Objectives: We aimed to compare the diagnostic accuracy of two intraoral digital X-ray sensors-the charged-coupled device (CCD) and complementary metal-oxide-semiconductor (CMOS)-for proximal caries detection in permanent molar and premolar teeth. Micro-CT served as the reference standard.
Methods: 250 samples were mounted in three-dimensional (3D)-printed phantoms, and their proximal surfaces were evaluated by ICDAS criteria directly to create a balanced sample. Bitewing radiography was conducted using 3D-constructed X-ray phantoms with a CCD sensor at a 0.08 s and a CMOS sensor at 0.12 and 0.16 s exposure time. Two examiners determined the diagnostic decisions twice at appropriate intervals. Three diagnostic thresholds for sound surfaces and enamel and dentin caries were defined and presented in a cross-table. Sensitivity and specificity values and overall accuracy were calculated, and receiver operating curves were generated and compared. Reliability assessment was performed using linear weighted κ statistics.
Results: The overall accuracies between the reference standard and different sensors and exposure times were 63.1% (CCD), 67.1% (CMOS sensor at 0.12 s) and 70.7% (CMOS sensor at 0.08 s). High specificity but low sensitivity values were found for all examination conditions at all thresholds. The area under the curve comparison values revealed no significant difference between sensor types and exposure times. Linear-weighted κ analysis revealed almost perfect agreement for all assessments.
Conclusion: No significant difference was found for diagnostic performance of proximal caries detection between the different sensors and exposure times. The increased exposure time did not lead to a significant diagnostic benefit.
Keywords: Radiography, Bitewing; Reproducibility of Results; Sensitivity and Specificity; X-Ray Microtomography.