Weight-Bearing Computed Tomography Bone Density Better Predicts Gait Line Than Traditional Anatomical Landmarks

Research Type: Level 3 - Retrospective cohort study, Case-control study, Meta-analysis of Level 3 studies Introduction/Purpose: Gait causes body weight to travel along a posterior to anterior line over the foot while Wolf's law predicts that load stresses stimulate bone formation. Since Weight Bearing CT (WBCT) detects bone density under load, looking at the densest bone lines using WCBT should predict the position of the gait line. This could be more reproducible and predictable than using anatomical landmarks such as the 2nd metatarsal (M2), subject to many constitutional and pathological variations. The aim of the present study was to compare the gait lines given by dynamic baropodometric and WBCT datasets. We hypothesized that WBCT density line would agree with baropodometry, while M2 would not. Methods: In this retrospective comparative study, 14 asymptomatic feet from patients (BMI 29.16kg/m2±6.47, Age 57y±13.40, FAO 7.36%±3) who had undergone WBCT (CurvebeamAI, PA, USA) and baropodometry (T-Soles, Eindhoven, Netherlands) were analysed. Demographics and hindfoot alignment (Foot Ankle Offset (FAO)) were collected. The feet were automatically segmented to obtain the longitudinal axis of M2. Principal Componant Analysis and (Slicer3D v3.1) was used to identify the WBCT density line (PCA). Linear regression was applied to the 2-dimensional Centre of Pressure data to identify the line of best fit (COP)(T-Soles v10.1). Normality was verified with Shapiro-Wilk tests. Differences between M2, COP and PCA lines were assessed with Student's paired t-test and correlations with Pearson's. Alpha risk was set to 5% (α = 0.05). Results: Relative to the posterior anterior axis (y), COP, PCA and M2 angles were a mean 3.5°±2.2, 2.79°±1.69, and 2.13°±1.19, respectively. COP and PCA angles relative to the y axis were not different (p=0.207), correlated (ρ=0.55; r2=0.306; p=0.04) and in agreement (ICC: 0.52; 95%CI: [0.127; 0.778]; p=0.02), with a mean difference of 0.7°. Conversely, COP and M2 angles relative to y were different (p < 0.05), not significantly correlated (ρ=0.32; r2=0.104; p=0.26), and not in significant agreement (ICC: 0.22; 95%CI: [-0.14; 0.57]; p=0.167), with a mean difference of 1.37°. Conclusion: WBCT-derived bone density lines closely predicted the dynamic gait line, demonstrating agreement with baropodometric data. In contrast, the second metatarsal axis failed to align with the actual weight-bearing path, highlighting the limitations of anatomical landmarks such as the second metatarsal. These findings suggest that WBCT may provide a more reliable, subject-specific yet universal and biomechanically relevant method to define foot orientation. A small sample of asymptomatic but valgus feet may limit the generalizability of our findings. Future studies should investigate these initial findings throughout the full spectrum of varus, neutral and valgus feet, in normal and pathological cases.