In-vitro assessment of a registration protocol for image guided implant dentistry

In this study a computer aided navigation In this study a computer aided na vigation technique for accurate positioning of oral implants was assessed. An optical tracking system with specially designed tools for monitoring the position of surgical instruments relative to the patient was used to regis ter 5 partially or completely edentulous jaw models. Besides the accuracy o f the tracking system, the precision of localizing a specific position on 3 -dimensional preoperative imagery is governed by the registration algorithm which conveys the coordinate system of the preoperative computed tomograph y (CT) scan to the actual patient position. Two different point-to-point re gistration algorithms were compared their suitability for this application. The accuracy was determined separately for the localization error of the p osition measurement hardware (fiducial localization error FLE) and the erro r as reported by the registration algorithm (fiducial registration error FR E). The overall error of the navigation procedure was determined as the loc alization error of additional landmarks (steel spheres, 0.5 mm diameter) af ter registration (target registration error - TRE). Images of the jaw model s were obtained using a high resolution CT scan (1.5 mm slice thickness, 1 mm table feed, incremental scanning, 120 kV, 150 mAs, 512 . 512 matrix, FOV 120 mm). The accuracy of the position measurement probes was 0.69+/-0.15 m m (FLE). Using 3 implanted fiducial markers, FRE was 0.71+/-0.12 mm on aver age ano 1.00+/-0.13 mm maximum. TRE was found to be 1.23+/-0.28 mm average and 1.87+/-0.47 mm maximum. increasing the number of fiducial markers to a total of 5 did not significantly improve precision. Furthermore it was foun d that a registration algorithm based on solving an eigenvalue problem is t he superior approach for point-to-point matching in terms of mathematical s tability. The was 0.71+/-0.12 mm on average and 1.00+/-0.13 mm maximum. TRE was found to be experimental results indicate that positioning accuracy of oral implants may benefit computer aided intraoperative navigation. The ac curacy achieved compares well to the resolution of the CT scan used. Furthe r development of point-to-point/point-to-surFace registration methods and t racking hardware has the potential to improve the precision of the method e ven further. Our system has potential to reduce the intraoperative risk of causing damage to critical anatomic structures, to minimize the efforts in prosthetic modelling, and to simplify the task of transferring preoperative planning data precisely to the operating room in general.