Objectives: The aim was to compare the caries diagnostic accuracy of two cone beam CT systems (CBCT) with two intraoral receptors, one digital and one film.
Methods: 100 non-cavitated extracted human teeth were placed with approximal contacts: each row of teeth comprised 1 canine, 1 first and second premolar, and 1 first and second molar. Radiographs of each tooth were recorded using two intraoral modalities: Digora-fmx (Soredex) and film (Kodak Insight), and two CBCT systems: NewTom 3G (Quantitative Radiology) in three fields of view (FOVs): pixel size of 0.36 mm (FOV 12 inches), pixel size of 0.25 mm (FOV 9 inches), pixel size of 0.16 mm (FOV 6 inches); and 3DX Accuitomo (Morita), pixel size of 0.125 (FOV 4 cm). The volumetric data from the CBCT systems were reconstructed and sectioned (0.5 mm) in the mesiodistal tooth plane. Six observers scored approximal and occlusal surfaces for the detection of caries lesions. The teeth were clinically sectioned and microscopy served as the validation tool. Two-way ANOVA tested differences in sensitivity, specificity and overall true score (true positives plus true negatives) between the modalities.
Results: Microscopy of approximal surfaces found 63% sound, 31% enamel and 6% dentin lesions; of occlusal surfaces, 6% sound, 59% enamel, 19% dentin lesions and 16% had fillings. For approximal surfaces, NewTom 12 inch and 9 inch images had significantly lower sensitivities than Accuitomo (P < 0.02); and NewTom 9 inch and 6 inch had significantly lower specificities than film and Digora-fmx (P < 0.04). Accuitomo images were not significantly different from film or Digora-fmx images for any of the tested variables (P > 0.2). For occlusal surfaces, the Accuitomo presented a higher sensitivity than the other systems. Specificity and overall true score did not differ (P > 0.06) among the modalities.
Conclusions: In conclusion, the NewTom 3G CBCT had a lower diagnostic accuracy for detection of caries lesions than intraoral modalities and the 3DX Accuitomo CBCT. The Accuitomo CBCT had a higher sensitivity than the intraoral systems for detection of lesions in dentin, but the overall true score was not higher.