an emergency surgery. In the recent years, three-dimensional (3D) images can be created easily by advanced imaging modalities and computer technologies, and 3D medical imaging has become indispensable to surgical planning. Although imaging modalities other than computed tomography (CT) and a surgical navigation system can be used in a scheduled surgery, in an emergency surgery, they cannot be used owing to shortage of preparation time. The imaging modalities that can be used in an emergency surgery are 3D-CT angiography (CTA) and/or digital subtraction angiography (DSA). Thus, it is important to perform accurate surgical simulations using the data of 3D-CTA and/or DSA. Although DSA remains the gold standard for the diagnosis of intracranial aneurysms, it is invasive and time consuming, and has the risk of aneurysmal rerupture and associated complications, with approximately 0.7% occurrence of permanent neurological injury and less than 0.1% mortality rate.  Considering these situations, we aimed to create realistic surgical simulation images only using 3D-CTA data. Meanwhile, stereoscopic imaging using advanced technology to generate more realistic surgical simulation should be pursued. Then, we aimed to develop a low-cost technique for stereoscopic viewing of precise surgical simulation images, with a low set-up time to enable its use in emergency surgeries.
Between January 2009 and June 2011, 54 patients (25 men and 29 women) with ruptured intracranial aneurysm treated by craniotomy (neck clipping) were enrolled in this study. Patients treated with endovascular procedures were excluded. The mean age of the patients included was 58.4 (range, 19-79) years. The locations of the aneurysms were as follows: internal carotid artery-posterior communicating artery (IC-PC) (n = 24), anterior communicating artery (Acom) (n = 21), and middle cerebral artery (n = 9). All patients in this series underwent CTA. It is our policy to treat ruptured basilar artery aneurysm and IC-ophthalmic aneurysms with endovascular procedures and to treat other aneurysms with craniotomy. In patients with suspected carotid stenosis due to neck bruit, we considered it necessary to evaluate the real circulatory status. Therefore, we undertake angiography only for these patients.
Multislice helical CT scanner of 64 detector rows (Aquilion 64; Toshiba Medical Systems, Tochigi, Japan) was used for CTA. All patients were positioned supine with the head maintained in a neutral position; 80 mL of nonionic iodinated contrast medium (Iopamiron 300; Bayer, Leverkusen, Germany) was injected into the cubital vein with an automated injector at a flow rate of 5 mL/s. Then, enhanced images (140 kV and Auto mA; gantry rotation speed, 0.6 s/rotation) were obtained by helical scanning (slice thickness, 0.625 × 64 collimation). The scan delay was set using the Real Prep mode. The scan range included the craniovertebral junction up to the vertex.
Multilayer image fusion
CTA data were transferred to the GE Advantage workstation (GE Medical Systems, Waukesha, WI, USA) equipped with an AW VolumeShare™ (GE Medical Systems). To shorten image processing time, an intra-hospital local area network (LAN) system was built so that our department could acquire Digital Imaging and Communications in Medicine (DICOM) data directly from every imaging modality. This system facilitated us to respond to emergency cases. Image processing began with the segmentation and optimization of each structure, such as the arterial system, venous system, brain tissue, skull, and soft tissue, using AW VolumeShare [Figure 1]d-h. First, the image of the arterial system was segmented [Figure 1]d. To secure the accuracy of volume rendering, the images were segmented while referring to a maximum intensity projection (MIP) image [Figure 1]a-c. After the opacity and color of the image of the arterial system were optimized for detailed observation, the image data were preserved. Visual observation was used to ensure that the aneurysm and surrounding vessels were clearly described in the images. Then, by using a semiautomatic threshold-based segmentation technique, volume images of the brain tissue, skull, and soft tissue were segmented and optimized. Necessary image components were merged and again optimized to obtain the final multilayer fusion image [Figure 1]i-k. Processing of segmentation and fusion of volume data from CTA were performed only by neurosurgeons.
Data acquisition for arbitrary directional stereoscopic images
First, using the newly obtained multilayer fusion image, the most suitable rotational trajectory to observe the aneurysm was set and the image data from 50 directions around the aneurysm were collected while rotating by 7.2° [Figure 2]a and b. Next, the rotational trajectory was tilted 10° up and down, and data were collected similarly. In all cases, image data from 150 or more directions were collected from rotational trajectories around the aneurysm. When image data from further directions were necessary, rotational trajectories for the observation were added [Figure 2]c. Theoretically, image data for stereoscopic viewing from arbitrary directions can be obtained by increasing the rotational trajectories to collect image data. Finally, the collected image data were stored as an MPEG file.
