Background: Patellar instability (PI) is relatively rare but occurs most often in younger patients with underlying pathoanatomy. Trochlear dysplasia (TD) is one of many identified PI risk factors, but consensus is lacking on ideal radiographic measurements. The Dejour classification of TD on lateral radiographs is widely accepted but has suboptimal intra and interrater reliability and does not allow quantification of TD. Lateral trochlear inclination (LTI) measured on the most proximal axial magnetic resonance image (MRI) of the trochlear chondral surface is another described measurement of TD. LTI has historically been described with reference to the posterior aspect of the femur at the same axial level at which the proximal trochlea is measured. However, given the transitional anatomy of the distal femur, the LTI may be better represented by referencing the axis of the fully formed posterior femoral condyles. The posterior condyles represent a true axis of rotation that serves as an internal reference for knee motion and is clearly visible on MRI. We hypothesized that modified LTI measurements (LTI) referencing the posterior condylar axis would differ from the apparent LTI (ALTI) in a pediatric and adolescent population. We also hypothesized that the LTI would have stronger intra and inter reliability than the ALTI measurement and Dejour classification. Lastly, we hypothesized that the most proximal level of the trochlea would adequately represent overall proximal trochlear morphology. This is clinically relevant because dysplasia is most severe on the proximal trochlea and normalizes distally towards the intercondylar notch. Methods: Patients aged 9 to 18 years treated for PI at our tertiary referral center between January 2014 and August 2017 were identified. The Dejour classification was determined on lateral knee radiographs. The ALTI was measured as previously described on axial MRI images (Figure 1A). The LTI (also referred to as LTI #1) was measured on the same MRI image with respect to the angle of the posterior condyles (Figure 1B-C). The LTI was measured again in this fashion at the three subsequent, consecutive axial levels (LTI#2, LTI#3, LTI#4) to capture the first 12 mm of the proximal trochlea. The average of these measurements (LTI-avg) was calculated for each patient. All measurements were performed by two independent observers. A cohort of 30 patients were randomly selected for reliability analysis which was performed twice by three independent observers at least two weeks apart. Inter- and intra-rater correlation coefficients were calculated on this subgroup. Regression analysis was performed on the entire cohort. Results: Sixty-five patients met inclusion criteria for this study, and thirty patients were randomly selected for reliability analysis. Inter- and intra-rater reliability for ALTI showed good agreement but trended towards more variability than the inter- and intra-rater reliability for LTI#1 which had near perfect agreement (Table 1). Inter- and intra-rater reliability for all subsequent LTI measurements and LTI-avg had high or near perfect agreement (Table 2). The Dejour classification had poor to moderate inter-rater and good to near perfect intra-rater reliability. The crossing sign was the most reliable radiographic feature (Table 3). In the entire cohort of 65 patients, the average ALTI (9.2+/-12.6 degrees) was 7.0+/-3.4 degrees greater (less dysplastic) than the average LTI #1 (4.2+/-11.9 degrees) (p = 0.013). Referencing the 11 degrees LTI threshold value for trochlear dysplasia reported in the literature, the ALTI was below 11 degrees in 60% of our PI patients indicating dysplasia, while the LTI was less than 11 degrees in 71% of our PI patients. Regression analysis demonstrated statistically significant positive correlation between LTI#1 and LTI#2 (r=0.88, beta=0.81, p<0.0001), LTI#1 and LTI#3 (r=0.67, beta=0.54, p<0.0001), LTI#1 and LTI#4 (r=0.65, beta=0.43, p<0.001), and LTI#1 and LTI-avg (r=0.91, beta=0.70, p<0.0001). Conclusion: LTI has higher intra and interrater reliability when performed with reference to the posterior condyles compared to the historical measurement (ALTI) and the Dejour classification. The significant and strong correlation between LTI#1 and subsequent LTI measures as well as LTI-avg shows that 90% of TD is represented on the first, most proximal axial image and thus provides an appropriate, reliable and quantifiable measurement of TD in children and adolescents with PI. The significant difference found between LTI and ALTI shows that the historical measurement appears to underestimate dysplasia. Thus, previously described threshold values should be re-examined using this new technique to appropriately characterize trochlear dysplasia in patients with patellar instability as this can have implications for treatment algorithms for these patients. [Table: see text][Table: see text][Table: see text][Figure: see text]
Interrater Reliability of the Berg Balance Scale When Used by Clinicians of Various Experience Levels to Assess People With Lower Limb Amputations
