STATIONARY INTRAORAL TOMOSYNTHESIS IN PROXIMAL CARIES DETECTION: AN EX-VIVO STUDY

BACKGROUND Research on radiographic detection of proximal caries has shown high specificity but low sensitivity, especially for early caries lesions. Stationary intraoral tomosynthesis (s-IOT) shows promise in caries detection, utilizing multibeam carbon nanotube source array technology and a direct intraoral detector to acquire images at similar radiation dose to standard bitewings and generating stacks of parasagittal image slices with an iterative reconstruction algorithm. OBJECTIVE(S) This study aimed to compare the diagnostic accuracy of s-IOT and conventional digital bitewing modalities for the detection of proximal caries lesions. STUDY DESIGN Extracted permanent posterior teeth (n = 64) with proximal surfaces scored between 0 and 4 in the International Caries Detection Assessment System were mounted in phantom heads for imaging. Eight observers, each with 2 or more years of radiology specialty training, scored surfaces on a 5-point Likert scale for caries presence and depth using s-IOT images and digital bitewings, in a standardized viewing format. Fifty percent of specimens will be repeated for intra-examiner reliability assessment. Results will be compared with a histologic gold standard and analyzed by logistic regression and receiver operating curve (Az value). The study is currently in progress, with final results expected by August 2020. RESULTS Pilot test results with 3 observers showed s-IOT to have a sensitivity of 49.6% and specificity of 86.1%, with interobserver agreement of 0.433. The conventional bitewing modality had a sensitivity of 47.9% and specificity of 83.3%, with interobserver agreement of 0.380. DISCUSSION/CONCLUSIONS Based on preliminary data, the 2 imaging modalities showed small differences between sensitivity, specificity, and interobserver kappa reliability, indicating that s-IOT is comparable to digital bitewings for the detection of proximal caries.