Inter- and intra-rater reliability of patellofemoral kinematic and contact area quantification by fast spin echo MRI and correlation with cartilage health by quantitative T1ρ MRI

Abstract

Background Patellar maltracking is a leading cause of patellofemoral pain syndrome (PFPS). The aim of this study was to determine the inter- and intra-rater reliability of a semi-automated program for magnetic resonance imaging (MRI) based patellofemoral kinematics. Methods Sixteen subjects (10 with PFPS [mean age 32.3; SD 5.2; eight females] and six controls without PFPS 19 [mean age 28.6; SD 2.8; three females]) participated in the study. One set of T2-weighted, fat-saturated fast spin-echo (FSE) MRIs were acquired from each subject in full extension and 30° of knee flexion. MRI including axial T1ρ relaxation time mapping sequences was also performed on each knee. Following image acquisitions, regions of interest for kinematic MRI, and patellar and trochlear cartilage were segmented and quantified with in-house designed spline- based MATLAB semi-automated software. Results Intraclass Correlations Coefficients (ICC) of calculated kinematic parameters were good to excellent, ICC > 0.8 in patellar flexion, rotation, tilt, and translation (anterior -posterior, medial -lateral, and superior -inferior), and contact area translation. Only patellar tilt in the flexed position and motion from extended to flexed state was significantly different between PFPS and control patients (p = 0.002 and p = 0.006, respectively). No significant correlations were identified between patellofemoral kinematics and contact area with T1ρ relaxation times. Conclusions A semi-automated, spline-based kinematic MRI technique for patellofemoral kinematic and contact area quantification is highly reproducible with the potential to help better understand the role of patellofemoral maltracking in PFPS and other knee disorders.

Patellar Malalignment Is Associated With Patellofemoral Lesions and Cartilage Relaxation Times After Hamstring Autograft Anterior Cruciate Ligament Reconstruction

“…These regions of interest were then used to estimate patellar alignment using a previously developed software program, with good intraclass correlation coefficient values (>0.84) for intra- and interuser reliability. 16 First, the patellar and femoral coordinate systems were defined anatomically as previously described. 16 The segmented femoral bone in the knee-flexed position was then rigidly registered to that in the fully extended position by the use of an iterative closest-point shape-matching algorithm.…”

“…16 First, the patellar and femoral coordinate systems were defined anatomically as previously described. 16 The segmented femoral bone in the knee-flexed position was then rigidly registered to that in the fully extended position by the use of an iterative closest-point shape-matching algorithm. Thus, the femur was held fixed to allow calculation of patellar alignment relative to the femoral coordinate system in each knee-extended and knee-flexed position.…”

Patellofemoral and tibiofemoral alignment in a fully weight‐bearing upright MR: Implementation and repeatability

“…Reporting SEM with the unit‐less ICC(2,1) enhances interpretation of repeatability, providing 95% confidence that true alignment is within one SEM of the point estimate. Lau et al evaluated 3D patellar alignment in participants with patellofemoral pain who were positioned supine and partially weight‐bearing (25% body weight). They reported SEM values that are comparable to our results (flexion 2.40°, spin 0.94°, tilt 0.98°, proximal translation 1.43 mm, lateral translation 1.10 mm, anterior translation 1.67 mm) .…”

“…Lau et al evaluated 3D patellar alignment in participants with patellofemoral pain who were positioned supine and partially weight‐bearing (25% body weight). They reported SEM values that are comparable to our results (flexion 2.40°, spin 0.94°, tilt 0.98°, proximal translation 1.43 mm, lateral translation 1.10 mm, anterior translation 1.67 mm) . This suggests that acquiring images in upright positions does not affect repeatability.…”

“…They reported SEM values that are comparable to our results (flexion 2.408, spin 0.948, tilt 0.988, proximal translation 1.43 mm, lateral translation 1.10 mm, anterior translation 1.67 mm). 22 This suggests that acquiring images in upright positions does not affect repeatability. We reported the additional statistic of SDC 95 , which estimates the smallest amount of change that could be measured on two occasions that would provide 95% confidence that true change has occurred.…”