The verification of the alignment of the lower limb is critical for reconstructive surgery as well as trauma surgery in order to prevent osteoarthritis. The mechanical axis is a straight line defined by the center of the femoral head and the center of the ankle joint, ideally passing the knee joint in its center.

Whereas the usual preoperative method to determine the mechanical axis of the lower limbs is still the long standing radiograph, common intra-operative methods are the use of an electrocautery cord or an X-ray grid consisting of wire lines underneath the patient. Both methods require the surgeon to bring the femoral head and the ankle joint exactly to overlay with a radiopaque line that passes through both points. The distance of the knee center from this line is defined as the mechanical axis deviation (MAD). In order to reduce the errors introduced by perspective projection effects, the joint centers must be placed in the center of the c-arm images, which definitely requires time, experience and additional radiation.

We propose a computer aided X-ray stitching method that puts individual X-ray images into a panoramic image frame combining the Camera Augmented Mobile C-arm (CamC) system, which features a video camera with its optical center virtually coinciding with the origin of the X-rays, with an optical tracking marker pattern underneath the operating table. The camera image of the marker pattern is used to perform pose estimation of the C-arm, allowing the calculation of the x-ray source motion between the positions in which the individual X-rays were taken. By estimating the homography, the different X-rays can be registered into a panoramic frame, enabling perfect alignment and metric measurements.

In order to reduce parallax effects that lead to axis and metric measurement errors, we applied a method requiring two constraints: The bone plane has to be roughly parallel to the planar marker pattern and the distance between the marker plane and the bone plane has to be estimated.

In order to evaluate the method, we used a life-size synthetic skeleton leg. After tightening a straight wire between the centers of the hip and ankle joint, the knee joint was bent into a MAD of 55 mm, which was confirmed by measuring the distance between the knee center and the wire with a ruler. The leg phantom was then placed on a radiolucent operating table, parallel to the pattern plane 130 mm underneath. The operating table was moved through the C-arm while acquiring the three desired X-ray images. which were registered into a panoramic image frame. The centers of the femoral head, the ankle, and the knee were manually determined on the generated panoramic image by a surgeon. The mechanical axis was automatically displayed and the MAD was visualised in the image and computed as 55.23 mm.

We presented a new solution to intra-operatively verify alignment of the lower extremity. When using the CamC system, only a marker pattern has to be used for tracking. No additional tracking devices and calibration procedures are needed. Furthermore, the presented method only requires three x-rays that cover the femoral head, the knee and the ankle and marking of the three spots. Due to the parallax correction, these spots do not have to be exactly in the center of the picture. For this reason, compared to using an X-ray grid or an electrocautery cord, our method allows the procedure to be much faster and reduces the number of x-ray images. However, for clinical evaluation, a patient study will be conducted in the future.