Stereo surgical simulation using MPEG data
We adopted two approaches for stereoscopic imaging. In the first method, MPEG data from the freeware named Stereo Movie Maker (http://stereo.jpn.org/eng/stvmkr/) were used. This software enables various modes of stereo imaging, such as cross-eyed viewing mode, parallel viewing mode, anaglyph mode, and liquid crystal (LC) shutter glass mode [Figure 3].
Stereo surgical simulation using QuickTime Virtual Reality (QTVR) format data
In the second method, we used the QTVR format. First, screen shots of each frame were collected from the MPEG file. These screen shots were converted to the QTVR format using Object 2VR software (Garden Gnome Software e.U., Vienna, Austria), which created a stereoscopic image viewable with QuickTime player (Apple Computer Inc., Cupertino, CA, USA) [Figure 4].
Viewing of stereoscopic surgical simulation image in the operating room
Viewing of stereoscopic surgical simulation images aseptically in the operation room was achieved by wrapping a touch-screen tablet PC (HP Elite Book 2730 p; Hewlett-Packard Development Company, Palo Alto, CA, USA) within a sterile transparent sheet [Figure 5]a. The displayed image could be rotated freely with a digital eraser pen also wrapped in a transparent sterile film [Figure 5]b.
DSA was performed using a biplane DSA unit (Infinix Celeve; Toshiba Medical Systems, Otawara city, Japan) with a matrix resolution of 1024 × 1024 pixels.
Multilayer fusion images
Multilayer fusion images were successfully created in all cases. The mean time needed for image processing after DICOM data acquisition was as follows: artery, 5 min; artery and venous system, 12 min; artery, venous system, and skull, 15 min; and artery, venous system, skull, and brain tissue, 23 min. More complicated fusion images, such as a 5-layer fusion image simulating craniotomy, were completed in less than 40 min. Volume rendering images created only from the artery information provided a satisfactory description of aneurysm in all 34 cases. In the cases in which DSA was performed, images of the arterial layer were equivalent to those obtained by DSA for evaluating the characteristics of each aneurysm (location, size, and direction). Arterial layer images also provided an adequate description of the main perforating arteries such as the anterior choroidal artery and Heubner's artery. The venous system consisting of the sylvian vein, superior sagittal sinus, and cortical vein that are encountered during surgery was also described clearly. Even in aneurysms located at the skull base, clear separation and depictions of the aneurysms and bone were possible in three-layer fusion images of the artery, venous system, and skull.
Stereoscopic visualization and virtual surgical simulation
Stereoscopic surgical simulations were performed in all patients scheduled for craniotomy. In a usual emergency surgery, we first performed a stereoscopic surgical simulation using MPEG data. In most cases, image data from 150 directions collected from three rotational trajectories were sufficient for stereoscopic surgical simulation of a ruptured aneurysm. The time needed for the export of an MPEG file from collected image data was about 2 min. In this method, virtual realistic surgical simulation was attained by combining minute multilayer fusion images and stereoscopic viewing. Then, the stereo images were converted to the QTVR format. Subsequently, stereoscopic surgical simulation using QuickTime player was also achieved. The QTVR format enabled us to rotate or enlarge images freely and to observe them from particular angles.
In aneurysms of complex shapes, or in Acom aneurysm surrounded by blood vessels of complicated structures, stereoscopic viewing was especially useful. For Acom aneurysms, there are usually three kinds of surgical approaches: right pterional, left pterional, and interhemispheric. The selection of an appropriate surgical approach using 2D-DSA data is usually difficult; however, this selection was greatly aided by stereoscopic reality simulation based on multilayer fusion images.
We used a wide screen to obtain more accurate surgical simulation. Anaglyph or viewing by LC shutter glasses was useful for the doctor who cannot perform cross-eyed viewing.
Although the photographs of the images shown are monoscopic, the actual surgical planning was carried out in stereo.
Case 1: A 63-year-old woman presented with subarachnoid hemorrhage. 3D-CTA revealed Lt IC-PC aneurysm. Stereoscopic simulation was used to better understand the 3D structure around the aneurysm [Figure 6]. Neck clipping of the aneurysm was performed successfully, and her hospital course was uneventful.
