Abstract
AIMS: Weightbearing CT (WBCT) has set a new standard for the assessment of foot and ankle alignment in patients with progressive collapsing foot deformity (PCFD) under physiological loading conditions compared with conventional CT. Principal component analysis (PCA) models are currently used for a detailed 3D shape analysis, but are not able to take into account non-linear (e.g. rotational) anatomical variance, which is particularly relevant in PCFD. Innovative advances in geometrical morphometrics by principal polynomial shape analysis (PPSA) are now able to overcome this challenge. Therefore, the objective of this study was to evaluate the use of PPSA in identifying distinct morphological patterns in patients with PCFD under weightbearing conditions. METHODS: In this retrospective comparative study, 40 feet from 20 PCFD bilateral patients imaged by WBCT were confirmed eligible for analysis. Subsequently, matched controls were selected from a cohort of patients who underwent WBCT imaging for clinical follow-up of disorders unrelated to the foot. From the WBCT images, 3D models were reconstructed and registered. PPSA was applied to the 3D foot models to identify and delineate morphology variations in foot shape between the PCFD and control group. RESULTS: Automated classification of PCFD by linear discriminant analysis using the PPSA model yielded a sensitivity of 92.5% and specificity of 92.5%. Furthermore, PPSA revealed distinct foot morphology components in the PCFD group. Anatomical differences were significant and most pronounced at the level of the talocalcaneonavicular joint, with prominent internal and plantar rotation of the talar bone (p < 0.001). CONCLUSION: This study is the first to apply PPSA in patients with PCFD. The findings validate distinct 3D spatial position alterations compared with control subjects. More specifically, they demonstrate that the talocalcaneonavicular joint complex is the most affected structure.