|Home | About | Journals | Submit | Contact Us | Français|
Computer–aided surgery (CAS) based on high–resolution imaging techniques represents an important adjunct to precise intraoperative orientation when anatomical landmarks are distorted or missing. Several commercial systems, mostly based on optical or electromagnetic navigation principles, are on the market. This study investigated the application of EasyGuide®, VectorVision®, and InstaTrak® CAS systems in ENT surgery under practical and laboratory conditions. System accuracy, time required, handling, individual features, and practicality were examined in 155 patients who underwent endonasal sinus surgery and in 23 patients who underwent anterior or lateral skull base procedures. The VectorVision® and InstaTrak® CAS systems proved to be suitable for routine application in surgery involving the paranasal sinuses and various regions of the anterior skull base by helping to avoid critical structures and to determine minimally invasive approaches.
Endonasal sinus surgery as well as surgery to various regions of the anterior skull base are challenging because of the presence of a variety of endangered vascular and neural structures in a very confined space. Due to previous surgical procedures and the destructive nature of some of the diseases affecting the skull base, surgical landmarks are in some cases distorted, thus increasing the risk of immediate intraoperative complications and long–term postoperative defects.1, 2, 3, 4 In such cases, computer–assisted surgery (CAS) is an important tool for increasing the precision of orientation during surgery, which could become mandatory for specific cases. Different groups have already applied techniques of CAS to neurosurgery and ear–nose–throat (ENT) surgery and gained first encouraging experiences.5, 6, 7, 8, 9, 10, 11, 12 Even though different commercially available CAS systems are based on different navigation principles (e.g., optical or electromagnetic), precise intraoperative orientation is provided by high–resolution imaging techniques (e.g., computed tomography (CT), magnetic resonance imaging (MRI)). Current indications for CAS in ENT are endoscopic or microscopic revision endonasal sinus surgery; the surgeon's experience; and surgery of the anterior and middle cranial fossae, infratemporal fossa, and retromaxillary space. In otorhinolaryngology, both electromagnetic and optical systems have been found to be suitable for the demands of intraoperative navigation.12, 13, 14, 15 Both EasyGuide® and VectorVision® image–guided surgery systems, which use an active or passive optic navigation technique, were initially designed for neurosurgery. In contrast, InstaTrak (Vti)®, which uses electromagnetic technology, was created specifically for endonasal sinus surgery. This prospective study compared these three CAS systems in both clinical and laboratory conditions.
The following commercially available CAS systems were used in the study: EasyGuide® (version 1.1, Philips, Eindhoven, The Netherlands); VectorVision® (version 3.56, BrainLAB, Munich, Germany), and InstaTrak® (version 2.4, Vti®, Woburn, MA). Based on either active or passive optic or electromagnetic navigational principles, these systems are available commercially. Although these systems use different navigational principles all require general preoperative steps: the placement of fiducial markers, CT or MRI imaging, data transfer to the operating room (OR), and identification and registration of the fiducials. Registration is the correlation between the position of the instrument in the surgical field and the corresponding location on the CT or MRI images. Orthogonal images, obtained by multiplanar reformatting of axial CT images, provide detailed anatomic views of the anatomy involved. When performing these steps, each platform has individual needs and a number of potential deficiencies influencing the accuracy of the system.
The EasyGuide® system is based on an active optical navigational principle. It consists of three main parts: a mobile workstation, a position digitizer, and a straight pointer equipped with infrared light emitting diodes (IREDs). The DICOM 3.0–protocol allowed us to use the source images acquired by a Picker® Helical CT scanner. During the period of application, in cooperation with Philips, we developed a goggle equipped with three additional LEDs that allows movement of the patient's head during sinus surgery with constant tracking by the optical position digitizer. Image–guided surgery with EasyGuide® was performed in 73 cases of microscopic endonasal sinus surgery.
In contrast to the active optical tracking technique, VectorVision® utilizes a freehand probe, tracked by a passive–marker sensor system. It has a mobile workstation and two cameras that carry IREDs and that register reflected infrared light from the marker spheres. Moreover, these marker spheres can be mounted on a special bracket and therefore are easily adapted to an already existing instrument in the OR. VectorVision® was used in 23 cases of surgery to the anterior and middle cranial fossae.
The InstaTrak® CAS system was used in 82 cases of microscopic endonasal sinus surgery. This device employs electromagnetic tracking technology attached to a standard suction. Two electromagnetic sensors provide positional information on the CT images. One sensor is incorporated into the standard suction; the second sensor is located on a headset worn by the patient during the surgical procedure. The headset compensates for head movement and is a means of obtaining automatic registration.
