Abstract
OBJECTIVE: To investigate the correlation between nine commonly used cephalometric analyses and facial profile attractiveness and to explore an optimized combination of cephalometric measures. METHODS: Sixteen nonprofessional evaluators assessed the profile attractiveness of 210 untreated Chinese adults using a visual analog scale. Eighty-seven cephalometric measures were obtained from nine analyses (Burstone, Downs, Holdaway, Jarabak, McNamara, Ricketts, Steiner, Tweed, and Wylie). Quadratic regression analysis was employed to identify measures significantly correlated with facial profile attractiveness and to calculate their maximum attractiveness values (MAVs). Stepwise regression was applied to assess the explanatory power of each analysis for profile attractiveness and to construct optimized predictive models. RESULTS: The explanatory power of the nine analyses for attractiveness variation was ranked as follows: Holdaway (41.5%) > Ricketts (37.6%) > Steiner (36.8%) > Burstone (35.7%) > Tweed (35.6%) > Downs (33.9%) > McNamara (24.3%) > Wylie (13.2%) > Jarabak (6.1%). Among individual measures, the H-angle, ANB (°), A-Npog (mm), and NA-APo (°) accounted for more than 26% of attractiveness variation. A five-indicator model comprising H-angle (28.8%; MAV = 17.2°), L1-APog (14.6%; MAV = 0.5 mm), Wits appraisal (4.5%; MAV = 0.1 mm), ANS-Me/N-Me (4.2%; MAV = 54%), and ANS-Ptm (3.3%; MAV = 46.7 mm) explained 55.4% of the variation. CONCLUSION: Among the nine cephalometric analyses, the Holdaway method exhibited the strongest explanatory power for variation in profile attractiveness. The newly constructed five-indicator model may provide more precise aesthetic references for orthodontic and orthognathic treatments.