Summary
The new sagittal measurements described in this study were able to distinguish between patellar instability and control groups.
Abstract
Objectives: The primary aim of this study was to compare the trochlear morphology of patients with primary patellar dislocations to age and sex-matched controls with ACL injury using sagittal-based MRI measurements to identify any anatomical differences between the two groups.
Methods
A retrospective case-control study was performed comparing consecutive patients presenting with first-time lateral patellar instability (PI) event to an age and sex-matched control population sustaining ACL injury without patellar dislocation (control). Patients without a knee MRI, with a prior PI event, or with a history of ipsilateral knee surgery were excluded. Sagittal MRI measurements were performed at the medial, central, and lateral trochlea. Measurements assessed the proximal trochlear bump height, length, and angle from the anterior femoral cortex. In addition, measurements were obtained pertaining to the trochlear length. Measurements were normalized to each patient’s epicondylar width to adjust for differences in knee size. Summary statistics were computed and comparative univariate analysis using appropriate parametric and nonparametric tests was performed. To quantify the ability of each sagittal measurement to discriminate between PI and control patients, a series of logistic regression models were trained and evaluated with a single sagittal measurement type as the sole predictor. For each model, the outcome was whether a patient was part of the PI or control group. A training and validation process was repeated 100 times and each model’s average area under the receiving operating characteristic (ROC) curve (AUC) was recorded; this was used as a metric of overall model performance.
Results
Measurements were completed in a total of 284 patients: 142 patients with PI and 142 control patients. There were no significant differences in age or sex between groups. Compared to controls, patients with PI had a significantly smaller sagittal bump height in the lateral and medial trochlea (p<0.001, p<0.001), sagittal bump length in the lateral and medial trochlea (p<0.001, p<0.001), and proximal sagittal bump angle in the lateral and medial trochlea (p<0.001, p=0.001). By contrast, patients with PI demonstrated significantly greater linear central and medial trochlear length (p<0.001, p=0.004), curvilinear lateral and central trochlear length (p<0.001, p=0.001).
The three individual measurements that were best at discriminating between control and PI patients were normalized lateral sagittal bump height (Cutoff Point: Classify to PI if <7.69, AUC: 0.77, Sensitivity: 0.70, Specificity: 0.82), central linear trochlear length (Cutoff Point: Classify to PI if >37.31 mm, AUC: 0.77, Sensitivity: 0.52, Specificity: 0.91), and normalized lateral sagittal bump length (Cutoff Point: Classify to PI if <38.22, AUC: 0.75, Sensitivity: 0.71, Specificity: 0.58).
Conclusion
The sagittal measurements described in this study were able to distinguish between PI and control groups. The differences demonstrated a smaller bump height, bump length, and bump angle in the medial and lateral trochlea for the PI group. For trochlear length, there was greater linear and curvilinear length in the central trochlea for the PI group relative to the controls. Improved understanding of sagittal trochlear morphology may provide additional insight into anatomic factors that contribute to patellar instability, and may aid surgical decision making