Three knee OA features can predict TKR

A model utilising three features of knee osteoarthritis can accurately predict which patients will progress to total knee replacement (TKR) and the time to TKR, its developers say.

A Canadian study used machine learning and data from the Osteoarthritis Initiative (OAI) cohort to develop a prediction model with high accuracy (concordance index of 0.85, Brier score of 0.02, and AUC of 0.86).

The data comprised baseline features from 7,589 knees of which 413 progressed to TKR after a mean follow-up of 108 months.

The top ten features selected for TKR prediction came from imaging (radiograph and MRI), medical history, 400m walk performance, use of intra-articular medication, and knee symptoms.

In the final model, a combination of Kellgren/Lawrence (K/L) grade on radiograph, bone marrow lesions (BML) in the medial condyle on MRI, and the presence of knee-related symptoms demonstrated the most significant impact in prediction time to a TKR event.

The researchers from the University of Montreal and Laval University Hospital Research Centre, Quebec, said the K/L grade and BML findings were not unexpected.

“Moreover, BMLs in the medial condyle as an indication of the likelihood of a TKR event is well in line with findings showing that this is the area in knee OA where both BMLs and cartilage degeneration are the most frequently affected,” they wrote in Arthritis Care & Research.

“There are medications that have been shown to prevent or reduce the severity of BMLs, but to the best of our knowledge, there is only one study showing that a bone remodelling therapeutic agent, bisphosphonates, was associated with about a 25% lower risk of TKR.”

The study authors acknowledged that MRIs may not be customary investigations in clinical practice yet but would become more routine as costs came down and access improved.

They said their model was the first to “predict with high accuracy if a given osteoarthritic knee would require TKR, when a TKR would be required, and who would likely progress fast toward this event.”

However validation in another cohort was required before it could be adapted into a useful tool to guide clinical decision-making, the said.

Already a member?

Login to keep reading.

Email me a login link

© 2022 the limbic