Date of Award

January 2025

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)

Department

Medicine

First Advisor

John P. Fulkerson

Abstract

The relationship between anatomic risk factors for patellar instability and patellofemoral cartilage loss in the general population remains unclear. This study aimed to determine whether common measures of patellar instability are associated with lateral patellofemoral cartilage loss. A secondary objective was to evaluate the inter- and intra-rater reliability of 21 anatomical landmarks, initially placed using an artificial intelligence algorithm and subsequently verified with human input. We analyzed 389 non-weight-bearing computed tomography scans of knee joints from participants enrolled in the Multicenter Osteoarthritis Study. Baseline and 2-year follow-up knee magnetic resonance images were scored for patellofemoral cartilage morphology using the magnetic resonance imaging osteoarthritis knee score. The tibial tubercle–trochlear groove distance, sagittal tibial tubercle-trochlear groove distance, patellar tilt angle, tibiofemoral rotation, patellar height, entry point to trochlear groove angle, and entry point transition point angle were measured for each knee. Inter- and intra-rater reliability of three-dimensional landmark placements was assessed in 30 knees. We conducted a multivariable logistic regression analysis informed by univariate results and correlation structures. Generalized estimating equations logistic regression models with continuous, quintile, and natural cubic splines were fitted for each variable. All 19 semi-automated landmarks demonstrated excellent inter-rater reliability (ICC ≥ 0.87), with tibial and femoral shaft centers achieving perfect agreement (ICC = 1). Analysis of lateral patellofemoral osteoarthritis progression (n = 389 knees, mean age 53.79 ± 5.51 years, body mass index 28.48 ± 5.13 kg/m²) identified significant associations between specific 3D metrics and disease progression. The tibial tubercle-trochlear groove distance model (estimate: 0.062, 95% CI: 0.004–0.120, P = 0.036), tibiofemoral rotation model (estimate: 0.049, 95% CI: 0.0004–0.098, P = 0.048), and entry point-transition point spline model (estimate: 0.0014, 95% CI: 0.0001–0.0027, P = 0.036) all demonstrated significant positive associations. These findings highlight the reliability of semi-automated landmark placement and their potential role in advancing the understanding of patellofemoral arthritis. All 19 semi-automated landmarks demonstrated excellent inter-rater reliability and were successful in calculating three-dimensional metrics for the subsequent clinical study. The results also demonstrate significant associations between 3D anatomical metrics and lateral patellofemoral osteoarthritis progression. Tibial tubercle-trochlear groove distance, tibiofemoral rotation, and entry point-transition point angles show promise as valuable measures for understanding and tracking disease progression.

Comments

This is an Open Access Thesis.

Download
Open Access

This Article is Open Access

Share

COinS