Linear Anterior-Posterior Computed Tomography Parameters Used to Quantify Trochlear Dysplasia Are More Reliable Than Angular Measurements.

PURPOSE:(1) To evaluate the reliability of 9 commonly used quantitative parameters of the trochlear morphology on computed tomography (CT) and (2) to analyze for differences in the reliability regarding patient subgroups (patellofemoral instability [PFI] vs non-PFI). METHODS:A retrospective analysis of lower-limb CT scans performed between August 1996 and February 2013 was performed. The CT scans of all patients with PFI and 30 randomly selected cases without a history of PFI (non-PFI) were included. The following measurements were performed on 1 proximal axial CT slice at the entrance of the trochlear groove and 1 slice 5 mm further distal: relative medial, central, and lateral trochlear height; trochlear depth; relative transverse trochlear shift; trochlear facet asymmetry; sulcus angle; and medial and lateral trochlear slope. Four investigators performed the measurements independently, and intraclass correlation coefficients (ICCs) were calculated for the entire study group, as well as for the PFI and non-PFI groups separately. RESULTS:In total, 66 cases (36 PFI cases) were included in the study. We found almost perfect inter-rater and intrarater agreement for the trochlear height on both axial CT slices (ICC, 0.831-0.977). For the other measurements, we found only fair reliability (ICC < 0.4) on the proximal CT slice, whereas on the distal CT slice, at least moderate reliability (ICC > 0.4) was observed. ICCs were lower for many parameters in the PFI group. Angular values were less reliable than linear values. In particular, measurements involving the medial facet (i.e., sulcus angle, medial trochlear slope, and trochlear facet asymmetry) were less reliable. CONCLUSIONS:When interpreting quantitative parameters defining the trochlear morphology, one must taken into account the considerably lower reliability of angular parameters such as the commonly used sulcus angle compared with linear measurements. Radiologic measurements are less reliable in cases of PFI than in subjects without instability. LEVEL OF EVIDENCE:Level III, retrospective case-control study.