Aim: To develop a periodontal disease prediction (PDP) software program and a patient-based gingival recession simulator for clinical practice with the aim of improving the oral hygiene motivation of patients with periodontal problems.
Materials and methods: The developed PDP software has three components: a) A data loading window (DLW), b) A three-dimensional mouth model (3DM), and c) a periodontal attachment loss indicator (PLI). The demographic and clinical examination details of 1057 volunteers were recorded to the DLW. An unsupervised machine learning K means clustering analysis was used to categorize the data obtained from the study population and to identify the periodontal risk groups. An intraoral scanner was utilized to capture the direct optical intraoral data of the patients, which was transferred to the 3DM. The intraoral model underwent two algorithm steps to obtain a recessed model: First, the gingival curves separating the gingiva and tooth were extracted using a Dijkstra’s algorithm. Then, the limit curves determining the boundaries of the recessed regions in the intraoral model were obtained using the gingival curves.
Results: Study participants were divided into three different periodontal risk categories: low- (n = 462), medium- (n = 336), and high-risk (n = 259) groups. The gingival curves separating the gingiva and tooth were extracted, and recessed models were obtained and given inputs for the expected amount of recession via the here-proposed method/algorithm. Furthermore, the user can also demonstrate the gingival recession gradually via the slider method incorporated into the developed program.
Conclusions: A user-friendly computer-based periodontal risk estimation tool that is also a patient-specific gingival recession simulator was developed and presented for clinical use by dentists.
Keywords: computer-aided design, Dijkstra’s algorithm, gingival recession, oral hygiene motivation