Characteristics of proportional analysis for soft tissue facial profile: epidemiological possibilities of measurement item reduction.

Lundström et al. proposed a proportional analysis system for the soft tissue facial profile in the natural head position. To use this method for further epidemiological investigation and to interpret the characteristics of this analysis, each measurement (index) was identified in comparison to the other indices using cluster and factor analyses. Facial profiles of 111 (mean age: 22.9 years) Japanese males were measured and 11 indices (8 horizontal, 2 vertical and 1 horizontal/vertical) were calculated. Almost all internal co-relationships between each index were statistically significant (p < 0.05, 0.01). Variable cluster analysis classified indices into four major clusters and clarified the attributes of the 11 indices. The first cluster was index No. 1, 2, 3 and 7. The second cluster was index No. 6. The 3rd cluster was index No. 4, 5, 11, 8 and 10. The 4th cluster was index No. 9. These clusters are thought as vertical facial balance, upper and lower jaw relation or horizontal/vertical balance, chin morphology, and horizontal facial balance. From factor analysis, three factor axes that explained the characteristics of 11 indices were found (accumulated contribution rate: 76.5%). The heaviest loading factor was index No. 1,2 (0.95) on axis I, 5 (0.83) on axis II and 6 (0.78) on axis III. Therefore, axis I, axis II and axis III are thought to be based on the position of the soft tissue Nasion, SLI and Pogonion, respectively. Common indices which are included in both analyses are thought to be valid as a clue to reduce the number of measurement parameters.