Contribution of orthodontic analysis to lip morphology prediction in forensic facial reconstruction

In forensic facial reconstruction, lip morphology is difficult to estimate because of the lack of direct bony support. Methods have been developed to evaluate not only average facial soft tissue thickness (FSTT) but also its relationship with the distance between cranial landmarks. In recent years, improved prediction accuracy of FSTT regression models using orthodontic analysis items has been reported, suggesting that these items may help improve the prediction accuracy of lip morphology. In this study, 201 multidetector computed tomography images from living individuals (training group: n = 161, test group: n = 40) were used. Horizontal and vertical distances, FSTT measurements of the lip region, and Northwestern analysis items were measured. Multiple regression models were then constructed using the training group, including models with orthodontic analysis items as explanatory variables (Orthodontic model) and models without them (Non-orthodontic model). When accuracy was validated in the test group, the developed models showed acceptable prediction accuracy and predictive performance in internal validation. Furthermore, when the absolute errors calculated by each model were compared, the Orthodontic model showed significantly lower values, mainly in the items indicating the height of the vermilion lip, as well as in the FSTT of the Incisor superior (p < 0.05). Therefore, considering the relationship with orthodontic analysis items may contribute to improving the accuracy of predicting specific aspects of lip morphology, particularly those related to the vermilion.