A Machine Learning Model to Predict Knee Osteoarthritis Cartilage Volume Changes over Time Using Baseline Bone...

Area of Science:

• Orthopedics
• Radiology
• Biomedical Engineering

Background:

• Osteoarthritis (OA) is a leading cause of musculoskeletal disability, characterized by cartilage degradation.
• Early identification of patients at risk for progressive knee OA is crucial for timely intervention.

Purpose of the Study:

• To evaluate if baseline knee bone curvature (BC) can predict one-year cartilage volume loss (CVL) in osteoarthritis (OA) patients.
• To develop and validate gender-specific machine learning (ML) models for predicting CVL.

Main Methods:

• Utilized magnetic resonance imaging (MRI) data from 1246 participants to assess baseline BC and cartilage volume.
• Employed five machine learning algorithms to predict one-year CVL across 12 knee regions.
• Included age, body mass index, and eight BC regions as predictive variables.

Main Results:

• Machine learning models demonstrated high predictive accuracy (R ≥ 0.78) for one-year CVL in both genders.
• The model showed excellent performance across most knee regions, with a noted exception for the medial tibial plateau in women.
• Successfully identified five baseline BC regions as key predictors of future cartilage loss.

Conclusions:

• Baseline knee bone curvature is a significant predictor of one-year cartilage volume loss in OA.
• Developed accurate, gender-based ML models for predicting progressive knee OA, offering a novel tool for risk stratification.
• This predictive capability can significantly benefit patients identified as being at risk for structural progressive knee OA.