Machine Learning to Predict Osteoarthritis

We developed a machine learning-based prediction model for incident radiographic osteoarthritis (OA) of the knee over eight years using MRI-based cartilage biochemical composition and knee joint structure, demographics, and clinical predictors, including muscle strength and symptoms. We found that a 10-predictor model including MRI parameters coupled with demographics, symptoms, muscle, and physical activity scores provided good prediction of incident radiographic OA over eight years.

Flowchart showing a machine learning framework for predicting knee osteoarthritis development using Osteoarthritis Initiative data, starting with 112 predictors across demographics, symptoms, physical activity, muscle, alignment, medication, and MRI measures, and narrowing to a top 10 clinically informed predictors to model incident radiographic osteoarthritis over eight years.