INTRODUCTION: The purpose of the present study was to use facial axis (FA) points to classify dental arch form generated from an analysis of 3-D virtual models of a sample of normal occlusions. A secondary aim was to introduce a new arch form template based on this classification for clinical application. METHODS: One hundred and twenty five plaster models of Class I occlusions were 3-D scanned (Orapix Co., Ltd, Seoul, Korea) and FA points digitized on the virtual models using Rapidform 2006 software (INUS Technology Inc., Seoul, Korea). Following intercanine and intermolar arch width and depth measurements, K-means cluster analysis was applied on 77 cases (Dataset 1) to classify the sample into arch form types. A curve of best fit of the mean arch form of each type was generated. The remaining 48 cases (Dataset 2) were mapped into the clusters and a multivariate test was performed to assess the differences among the clusters. RESULTS: Classification into five clusters demonstrated maximum inter-cluster distance in the arch parameters and produced the most homogeneous cluster size. The differences between the 5 cluster types were statistically but not clinically significant and so they were recombined to form three clusters representing 'narrow', 'moderate' and 'wide' arch forms. CONCLUSIONS: A template with three arch form types based on anterior and posterior dimensions has been proposed through 3-D analysis of FA points for more accurate arch form identification and arch wire selection.
INTRODUCTION: The purpose of the present study was to use facial axis (FA) points to classify dental arch form generated from an analysis of 3-D virtual models of a sample of normal occlusions. A secondary aim was to introduce a new arch form template based on this classification for clinical application. METHODS: One hundred and twenty five plaster models of Class I occlusions were 3-D scanned (Orapix Co., Ltd, Seoul, Korea) and FA points digitized on the virtual models using Rapidform 2006 software (INUS Technology Inc., Seoul, Korea). Following intercanine and intermolar arch width and depth measurements, K-means cluster analysis was applied on 77 cases (Dataset 1) to classify the sample into arch form types. A curve of best fit of the mean arch form of each type was generated. The remaining 48 cases (Dataset 2) were mapped into the clusters and a multivariate test was performed to assess the differences among the clusters. RESULTS: Classification into five clusters demonstrated maximum inter-cluster distance in the arch parameters and produced the most homogeneous cluster size. The differences between the 5 cluster types were statistically but not clinically significant and so they were recombined to form three clusters representing 'narrow', 'moderate' and 'wide' arch forms. CONCLUSIONS: A template with three arch form types based on anterior and posterior dimensions has been proposed through 3-D analysis of FA points for more accurate arch form identification and arch wire selection.