Wear patterns in knee OA correlate with native limb geometry.

Area of Science:

• Biomedical Engineering
• Orthopedics
• Radiology

Background:

• Current knee osteoarthritis (OA) grading uses discrete scoring systems on predefined regions.
• Anatomically standardized cartilage thickness assessment is crucial for robust statistical comparisons.
• This study addresses the need for standardized analysis to understand OA risk factors.

Purpose of the Study:

• To develop a method for anatomically standardized cartilage thickness assessment.
• To compare morphotypes of healthy individuals and virtual reconstructions of OA patients.
• To assess shape-related knee OA risk and correlate phenotype with cartilage loss patterns.

Main Methods:

• Utilized a dataset of 798 end-stage knee OA cases.
• Computed joint space width and translated it into 2D pixel images.
• Employed a principal polynomial autoencoder for non-linear encoding of cartilage wear patterns and virtual healthy twin reconstruction.

Main Results:

• Identified 4 dominant wear patterns accounting for 94% of variance: medial (54.8%), bicompartmental (25.2%), and lateral (9.1%).
• Medial wear further subdivided into anteromedial (11.3%) and posteromedial (10.4%).
• Pre-diseased limb geometry predicted OA incidence with a positive predictive value of 0.80 (r=0.58, p<0.001).

Conclusions:

• Presented an innovative workflow to correlate cartilage wear with knee joint phenotype.
• Assessed distinct knee OA risk based on pre-diseased lower limb morphology.
• Confirmed the importance of alignment and joint geometry in knee OA onset and progression.