Comparison of the accuracy of invasive and noninvasive registration methods for image-guided oral implant surgery

• Published in: The International journal of oral and maxillofacial implants Vol. 25; no. 3; p. 491
• Main Authors: Widmann, Gerlig, Stoffner, Rudolf, Schullian, Peter, Widmann, Roland, Keiler, Martin, Zangerl, Antoniette, Puelacher, Wolfgang, Bale, Reto Josef
• Format: Journal Article
• Language: English
• Published: United States 01.05.2010
• Subjects: Dental Implantation, Endosseous - methods; Dental Models; Humans; Image Processing, Computer-Assisted; Imaging, Three-Dimensional - methods; Models, Anatomic; Pattern Recognition, Automated - methods; Reproducibility of Results; Surgery, Computer-Assisted - instrumentation; Surgery, Computer-Assisted - methods; Tomography, X-Ray Computed
• ISSN: 0882-2786

Registration refers to the linkage of coordinates from an image with actual patient coordinates and has been shown to be the most influential factor in the accuracy of image-guided surgery. Invasive bone markers are the gold standard but require surgical placement prior to imaging. In contrast, registration templates or external registration frames are noninvasive, but their repositioning during imaging and surgery can be a source of error. The purpose of the present study was to determine whether noninvasive registration methods can achieve an accuracy similar to that of invasive bone marker registration. Computed tomographic slices (1 mm each) of a maxillary and mandibular dental stone cast that had been prepared with target markers on the buccal and oral surfaces were registered with an optical-based navigation system simulating invasive bone markers, noninvasive registration templates, and a noninvasive external registration frame. Predicted error of the navigation system, fiducial registration error, and target registration error were evaluated. The use of five and seven registration markers was compared. A total of 696 error measurements was performed. The external registration frame showed significantly worse fiducial registration error compared with the other methods, but there was no significant difference in target registration error between invasive and noninvasive registration methods. The predicted error given by the navigation system significantly underestimated target registration error. Increasing the number of registration markers from five to seven resulted in no significant differences. Noninvasive registration based on registration templates or external registration frames showed accuracy that was equivalent to that of invasive registration. The use of five registration markers was sufficient. The predicted error given by a navigation system should not be mistaken as "navigation error" during clinical application.