Quantification of facial asymmetry: A comparative study of landmark-based and surface-based registrations.

Abstract

To compare the validity and reproducibility of four different methods for the quantification of soft tissue facial asymmetry. Twenty 3D-photographs were randomly selected from the healthy control database. To assess the validity of the one landmark-based and three surface-based methods for measuring asymmetry, artificial facial asymmetry was created on two 3D photographs. The discrepancy between the artificial facial asymmetry and the asymmetry quantified by different methods was calculated. To evaluate the reproducibility of the methods, they were applied three times by two independent observers. The intraobserver and interobserver variations were calculated. The mean absolute differences between the measured asymmetry and the artificial asymmetry for the landmark-based method, forehead method, 5mm method and 1mm method were 1.4mm, 0.3mm, 0.09mm and 0.06mm respectively. The intraobserver and interobserver variations for the forehead, 5mm and 1mm methods were between 0.18mm and 0.34mm. The surface-based methods were valid and highly reproducible in comparison to the landmark-based method for the quantification of soft tissue facial asymmetry. The 1mm surface-based method can be implemented in a clinical and research setting to quantify facial asymmetry.