System accuracy, time needed for preparation, imaging and system set–up, applicability to requirements of ENT surgery, and potential complications related to the use of a CAS system of all three CAS devices were evaluated. System accuracy was analyzed for both laboratory and intraoperative conditions, and preoperative axial CT scans were obtained using a Picker® PQ 5000 CT scanner. The image matrix was 512 × 512 using a slice thickness of 3 mm and a reconstruction increment of 1 mm. A spiral standard algorithm was used and the gantry tilt was set to 0 degrees.
The only available methods to describe intraoperative or laboratory accuracy are subjective measurement techniques. However, the CAS systems investigated in this study provide the user with a nearest marker function (NMF). The NMF represents the distance between the actual position of the probe to the stored position of the next fiducial in the images. Image–slice thickness and number, spread of anatomic or artificial fiducials, and root mean square error (RMSE) influence the result of NMF.
To evaluate accuracy in the laboratory, 10 fiducial markers were applied to a perspex phantom at predefined geometric coordinates (Fig. 1). An additional five markers for the optic system were used in conjunction with a headset for the electromagnetic system, and a standardized protocol was performed. After an axial CT of the model was acquired, it was moved to four specific positions using a remote control–operated platform to simulate possible head movements during a standard sinus procedure. The simulated movements were as follows: 10–cm shift to the left, 20–cm elevation of the platform, 30–degree rotation to the right, and 90–degree rotation to the left. Each marker was registered by its x–, y–, and z–coordinates. Each passage included 50 values and was performed five times (250 values total). The coordinates of the markers at the model were predefined, and three–dimensional (3D) geometric transformation formulae were used to compare the actual coordinates with those measured by the system.
To ensure correct functioning of the CAS systems during clinical applications, we established a standardized work flow. The fiducial markers or headsets were fitted to the head of the patients immediately before the CT or MRI images were obtained. The resulting images were transferred directly to the OR from the CT or MRI scanner via a digital network using the DICOM 3.0 protocol. The same person placed the markers and performed the registration procedure to reduce potential variations that could affect intraoperative accuracy.
To gain reliable values describing intraoperative accuracy of the various systems, we visually checked bony landmarks and selected surface points near the region of the paranasal sinuses and the anterior skull base three times for each case utilizing the NMF (Table 1).
As the overall deviation for the registration set and nearest marker function, RMSE was used as the standard for accuracy for the EasyGuide® and VectorVision® CAS systems.
Statistical analyses with a nonparametric test (Wilcoxon Matched–Pairs Signed Rank test) were performed using SPSS® statistical software (SPSS Inc., Munich, Germany). A p value less than 0.05 was regarded as statistically significant.
During laboratory conditions and in clinical applications, the accuracy of the active and passive optic systems differed considerably. In contrast, the electromagnetic system InstaTrak® provided similar precision during both experimental and clinical situations (Table 2).
The CAS systems required more time for patient preparation, image aquisition, and system set up. Usually, additional manpower is also needed. This new technique was not integrated into daily routine without additional effort (Table 3).
During the study, use of the CAS systems was associated with no intraoperative complications. Handling the CAS systems also was associated with no major problems. After the staff were trained, they found it simple to use. The following examples demonstrate the applicability of the three CAS systems to specific requirements of ENT surgery.
A 41–year–old white female with recurrent maxillary, sphenoidal, and frontal sinusitis underwent endonasal microscopic revision sinus surgery. The EasyGuide® CAS system was connected to an IRED–equipped pointer and to a Blakesley forceps. These instruments allowed identification of various landmarks, such as the skull base (Fig. 2), during the procedure.
A 46–year–old white male suffering from a meningioma of the left infratemporal fossa, orbit, sphenoid sinus, and nasal cavity underwent revision surgery via an infratemporal fossa approach. The VectorVision® CAS system was used to remove as much of the tumor as possible by following the floor of the middle cranial fossa medial to the carotid artery, into the sphenoid sinus and nasal cavity (Fig. 3). To achieve a reliable registration, the patient had to be immobilized in a Mayfield clamp. This case illustrates some of the differences between laboratory and clinical accuracy such as rough representation of external anatomical landmarks and a certain degree of soft–tissue shift.
A 68–year–old woman suffering from frontal sinusitis with a mucocutaneous fistula had undergone a previous attempt at obliteration in her home country. The InstaTrak® system was used to determine the optimal site for entering the frontal sinus during endonasal microscopic sinus revision (Fig. 4), which included drainage of the frontal sinus and closure of the fistula. This CAS system provided highly accurate instrument localization, allowing the surgeon to navigate close to critical structures like the orbit.
Surgery of the paranasal sinuses and of the anterior and lateral skull base requires detailed anatomical knowledge and a good 3D “imagination” of the individual case based on preoperative imaging studies. During surgery this knowledge must match microscopic or endoscopic findings and a surgeon's tactile findings. Most intraoperative complications happen when surgeons fail to understand a given pathology sufficiently and therefore are unaware of the exact position of the surgical instruments.
