Poster 257: Assessing the Likelihood of Symptomatic Patellar Instability via the Two-Image Lateral Trochlear Inclination Angle as a Measurement of Trochlear Dysplasia

Objectives: Trochlear dysplasia is the most common anatomic risk factor for patellar instability (PI) and it is present in roughly 90% of patients with recurrent PI. Qualitative descriptions/categorizations of dysplasia include the Dejour and Oswestry-Bristol classifications. Quantitative measurements include the sulcus angle and the lateral trochlear inclination angle (LTI). The LTI has shown utility given that it is a numeric representation of dysplasia at the most proximal aspect of the trochlea where dysplasia is most severe. Recently, the two-image LTI has been described as a more accurate and reliable method to quantitate dysplasia compared to the original one-image LTI with high inter and intrarater reliability. However, since the advent of the two-image LTI, no analysis has been done to determine the likelihood of developing (PI) once the LTI reaches a certain value. A prior analysis of the single image LTI showed a value of 11° as a threshold value between patients with and without PI. The purpose of this study is to determine a two-image LTI threshold value at or below which a patient is likely to develop recurrent PI. Methods: Ninety-nine patients with PI and 91 age and sex matched control patients were identified. All patients had available magnetic resonance imaging (MRI) and any patients with unreadable imaging, questionable diagnosis or neuromuscular or arthritic conditions were excluded. Trochlear dysplasia was measured via the two-image LTI (Figure 1). In the study group, any negative LTI values (patients with dysplasia so severe that the trochlea is convex), were excluded from analysis. This left 64 study patients. Means and standard deviations were calculated, and an independent t-test determined significant differences between groups. A receiver-operating characteristic (ROC) curve was plotted to determine the area under the curve (AUC) and a threshold value was determined to best discriminate between study and control patients. Results: The mean two-image LTI in the study group was 9.4 ± 5.2 and 17.3 ± 6.77 in the control group (p < 0.001). Intrarater reliability of 15 randomly selected study patients and 15 randomly selected control patients was excellent (Intraclass Correlation Coefficient: 0.963, 0.924-0.982). ROC AUC was 0.825 (95% CI: 0.758-0.891) as depicted in Figure 2. A threshold value of 14.5° (Table 1) was chosen as the best demarcating value between controls and PI patients with a sensitivity of 0.828 (53/64), specificity o