Paper CAR95 Berlin

Computer Assisted Radiology
9th International Symposium and Exhibition
Berlin, June 21-24, 1995

Session 4
Tracking and Guidance Systems
Saturday, June 24, 1995, 11:00
Venue: Universitaetsklinikum Charité, Schumannstr. 20/21, Berlin-Mitte, with
special support by Prof. D. Bernd Hamm, MD, Director Inst. f. Roentgendiagnostik

Session Chairs:
D.J.H. Hawkes, PhD, UMDS, Guy's Hospital London (UK)
F.A. Jolesz, MD, Brigham and Women's Hospital, Boston (USA)

Application of the ARTMA Image-Guided Navigation System to Endonasal Sinus
Surgery
A.R. Gunkel, W. Freysinger, W.F. Thumfart, Univ.Klinik f. HNO, Innsbruck; M.J.
Truppe, ARTMA Medizintechnik GmbH, Wien (A)
(link to CAR´95)

Application of the ARTMA image guided navigation system to endonasal sinus surgery

Gunkel A.R., Freysinger W., Thumfart W. F., Truppe M. J.*
HNO-Universitaetsklinik Innsbruck, Anichstrasse 35, A-6020 Innsbruck, Austria
* ARTMA Medizintechnik GmbH, Am Kanal 27, A-1110 Wien, Austria

Summary

We solve the problem of orientation in revision operations with the ARTMA Virtual Patient® System. It uses state of the art image processing techniques and a 3D digitizer to find the position of instruments. The realization of Interventional Video Tomography allows mutual correlation of a video photo, corresponding anatomical landmarks with CT and/or MRI images. Structures and access paths to the regions of interest can be defined and fused with the endoscopic video image in real time. We report our first experiences with this comfortable an highly reliable tool.

Introduction

Many surgical procedures in the field of ENT take place in close proximity to vital structures like the orbit, the skull base, the internal carotid artery etc. In our clinic we have decided to use computer-assisted navigation systems to reduce the risk of trauma to vital structures during endonasal endoscopic procedures. The ARTMA Virtual Patient System finds the position of a surgical instrument (endoscope) by permanently measuring magnetic fields. It correlates the position of the surgical instrument, ideally in the range of 1 mm, to CT- or MR-images and other possible data sources. An important prerequisite is that the use of Computer-Assisted-Surgery (CAS) does not extend the time needed for surgery significantly. However, the ARTMA System is one of several 3D-orientation systems that operate on different technologies [1,2,3]. Principally, the actual position of the surgical instrument is shown in the correlating CT- or the MR- images by exploiting the inherent spatial information of the CT-/MR- images. If properly used, CAS-systems can be a very helpful tool in the hand of an experienced surgeon but will never replace his expertise and knowledge.

Material and Method

We used the ARTMA 3D-navigation system mainly in patients having chronic polyposis of the paranasal sinuses. Preoperatively, an axial CT-scan of the head was performed extending from the forehead to the hard palate, especially showing the para-nasal sinuses. The helical scan (3 mm table increment) was reconstructed in 1 mm non overlapping slices, stored on a WORM optical disk and transferred to the navigation system. The ARTMA Virtual Patient [4] system uses these data without postprocessing. Additionally, it can handle images provided by other imaging modalities, such as MR, ultrasound, or X-ray, without modification. Orientation in space is achieved by use of a 3D-digitizer [5] which measures magnetic fields with Hall sensors to determine spatial position and orientation. An electromagnet close to the head of the patient establishes a magnetic field with known orientation and strength (see Fig. 1) and defines the reference coordinate system. The CT data of the patient are calibrated by marking a structure in a slice, holding a stylus against the corresponding anatomical structure of the patient and matching it to the appropriate point in two different video photographs of the patient's face. On this base the patient image transformation can be realized through the direct linear transformation [6]. The principle of Interventional Video Tomography (IVT) allows to correlate two dimensional images supplied by the endoscope to their correlate in the CT or MR slices. Advanced image processing algorithms and the spatial information of the imaging modalities yield an exact orientation in space even for the endoscopic view. Hall sensors attached to the head of the patient, the surgical instrument, and/or the endoscope trace every change of the magnetic field. Therefore, the position of the patient and the surgical instrument is known with respect to the reference magnet near to the patient's head, see Fig. 1. The actual position of the instrument inside the patient is determined by the computer and displayed on the screen. Consequently, the patient can be moved in space freely.

During endonasal endoscopic operations the surgeon watches the live endoscopic video on the computer monitor. He can permanently check every movement of his instrument (rigid endoscope) by looking at the video image and the corresponding CT slices displayed by the system at any time. The most sophisticated feature, though, is that the Virtual PatientTM System allows to mark certain regions on the screen (either on the CT scan or on the video photograph of the patient) and to create virtual trajectories of access to the surgical field, leading the surgeon to the region he is aiming at (Fig. 2). The region of interest and the defined path are superimposed to the live endoscopic video. The task of the surgeon is to follow a series of rectangles floating in space. If he leaves the planned and direct path the rectangles are distorted or drawn out of his endoscopic view. This indicates that he is leaving the trajectory of the best approach. The accuracy of the ARTMA System was assessed by visually judging the quality of the match between the actual anatomical location of the instrument and its calculated position as displayed on the video-image. The next release of the software will make measuring of distances much easier.

