Patients whose legs are naturally curved are responding well to a novel form of knee-replacement surgery that uses AI to find the best alignment of the new components for their body.

The research involved patients randomized to receive either conventional knee-replacement surgery or the new ‘functionally aligned’ knee surgery.

“The results are positive for patients whose legs are naturally bowed,” says lead author Associate Professor Simon Young of Waipapa Taumata Rau, University of Auckland. See Journal of Arthroplasty.

Young is also an orthopedic surgeon at Health New Zealand Te Whatu Ora Waitematā. He operates at the Elective Surgery Center, which is part of Auckland’s North Shore hospital campus, where they started using a robot for knee operations in 2017, allowing greater accuracy.

Traditionally, knee-replacement surgery is not universally successful, with around one in five patients not fully satisfied with their knee following the procedure.

There are a number of factors that may contribute, but one reason could be because they had naturally bowed legs.

Young saw the potential to use the surgical robot to compare imaging of the patient’s knee with thousands of possible options for placement of the replacement knee, allowing the surgeon to select the best match for that person’s natural knee alignment.

Conventional surgery puts the knee on straight, whereas the patient’s knee may have never been perfectly aligned with the leg bones to begin with.

“We developed an artificial intelligence (AI) algorithm to use with the robot to help surgeons position the knee components in a way that better matches each patient’s natural anatomy,” says Young, an associate professor in the University’s Department of Orthopedics.

Around 30% of the population have bowed legs either naturally or because of conditions such as arthritis, which cause the knees to wear out, Young says.

“If you look around on a football field, you will see a lot of people who are young, fit, and healthy, who naturally have bowed legs. If, when they get older, they get arthritis and you put the new knee in straight, it will be in a position it has never been in their lives.

“For these people, usual knee replacements that assume the leg is straight may not work as well as ones that are functionally aligned.”

The new AI software takes the patient’s knee alignment and surrounding soft tissue and allows the surgeon to test different options—digitally.

“When we are in the operating room, we’re virtually positioning the components, then we consider that patient’s native alignment, and also their soft tissue tension. The computer model goes through and analyzes the 20,000–25,000 potential positions and ranks them according to what would be optimal for that patient. We then choose what we think is the best option.”

For the study, Young and colleagues randomized 244 knee-replacement patients to traditional or AI-assisted alignment and then followed the patients for two years, assessing their recovery with X-rays and questionnaires.

Overall, both groups had good outcomes and were happy with their knee replacements. However, patients who naturally had more bowed legs reported better results with the newer functionally aligned knees.

Based on the study, Young would recommend surgeons consider a patient’s natural leg shape when planning knee replacement surgery.

Young has developed an app for use by orthopedic surgeons wishing to use functional alignment in New Zealand, Australia, and Asia.

The research is ongoing to refine and expand the tool to ensure the best possible outcomes for all patients, Young says.