Accuracy of 2D facial profile photographs under routine clinical conditions compared with 3D imaging

To assess the accuracy of facial soft-tissue profile lines derived from two-dimensional (2D) photographs acquired under routine clinical conditions, using three-dimensional (3D) facial photographs as the reference standard.

This study analysed pre-existing 2D and 3D facial photographs of 160 orthodontic patients. True facial profile lines were generated from 3D photographs using an automated midsagittal plane based on best-fit superimposition of mirrored surface models. Corresponding profile lines were traced on 2D photographs. Fixed landmarks and sliding semilandmarks were placed on predefined profile curves, and configurations were analysed using geometric morphometrics. Pairwise Procrustes superimposition was performed, and Procrustes distances quantified discrepancies between 2D- and 3D-derived profiles. The effects of age and sex were evaluated.

Significant differences were found between 2D-derived and 3D reference profiles (P < 0.001). The median discrepancy was 0.75 mm (IQR 0.79 mm; range 0.01-9.36 mm), with the largest differences in the lip region. Although mean discrepancies were small, variability at the individual level was considerable. Differences exceeding ±1 mm and ±2 mm occurred in 35.2% and 8.2% of measurements, respectively. Errors were greater in males and younger subjects, with no significant interaction.

2D photographs show statistically significant and clinically relevant discrepancies compared with 3D-based profiles. While average errors are small, individual variability should be considered when precise assessment is required.

Routine 2D facial photographs may be insufficient for accurate individual-level soft-tissue profile assessment. Clinicians should be cautious when using them for diagnosis or treatment planning where precision is critical.