Predicting first molar width using virtual models of dental arches

In dentistry, large regions of dentition often are restored with minimal information about the original anatomy. The ability to predict missing anatomy from existing anatomy would aid such restorations. This study investigated the relationship between first molar mesial-distal width and arch shape using newly defined reference points and three-dimensional (3D) digital methods. Full-mouth dental stone casts from 167 dental students were scanned and rendered as 3D virtual models. Maxillary and mandibular arch lengths and widths and first molar mesial-distal widths were measured on the virtual models using new definitions incorporating virtual planes. A linear mixed model of the first molar width regressed on the other measurements was done. Intraobserver reproducibility was evaluated by means of intraclass correlation (ICC) and standard deviation of measurement error (SDME). All measured distances were averaged as a combined group and as gender groups. The correlation coefficients between the maxillary and mandibular first molar widths were over 0.70 (P < 0.01). Intraobserver error was small. ICCs were over 0.92 and SDMEs were from 0.11 to 0.21 mm. Arch dimensions and first molar widths were defined and measured. Regression equations were calculated for predicting first molar width. The prediction of first molar width using arch dimension is essential for virtual designing of missing first molars. This approach also provides reliable reference point definitions for the virtual dental model which was impossible with traditional measurement methods. Therefore, this study would be helpful for understanding the 3D anatomy of dental arch and fabrication of automatic prosthodontic restorations.