机器人辅助全膝关节置换术中的股骨基轴线设计

Determination of Femoral Anatomic Axis in Robot-Assisted Surgery of Total Knee Replacement

在机器人辅助全膝关节置换术中,股骨基轴线的确定有着至关重要的作用。在传统人工全膝关节置换术中,这条基轴线主要依靠在股骨髓腔内插入髓内杆来获得。而在基于CT建模的机器人辅助全膝关节置换术中,在术前的模型中就必须首先进行股骨基轴线的确定。但是在术前模型中由于获得的股骨段的长度不一样,除受噪声点干扰外,股骨本身的上段切片以及下段髁部切片对基轴线的确定也有很大的影响。传统上,采用最小二乘法进行股骨基轴线的拟合。但是直接使用最小二乘法,将很容易受到噪声点的干扰。一个很小的干扰点就足以把回归直线拉离正确的位置。为了避免这些干扰,我们充分利用了最小中值二乘法鲁棒回归的特点进行基轴线的拟合,避免了最小二乘法的缺陷。另外为了自动寻找到股骨基轴线的最佳拟合直线,文中采用了遗传算法来寻找最优解。在最后的实验中,对真实的股骨数据分别用不同的方法进行拟合,获得了很好的拟合效果。

Determination of femoral anatomic axis plays an important role in robot-assisted surgery of total knee replacement. In traditional total knee replacement surgery, the axis is obtained by inserting a rod into femoral lumen. However, in the robot-assisted total knee replacement based on CT model, the femoral anatomic axis must be determined preoperatively. Because the lengths in femurs are quite different in different patients, besides noises, the upper segment and lower segment of the femur influence the design of the axis greatly. Traditionally, the femoral anatomic axis is obtained by using the least-squares method directly. However, this method is easily disturbed by noise. To avoid the noise disturbance of CT data, the least median of squares method is used to fit the femoral anatomic axis for its characteristic of robust regression. The least median squares method eliminates the disadvantage brought by the method of least squares. In finding the best-fit line, genetic algorithms are used in the paper. In our experiment, we use the proposed method to fit the femoral anatomic axis and obtain an excellent result.