Computer-Assisted Total Knee Arthroplasty Accuracy Influenced by Anatomic Landmark Identification

Nearly 1 million knee replacement surgeries are performed in the United States of America yearly. Robotic systems designed for knee arthroplasty have demonstrated comparable accuracy to  conventional methodologies lacking computer-assisted technology. Robot-surgical systems can be image-free or image-based. The two types yield similar surgical outcomes; however, image-based systems need potentially expensive medical imaging. 
 

Image-free robotic systems perform resections in a coordinate system formed by surgeon-probed anatomical landmarks across the femur and tibia. Resection accuracy is essential for improving the fitting of a replacement knee, which can yield greater patient satisfaction and implant longevity. This study aimed to quantify the influence of anatomical landmark identification on a robot’s ability to define the coordinate system in which bone resections are executed.
 

For the study, total knee arthroplasty was performed on a cohort of 40 cadaveric specimens using an image-free surgical robot. Data collected during the surgeries included the positions of anatomical landmarks necessary for defining a coordinate system. The difference between the landmark positions recorded by the surgical system and the ground truth positions was computed for each specimen. The standard deviation of the data was then determined, and error was randomly applied to each ground truth landmark through a Monte Carlo simulation. New coordinate systems were defined using the simulated data and overlayed with the ground truth coordinate system to quantify alignment error. Coordinate system error was then plotted against landmark error.
 

Different landmarks influenced different axes of a robot’s coordinate system. However, two landmarks on the bottom part of the femur induced the most error (greater than one degree) in the coordinate system per mm of landmark position error.
 

The study’s results can help develop novel surgeon training techniques and probing tools to improve the accuracy of image-free robotic surgical systems used in knee replacement surgery.