DETECTION OF PRIMARY OCCLUSAL CARIES USING STATIONARY INTRAORAL TOMOSYNTHESIS: A PILOT STUDY

Abstract

<h3>BACKGROUND</h3> The implementation of preventive strategies has aided in the noticeable decline of caries prevalence in developed countries. With the reduction of caries prevalence, there is a higher demand for more sensitive and accurate caries detection modalities. In this study we explore the utility of stationary intraoral tomosynthesis (sIOT) in occlusal caries detection. <h3>OBJECTIVE(S)</h3> In this ex vivo diagnostic accuracy study we aim to evaluate the accuracy and assess the sensitivity, specificity, and area under the ROC curve of s-IOT in detecting primary occlusal caries and to compare the results with digital intraoral bitewings, cone beam computed tomography (CBCT), and the histologic standard reference. <h3>STUDY DESIGN</h3> Sixty de-identified extracted human teeth were selected. The teeth selected included unrestored premolar or molar teeth with fully formed roots. The occlusal surfaces exhibited a status that ranged from sound to carious lesions extending into the inner third of the dentin. Teeth with developmental defects such as hypoplasia or fluorosis were excluded from the study. The teeth were coded and randomly mounted using polyvinyl siloxane (PVS) in dental dentoforms and imaged with the 3 imaging modalities (s-IOT, IO bitewings, and CBCT). A reader study was conducted in which 3 calibrated observers were presented with the 3 imaging modalities. A wash-out period of at least 2 weeks was allowed between sessions, with 1 modality displayed per session. The readers were asked to assess their confidence in caries presence using a 5-point Likert scale and to assess caries depth. After imaging, the teeth were sectioned and histologic examination was completed to establish the ground truth reference. <h3>RESULTS</h3> Accuracy, sensitivity, specificity, and area under the ROC curve of the 3 radiographic modalities will be compared with the histologic standard reference. Results are pending final statistical analysis.