RSNA 2003 Scientific Papers > Development of Three-dimensional Kinematic Analysis ...
  Scientific Papers
  SESSION: Physics (CAD IV: Skeletal, Liver)

Development of Three-dimensional Kinematic Analysis System for Total Knee Arthroplasty Using Single-Plane Fluoroscopy

  DATE: Tuesday, December 02 2003
  START TIME: 03:10 PM
  END TIME: 03:17 PM
  CODE: J18-834

Takaharu Yamazaki MS, RT
Suita Japan
Tetsu Watanabe MD
Yoshikazu Nakajima PhD
Kazuomi Sugamoto MD
Daisuke Maeda
Shinichi Tamura PhD

Computers, diagnostic aid
Fluoroscopy, technology
Knee, prostheses

Purpose: Quantitative assessment of three-dimensional (3-D) dynamic motion after total knee arthroplasty (TKA) allows objective evaluation of joint diseases and dysfunctions and provides useful information that can used to increase the limited life span of artificial knee joints. The purpose of this study is to develop a system for 3-D kinematic analysis of TKA implants using single-plane fluoroscopy.

Methods and Materials: Fluoroscopic image of knee implants was acquired as a series of digital images by 7.5 frame/sec serial shots. Since common fluoroscopic image occurs a geometric distortion due to the curvature of the screen in the image intensifiers, the image was corrected using a nonlinear distortion correction technique. In order to determine 3-D pose of knee implants using the single-plane fluoroscopic image, it is necessary to have an accurate geometrical model of knee implants and to know imaging geometry (X-ray parameters). A 3-D geometrical data of knee implant was acquired by computer assisted design (CAD) model. X-ray parameters were determined using a 217-maker 3-D calibration cube. A single-plane fluoroscopic image, together with this knowledge is sufficient for complete 3-D pose estimation of the model. In 3-D pose estimation of the model, an objective function was defined as the sum of Euclidean distance from point on the projection rays (corresponding to the point on the implant silhouette contours) to the closest point on the surface of implant CAD model. Therefore, the 3-D pose was estimated by minimizing the function using a nonlinear optimization technique. In order to validate the accuracy of this system, in vitro test was performed using images of knee implants taken in 10 different poses with respect to X-ray focus. The accuracy was assessed by comparing knee implants estimates with position measurements from the same knee implants obtained using an accurate 3-D digitizer. Finally, we performed 3-D kinematic analysis during dynamic motion of a TKA patient using this system. A sequence of 66 images was collected during walking.

Results: In the result of in vitro test, the estimation errors between knee implants (femoral and tibial components) in in-plane/out-of-plane translations and rotations were 0.68/3.61 mm and 0.59/0.49 deg, respectively. Result of clinical case was sufficient for carrying out 3-D kinematic analysis of TKA implants.

Conclusion: We have developed 3-D kinematic analysis system for TKA using single-plane fluoroscopy. This system could be useful for analyzing TKA 3-D kinematics during dynamic motions.