To overcome this problem associated with intraoperative orientation, preoperatively generated image data sets (e.g., CT or MRI) have been combined with sophisticated 3D coordinate–measuring devices and modern computer graphics. These techniques provide surgeons an almost unhindered movement of instruments in the surgical field while providing precise, real–time spatial registration of a patient's anatomy through the preoperative image set. In otorhinolarygology the term computer–aided–surgery has been widely accepted for image–guided navigation systems based on this technology.16 Optic and electromagnetic technologies have been found to have the widest applicability compared to other tracking principles, such as sonic or electromechanical systems developed for CAS.
Compared to electromechanic digitizers, optical digitizers offer more flexibility and are easier to use.17 The function of an optical digitizer is based on IREDs placed on the pointer device or digitizer itself (VectorVision®). A camera array fixed near the patient's head detects the position of the IREDs or registers the reflected infrared light from the marker spheres. To function correctly, none of the fiducial markers affixed to the skin should be removed from their initial position. This requirement was a major reason for the loss of accuracy in a number of cases. Furthermore, the line of sight of both cameras must be unimpeded. Otherwise, these cameras cannot detect the actual position of a pointer or instruments within the surgical field. The use of a microscope during endonasal sinus surgery requires sophisticated positioning of the IRED cameras to avoid intraoperative failure of the CAS system.
Electromagnetic digitizers superimpose a magnetic field around the surgical field. For this purpose, the tracking system uses a magnetic field–generating source, a magnetic field–detecting sensor, and processing software. Ferromagnetic materials and electromagnetic fields others than the field of the CAS system can cause such a CAS system to malfunction. Therefore, interferences with the system's electromagnetic field must be reduced as much as possible, for example, by using instruments free of ferromagnetic materials (e.g., titanium).
Both EasyGuide® (version 1.1) and VectorVision® (version 3.56) systems were developed and designed for application in neurosurgery. Therefore the use of EasyGuide® required a number of modifications to suit the demands of endonasal sinus surgery to insure the most suitable access to the region of interest in conjunction with the greatest possible accuracy. These described prototypical modifications have not yet been implemented in the systems. The attempt to integrate EasyGuide® into daily sinus surgery routine began with the development of the goggle as described previously. Our experiences hold promise because this goggle frees the head from rigid fixation. Further improvements are pending.
The VectorVision® (version 3.56) CAS was suitable for surgery involving the vulnerable region of the skull base. The overall intraoperative accuracy was less than 2 mm. The difference in the values achieved during laboratory conditions reflected factors like lost fiducial markers, approximations of external landmarks, dislocation of the reference marker array in respect to the Mayfield clamp, and shift of the soft tissue. With VectorVision®, any surgical instrument can be used as a measuring device by simply attaching a passive marker array with an adjustable bracket to the instrument. The amount of time needed for patient preparation, imaging, and system set up is acceptable. Except for an additional CT scan with fiducial markers in place, there is no additional burden to the patient. Software and hardware errors like program failures or camera malfunction were very limited, but occurred.18, 19
The InstaTrak® (version 2.4) is a CAS system primarily developed and designed for image–guided endoscopic and microscopic sinus surgery. The fiducial marker method requires placing skin markers for the aquisition of a CT scan. However, we found the headset method to be much easier to use and more reliable because dislocation of markers is ruled out from the beginning. The registration procedure is performed by the system itself immediately after the image data are transferred to the workstation, reducing the chances of human error and the need for additional time. Therefore the total time necessary to operate this device is much less compared to other systems. Only one patient claimed discomfort from wearing the headset, which caused him some pain in the glabellar region. Intraoperatively, accuracy decreased the further the surgeon proceeded toward the sphenoidal sinus. This finding mostly reflects that the magnetic field provided by the sensors is fixed at the headset and weakens at a distance of 8 cm or more from the headset.
The design of straight and long aspirators is appropriate for paranasal sinus surgery.
To make fixation of fiducial markers on the patient's skin unnecessary, BrainLAB has developed a new registration procedure that will be performed at a workstation immediately after the image data are transferred. This new feature will reduce the interval between image aquisition and surgical procedure to a minimum (VectorVision® compact, version 4.0).
The electromagnetic system InstaTrak® (version 2.4) is appropriate for paranasal sinus surgery whereas the present configuration of the optical system VectorVision® (version 3.56) is useful for surgery involving the vulnerable regions near the skull base. Overall, patients in both standard and complicated cases must spend 10 to 25 minutes more in the OR. Undoubtedly, however, CAS significantly improves a patient's safety by helping the surgeon to avoid critical structures and to determine minimally invasive approaches.