Model-based roentgen stereophotogrammetric analysis of orthopaedic implants
Model-based RSA Software
- Prof. Hans J.H.C. Reiber MSc. PhD (div. Image Processing, dept. Radiology)
- Edward R. Valstar MSc. PhD
- Berend C. Stoel PhD (div. Image Processing, dept. Radiology)
- Prof Rob G.H.H. Nelissen MD, PhD
- Emeritus Prof. Piet M. Rozing MD, PhD
- Bart L. Kaptein MSc. PhD
- Paul W. de Bruin MSc. PhD
December 1 2001 - June 1 2006
Technology Foundation STW
Division of Image Processing, department of Radiology, LUMC
Medis specials b.v.
In today’s ageing population, degenerative diseases, such as arthrosis, and rheumatoid arthritis will occur more frequently. In the treatment of these diseases by implantation of joint prostheses, adequate bone-prosthesis fixation is of greatest importance. The quality of this fixation can be assessed very accurately from two simultaneous X-ray exposures of the joint by a method called Roentgen Stereophotogrammetric Analysis (RSA). Because of its high accuracy, evaluation studies on the fixation of prostheses can be performed in a relatively short follow-up period and in a small patient group. Therefore, exposure of large groups of patients to new, not fully validated, implants and fixation techniques (e.g. new coatings and cements) can be prevented. Thereby, the number of revisions of prostheses, as a result of poor bone-prosthesis fixation, will be reduced resulting in a decrease of healthcare expenses.
In order to calculate changes in the position and orientation of a prosthesis in these RSA studies, metallic markers are attached to the surface of the prosthesis. These markers can then be identified in the two radiographs. Since attaching markers to the prosthesis is a difficult, time-consuming and expensive procedure for the manufacturer, RSA has only been applied in a limited number of clinical studies. To make clinical RSA studies more feasible and to make RSA more widely used, an alternative RSA method should be developed, that does not require the attachment of markers to the prosthesis.
We have implemented a very first prototype of such an alternative RSA method that allows the calculation of the position and orientation of a prosthesis from only the contours of the prosthesis - as they appear in the radiographs, without markers attached to these prostheses - by comparing these contours with those obtained from a simulated projection of a three-dimensional CAD model. Due to the casting and polishing process of the prosthesis during the production process, however, the actual implant may deviate from the CAD model. As a result, we found that our model-based RSA method is less accurate than the conventional RSA method based on the prostheses with markers.
In order to increase the accuracy of the model-based RSA method, so that it can be applied in clinical research, the matching algorithms should be able to cope with contours, which are partly unreliable. Therefore, new innovative matching algorithms and contour detection algorithms will have to be developed and implemented. These new techniques will be integrated in the RSA software package that has been developed in our institute to facilitate future validation studies.
As it is essential that improvements in model-based RSA are disseminated and accepted as feasible alternatives for conventional RSA, the developed methods will be extensively validated and evaluated in the final two years of the project.
The goal of the proposed project, therefore, is to develop an RSA method with which the bone-prosthesis fixation can be studied for any standard prosthesis without any modification in its manufacturing process. This will improve the applicability of the RSA method for clinical studies. By using RSA as a validation tool, poorly designed implants and poor fixation techniques will be prevented from being marketed, resulting in an improvement of quality of life of orthopaedic patients and a reduction in healthcare expenses.