2D Manual vs 3D Automated Assessment of Alignment in Normal and Charcot-Marie-Tooth Cavovarus Feet Using Weightbearing CT

Abstract

Background: The complex deformities in cavovarus feet may be difficult to assess and understand. Weightbearing computed tomography (WBCT) is increasingly used in evaluating complex deformities. However, the bone axes may be difficult to calculate in the setting of severe deformity. Computer-assisted 3D axis calculation is a novel approach that may allow for more accurate assessment of foot alignment or deformity. This study aimed to assess differences in measurements done manually on 2D slices of WBCT vs 3D computer models in normal and cavus feet. Methods: We retrospectively analyzed WBCT scans from 16 normal and 16 cavus feet in patients with Charcot-Marie-Tooth disease. Six measurements were assessed: talar–first metatarsal angle (axial plane), forefoot arch angle (coronal plane), and Meary angle, calcaneal pitch, and cuneiform and navicular to floor distances (sagittal plane). 2D measurements were performed manually and 3D measurements were performed using specialized software (Bonelogic, DISIOR). Bland-Altman plots and linear regression were used to analyze differences. Results: There were no significant biases in measured variables in normal feet. In the cavovarus group, automated assessment calculated increased sagittal plane deformity (fixed bias 7.31 degrees for Meary angle, 2.39 degrees for calcaneal pitch) and less axial plane deformity (fixed bias 10.61 degrees for axial talar–first metatarsal angle). The latter also demonstrated proportional bias indicating greater discrepancy with increasing adduction. Conclusion: Measurements were comparable in normal feet suggesting automated techniques can reliably assess the alignment of bony axes. However, automated calculations show greater sagittal and less axial deformity in cavovarus feet than measured by manual techniques. This discrepancy may relate to rotational deformity seen in cavovarus feet, which may not be easily assessed manually. 3D automated measurements may therefore play a role in better assessing and classifying the cavovarus foot, which may ultimately inform treatment algorithms. Level of Evidence: Level III, retrospective comparative study.