Case 2: A 57-year-old man presented with subarachnoid hemorrhage. 3D-CTA revealed Acom aneurysm of a complicated form. Stereoscopic simulation was used to select the surgical approach and understand the 3D structure around the aneurysm. Stereoscopic simulation of the aforementioned three surgical approaches indicated that the left pterional approach was reasonable [Figure 7]. Neck clipping of the aneurysm was performed successfully. The postoperative hospital course was uneventful.
The goal of treatment for ruptured aneurysms is to ensure complete, quick, and safe exclusion of the aneurysm. To achieve this, it is essential to understand the anatomy of the aneurysm and its surrounding structures before surgery. Only information about the aneurysm and the surrounding arteries is needed for endovascular surgery; however, more detailed anatomical information such as that of the venous system, skull, and soft tissues is required to reduce the surgical risk in craniotomy. In recent years, advancements have been made in 3D image processing technology and 3D images can be easily created from data obtained from various imaging modalities. As a result, 3D images are used not only for diagnosis but also for surgical simulation. Several simulation techniques can be used for cerebral aneurysm surgery, such as methods using 3D-CTA, 3D-DSA, and MR, and methods that involve the use of an expensive surgical simulator. The Dextroscope™ (Volume Interactions Pvt. Ltd., Singapore) is a surgical simulator with stereoscopic virtual reality using a hologram and is known to have high accuracy. There are a few reports on the use of the Dextroscope in cases of neurosurgical diseases, including intracranial aneurysm.  However, the Dextroscope is expensive, and therefore difficult to be introduced in many hospitals.
It is not feasible to use MR imaging immediately after the onset of aneurysmal subarachnoid hemorrhage because of body movement and difficulty in dealing with the rerupture of the aneurysm while obtaining the images. The simulation of surgery for intracranial aneurysm using 3D-DSA has also been reported.  The 3D-angio system can be a powerful tool in neuroradiology, especially for surgical intervention of intracranial aneurysms. 3D-DSA provides vivid 3D descriptions of the structure of blood vessels and is extremely effective for the evaluation of intracranial aneurysms. However, 3D-DSA has the disadvantage that data for bony structures cannot be obtained. Thus, 3D-DSA is the only method by which simulation of craniotomy is difficult. Moreover, 3D-DSA is invasive, and, at present, is not used by many hospitals. Currently, 3D-CTA is thought to be used for intracranial aneurysm surgery simulation in most hospitals. To date, there are many reports on the use of 3D-CTA for surgical simulation of intracranial aneurysms. ,,,, In most reports, however, surgical simulation was performed using only data of the arterial system, and it is far from simulation with virtual reality. Meanwhile, studies that reported surgical simulation using data of the bony structures and venous system, apart from the artery system, are rare.  3D-CTA image data contain information that is essential for craniotomy, such as information about the venous system, skull, and soft tissues, besides the arterial system. To obtain detailed anatomical information, it is important to set optimal viewing conditions for each structure. Therefore, we adopted an image-processing technique of multilayer fusing of two or more images after optimizing the image segmented from the DICOM data of CTA for each structure.
Meanwhile, the volume rendering image processed from CTA data is only a 2D image when it is still, and it becomes a pseudo-3D image on rotation. Stereoscopic viewing from arbitrary directions should be more effective in realistic surgical simulation. To achieve more realistic surgical simulation, we devised a new method for stereoscopic viewing from arbitrary directions using the Stereo Movie Maker or QTVR technology. By our method, realistic stereoscopic viewing was realized promptly at a low cost.
Up to now, there have been a few reports that have used QTVR technology in medicine. Friedl et al. used QTVR as an educational tool for aortocoronary bypass grafting, and their prototypical implementation into a database-driven and internet-based educational system in heart surgery.  Balogh et al. acquired intraoperative images during neurosurgical procedures for later reconstruction with a QTVR stereoscopic image system. They are using this system to learn microsurgical techniques.  Another application of QTVR in medicine is as an educational tool to study anatomy. The present study is the first report describing the application of QTVR movie maker in clinical medicine.
For emergency operations with a high degree of difficulty, such as aneurysmal clipping, preoperative realistic imaging and accurate surgical simulation is critical for a successful surgical outcome. Newly devised arbitrary directional stereoscopic viewing of multilayer fusion images enabled detailed surgical planning with 3D virtual reality. The surgical strategy selected on the basis of preoperative surgical simulation proved to be the correct choice during the actual operation, and the operation was carried out without complications in all cases. Stereoscopic virtual realistic surgical simulation is helpful in designing minimally invasive intracranial procedures. Our imaging technique has useful applications in surgery for not only cerebral aneurysms but also other neurosurgical diseases.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]