A Quantitative Analysis of Facial Asymmetry in Torticollis Using 3-Dimensional Photogrammetry

Abstract

Objective: To assess whether children with torticollis have quantifiably greater facial asymmetry than their age-matched controls using 3-dimensional (3D) photogrammetry. Design: We retrospectively analyzed patients diagnosed with torticollis and age-matched volunteers who underwent 3D photogrammetry of their faces. We calculated the root mean square deviation (RMSD) between native and reflected facial images, as a measure of asymmetry. Two observers independently measured RMSD values for all study participants. The Spearman correlation coefficient evaluated interobserver reliability. The Wilcoxon rank-sum test with Bonferroni adjusted P values for multiple comparisons. Setting: Institutional. Participants: Twenty patients diagnosed with torticollis and 12 age-matched volunteers. Patients were analyzed on a computer database and volunteers were selected and consented in the hospital. We excluded patients with a history of facial trauma, facial operations, or other craniofacial diagnoses. Interventions: Facial surface scans were obtained using the Canfield Vectra stereophotogrammetry system. The technology captures surface anatomy without radiation. Main Outcome Measures: RMSD comparisons between patients with torticollis and age-matched controls. Results: Compared to controls, patients with torticollis had statistically significant greater full face, upper third, and middle third facial asymmetry. There was a trend toward greater asymmetry of the lower facial third. Conclusions: We used 3D photogrammetry to quantitate facial asymmetry from torticollis. We found greater asymmetry in patients with torticollis than in their unaffected peers. All areas of the face appeared to be affected, though the asymmetry in the lower facial third just failed to reach significance.