Abstract
Background:
Patellar instability (PI) is a relatively rare knee condition but most commonly occurs in children and adolescents. The etiology of PI is complex and multi-faceted. Trochlear dysplasia, patellar tilt, tibial tuberosity-trochlear groove (TT-TG) distance, and patella alta have all been cited as contributory factors. Some authors have postulated that trochlear dysplasia (TD) alters the patellofemoral joint such that several of these radiographic parameters are changed relative to the amount of TD present. However, methods to quantify TD have not reached universal agreement. The lateral trochlear inclination (LTI) is one measure that has been described to characterize and quantify trochlear morphology on axial MRI images. Recently, a modified technique for measuring LTI that references the femur’s posterior condylar axis has demonstrated near-perfect inter- and within-rater reliability (ICC: 0.971, 95% CI: 0.88 – 0.97) and may serve as a new method to reliably quantify TD. The purpose of this study was to evaluate this modified LTI measurement against historically used radiographic measures of PI including TT-TG distance, lateral patellar inclination (LPI) angle, and the Caton-Deschamps (CD) ratio. We hypothesized that lower LTI measurements (i.e. more dysplastic trochleas) would be associated with higher TT-TG distances, increased patellar tilt, more patella alta and vice versa.
Methods:
Patients aged nine to eighteen years who were treated for PI at our tertiary referral center between January 2014 and August 2017 were queried. Radiographs and MRI exams of the symptomatic knee were obtained for measurement. On MRI, LTI measurements were made using the novel technique with reference to the orientation of the fully-formed posterior condyles. LPI measurements were conducted using a novel technique where the angle of the widest portion of the patella is measured relative to the posterior femoral condyles (ICC = 0.885, 95% CI: 0.77 – 0.95). TT-TG distances were measured on MRI from both the proximal (pTT-TG) and distal (dTT-TG) extents of the trochlea. The CD ratio was measured on lateral radiographs. All measurements were made by three independent observers. A separate cohort was randomly selected from the original list for the same measurements on two occasions, at least two weeks apart. Inter- and intra-rater correlation coefficients were then calculated using this subgroup and regression analysis was performed using data from the entire cohort.
Results:
Sixty-five patients met inclusion criteria for this study, from which thirty were selected for reliability analysis. Average pTT-TG distance (15.7±4.5 mm) was not significantly different from dTT-TG distance (15.7±4.2 mm) (p=0.94). Good inter-rater reliability was found for both pTT-TG (ICC=0.78, 95% CI: 0.59-0.89) and dTT-TG (ICC=0.83, 95% CI: 0.66-0.91). Both measurements also had excellent within-rater reliability (ICC=0.95, 95% CI: 0.90-0.98 for pTT-TG; ICC=0.90, 95% CI: 0.80-0.95 for dTT-TG). Linear regression analysis showed LTI to have a weak but significant inverse correlation with both pTT-TG (r = -0.40, b=-0.15 mm, p<0.001) and dTT-TG (r = -0.59, b=-0.21 mm, p<0.001). LTI also showed moderate and significant inverse correlation with LPI (r = -0.69, b=-0.54degrees, p<0.001). The average CD ratio was 1.3±0.3 (range 0.8-2.1). LTI did not show significant correlation with the CD ratio (r = -0.19, p=0.14). Moderate inter-rater reliability (ICC=0.62, 95% CI: 0.34-0.80) and high intra-rater reliability (ICC=0.82, 95% CI: 0.66-0.91) were found for CD ratio measurements.
Conclusion/Significance:
In our cohort, smaller LTI measurements (higher levels of dysplasia) were found to be significantly correlated with larger TT-TG measurements and greater patellar inclination angles which supports the notion that these radiographic measures are dependent variables relative to dysplasia and thus may not be consistent independent risk factors for patellar instability. Interestingly, pTT-TG and dTT-TG distances were not significantly different, suggesting that while TD tends to medialize the trochlear groove, once the groove is formed it proceeds distally in a fairly straight line. The CD ratio was not found to be correlated with dysplasia represented by LTI in this cohort which suggests that patellar height is independent of dysplasia. Taken together, the results of this study suggest that TD may be the principal anatomic risk factor for patellar instability and that the TT-TG and patellar tilt are altered secondary to TD. Patellar height did not have a correlation with TD and thus should be analyzed separately as to its role in patellar instability.