Haupt et al 1145 Figure 1. Inclusion/exclusion criteria. Figure 2. Axial computed tomographic scan demonstrating positioning of center-center and centroid axes. (A) Simulation of centercenter by dividing the tibial/fibular areas into quartiles. The center-center point at this position is the central line. (B) The centroid axis, plotted by tracing the cortical diameter of fibula and tibia, and computer tools connecting the geometric centers of each shape. (C) The angular deviation measurement of the center axis from (A) and the centroid axis from (B). 20-, and 30-mm levels for each subject. The values from all levels were pooled for this comparison. Additionally, the angular differences between centroid axes of differing levels and between center-center axes of differing levels, respectively, were measured. The published measurement error of the angle measurement tool is 1.1 degrees for angles between line segments of 100 pixels, and error decreases with increasing line segment length.18 The line segments used for this study were approximately 250 pixels or greater, indicating an error less than 0.45 degrees. Statistical analysis was performed using IBM SPSS v.25 statistical software. Observer consistencies were assessed using a 2-way, mixed-effect model intraclass correlation (ICC) looking for absolute agreement of exact measures. Interobserver consistency was assessed using a fully crossed design. Intraobserver consistency was assessed using repeated measurements made 1 month later (test-retest