Results

For a well-trained operating team (surgeon, technician, and nurse) the time needed for calibration data did not increase the operating time significantly. Good calibration gave an accuracy of 1 - 2.5 mm depending on whether we used radio- opaque fiducial markers attached to the patient's head or not. We found that the accuracy of was better with markers and that it was of crucial importance which anatomical locations we chose for calibration. It is best to take bony landmarks (i. e. spina nasalis anterior, glabella, teeth, etc.) instead of soft structures on the skin (cheek, lobule of helix, etc.).
The Virtual Patient System requires that the magnetic field should not be influenced by large masses of metal near to the operating field. A typical operating room easily meets this condition. Optimal results were obtained by using a custom- built wooden headrest.
Intraoperatively, it was indeed possible to move the patient's head without changing calibration. We found that the equipment did somehow naturally integrate into the usual operative procedure and did not hinder the surgeon by the introduction of additional instruments, like measuring probes. This means that with the ARTMA system standard endonasal instruments serve both as instrument and as probe.
The most useful feature of the system, though, is that the surgeon can follow a preoperatively defined path to the target. e.g. tumor, skull base, lamina papyracea, like a pilot. He is guided by colored rectangular frames floating in space superimposed to the live video-endoscopic images. It is important to note that this path can lie anywhere in the patient's head. It is not restricted to lie in "direct sight" of the surgeon; this allows him to use, say, a 70_ optics of the endoscope and bent instruments. We actually could measure distances, mark structures, and work in any region inside the nose or the paranasal sinuses as if we were performing a normal surgical intervention.

Discussion

One of the major advantages of the ARTMA Virtual Patient System is that both the patient and the instruments can be moved freely relative to each other without disturbing the calibration. This permits the surgeon to operate just as normal, moving the patient's head as he deems necessary. This system opens up a wide range of application possibilities as other media can be used for correlation, i.e. MRI images, ultrasound images, etc.. We worked with a first prototype of the ARTMA Virtual PatientTM System; a second release of the software version is under development and will enhace comfort and performance of the system. Many more tests are necessary to improve this systems and introduce it into regular clinical routine. Nevertheless, this 3D-navigation tool is an excellent aid during endoscopic endonasal surgeries.

Conclusion

The ARTMA Virtual Patient System can be termed a 3D video navigation system.. The arrival of even more powerful computers and "intelligent" software will never replace the experience and the skill of a surgeon. It is always he, who has to take the ultimate decision about the next step of a surgical procedure.

References

1 A. Kurzeja, M. Wenzel, B. Korves, R. Mösges, Dekompression des Nervus opticus nach Frakturen des Riechschädels mit Hilfe von CAS (Computer-Assisted Surgery), Laryngo-Rhino-Otol., 73, 274-276, 1994.

2 K.T. Kavanagh, Applications of Image-Directed Robotics in Otolaryngo-logic Surgery, Laryngoscope, 104,283-293.

3 S. J. Zinreich, S. A. Tebo, D. M. Long, H.Brem, D. E. Mattox, M. E. Loury, C. A. Vander Kolk, W. M. Koch, D. W. Kennedy, R. N. Bryan, Frameless Stereotaxic Integration of CT Imaging Data: Accuracy and initial Applications, Radiology, 188, 735-742, 1993.

4 M. J. Truppe, Augmented Reality in Endoscopic Surgery, presented at Medicine Meets Virtual Reality III, San Diego, USA, January 19 - 22, 1995.

5 C. F. Hildebolt, and M. W. Vannier, Three-Dimensional Measurement Accuracy of Skull Surface Landmarks, Am. J. Phys. Anthropol., 76, 497-503, 1988.

6 Y. I. Abdel-Aziz, H. M. Kamara, Direct linear transformation into object space coordinates in close range photogrammetry. In Proceedings of the Symposium on Close-Range Photogrammetry, University of Illinois, pp 2 -28, January 1971.

Figure 1: Schematic drawing of the intraoperative system set-up. A PC (1) is connected to a monitor (2), an endoscope(3), and a 3D-Digitizer (4). The origin of the overall coordinate system is generated by (5). The orientation of the field lines is shown schematically. The relative position of the patient (6) and the endoscope (3) or the instrument is measured by Hall-probes (black cubes). The relevant coordinate systems are shown for the patient (xp, yp, zp) and the endoscope (xe, ye, ze).

Figure 2: Intraoperative photograph of an endoscopic image, some marked structures (full points), and the rectangles symbolizing the preoperatively defined trajectory of approach (path).

Computer Assisted Radiology 9th International Symposium and Exhibition Berlin, June 21-24, 1995