Abstract: Objective: To assess the correlations between objective measurements of 108 finished orthodontic cases and subjective assessments made by 69 orthodontic specialists, to explore the statistically significant measuring categories of cast and cephalogram and to validate the regression model. Methods:  A stratified random sample of 108 cases was drawn from the large sample of 2 383 patients who finished orthodontic treatment between July 2006 and August 2008 in six orthodontic treatment centers around China. For each patient, the post-treatment information sources evaluated in this study included standardized plaster study casts and a lateral cephalometric X-ray image. These information sources were evaluated both singly and in combination by a panel of 69 orthodontic specialists. The average subjective grading scores of 69 orthodontists were regarded as the gold standard. Six examiners used the peer assessment rating (PAR) index and American board of orthodontics-objective grading system(ABO-OGS) to measure all the study casts respectively and three other examiners measured all the lateral cephalometric X-ray images by using customized software. The objective measuring data were correlated with the gold standard. The correlations between the objective measurement and the subjective evaluation were assessed, the statistically significant measuring categories of cast and cephalogram were explored and the regression model was validated. Results: The ABO-OGS scores of “occlusal relationship” correlated most strongly with the subjective scores of cast (r=0.655, P<0.01), and the secondarily correlated category with those were the PAR scores of “overjet” (r=0.525, P<0.01). The proclination of the lower incisors correlated most strongly with the subjective scores of cephalogram (r=0.446, P<0.01), and the secondarily correlated category with those was the protrusion of the lower lips (r=0.436, P<0.01). Nine components were predictive for the post-treatment model and lateral cephalometric film(Post-M+C) outcome: alignment (ABO-OGS), occlusal relationship (ABO-OGS), interproximal contact(ABO-OGS), L1/NB(°), overjet (PAR), SNB(°), occlusal contacts (ABO-OGS), U1/SN2(°) and centerline (PAR). These 9 components accounted for 72% of the variability in the average subjective grading scores. Conclusion: The objective regression model could replace the averaged opinion of Chinese orthodontic experts effectively, making objective assessment of orthodontic treatment outcome for Chinese patients.

Key words: Malocclusion, Orthodontics, Treatment outcome, Regression analysis