Evaluation of the Interproximal Surface Reconstruction in Digitally Segmented Tooth Models with Varying Levels of Misalignment: An In Vitro Study

Description :

Three-dimensional data has become the standard for treatment planning in orthodontics and maxillofacial surgery, with a shift to digital 3D methods. However, CBCT scans alone are insufficient for detailed analysis of dentition and interocclusal relationships (9). Recent advancements in deep learning-based automatic segmentation have further improved the ability to segment tooth and bone structures in CBCT images, as well as segmentation of IOS meshes. IOS systems can nowadays accurately capture fine occlusal morphology and tooth alignment. Combining CBCT with IOS can provide highly accurate crown-root-bone models for various dental applications (11). To our knowledge, no study has yet addressed the interproximal inaccuracy caused by closely spaced interdental contacts. This study aims to compare interproximal surface reproduction across three models: (1) automatically segmented crowns from intraoral scans (AIOS), (2) fused models combining CBCT roots with AI-segmented crowns (ACIF), and (3) ACIF models with manually adjusted interproximal areas based on CBCT axial views (ACIF adjusted). An in vitro experimental study will be conducted on twenty simulation models of the mandible containing 14 dry teeth. The anterior teeth, from right canine to left canine, will simulate varying levels of crowding. An initial intra-oral scan of each tooth will serve as the ground truth reference. This study seeks to fill the gap in the literature regarding the scarcely studied interproximal surface area providing new insights in the accuracy of the segmentation methods used in reproducing these hard-to-reach surfaces.

Titulaire :
MEDAWAR (EL) SKAF (EL) Lina

Contact USJ :
lina.medawar@usj.edu.lb

Chercheur(s) :
Pr Lina MEDAWAR (EL) SKAF (EL)

Projet présenté au CR, le : 15/06/2025

Projet achevé auprès du CR : 30/05/2026