Reproducibility of facial soft tissue landmarks on 3D laser-scanned facial images.

BACKGROUND: The three-dimensional (3D) measuring technology is useful to inspect facial shape in three planes of space (X, Y, and Z). Recent work has been directed to analyse craniofacial morphology using facial soft tissue landmarks to identify facial differences among population. The reproducibility of facial landmarks is almost necessary to ensure accurate 3D facial measurements.
OBJECTIVE: The aim of this study is to assess the reproducibility of facial soft tissue landmarks using laser-scan 3D imaging technology. SUBJECTS AND METHODS: Facial landmarks were assessed for 30 15(1/2)-year-old British-Caucasian children (15 males and 15 females). The sample was recruited from the Avon Longitudinal Study of Parents and Children (ALSPAC). The 3D facial images were acquired for each subject using two high-resolution Konica/Minolta laser scanners. Twenty-one facial landmarks (63 X, Y, and Z coordinates) were identified and recorded on each 3D facial image by two examiners. The reproducibility of landmarks identification at 2-week interval was assessed for one of the examiners (intra-examiner). In addition, the reproducibility of landmarks was assessed between the two examiners (inter-examiner). Using Bland-Altman plots, both intra- and inter-examiner assessments had evaluated landmarks reproducibility in three dimensions for the sample divided by gender. The reproducibility of the 3D-coordinates for each landmark was considered under three categories (< 0.5 mm, < 1 mm, and >1 mm) for both intra- and inter-examiner reproducibility assessments.
RESULTS: The distribution of coordinates at the three levels of reproducibility show the following percentages: intra-examiner: < 0.5 mm (38%), < 1 mm (51%), >1 mm (11%); inter-examiner: < 0.5 mm (35%), < 1 mm (48%), >1 mm (17%). Generally, 10 landmarks were reproducible to less than 1 mm for both intra- and inter-examiner reproducibility assessments. The Labiale Superius was the most reproducible and Palebrale Superius was the least reproducible landmark. Some landmarks showed greater reliability in certain planes of space; the Glabella was more reliable in the Z than the Y axis. Gender differences were found; Subnasale was more reproducible in the Y-axis in males compared with females.
CONCLUSIONS: The reproducibility of facial landmarks should be considered in the three planes of space. The majority of X-Y-Z coordinates taken to the 21 facial landmarks were reproducible to < 1 mm which is clinically acceptable. The accuracy of landmarks identification ranged from 0.39 to 1.49 mm. The reliability in identification depends on the clarity and definition of each landmark as well as gender characteristics. The different landmarks reproducibility should be considered when evaluating changes related to growth and healthcare interventions.