Effect of Filtration and Slice Thickness of Cone-Beam Computed Tomography Images on Occlusal Caries Detection: An Ex Vivo Study

Abstract

Objectives: The aim of this study was to evaluate the diagnostic accuracy of different filtrations and slice thicknesses of cone-beam computed tomography (CBCT) in the detection of occlusal caries.

Materials and methods: One-hundred teeth were selected for this ex-vivo experimental study. The CBCT images of the teeth were evaluated and scored by two observers in panoramic and cross-sectional views using different slice thicknesses and filtrations. Paired t-test, repeated-measures analysis of variance (ANOVA), and the least significant difference (LSD) test were used to compare the data with the histological gold standard. Receiver operating characteristic (ROC) analysis was used to determine the diagnostic accuracy of each slice thickness and filtration (P<0.05).

Results: The mean score of true caries detection in cross-sectional views was lower than that in panoramic views (P<0.05). Repeated-measures ANOVA showed a significant difference in the mean of true detections in different thicknesses of cross-sectional views, but this difference was significant only between 5 mm thickness and other thicknesses in panoramic views. On all the views, increasing the thickness decreased the accuracy of caries detection. Repeated-measures ANOVA showed a significant difference between different filtrations; on all the views, increasing the filtration increased the accuracy of caries detection.

Conclusions: An increase of filtration of CBCT images increases the accuracy of occlusal caries detection; however, an increase in slice thickness results in a lower diagnostic accuracy.

Keywords: Biomedical Research; Cone-Beam Computed Tomography; Dental Caries; Filtration.

Fig. 1:

Placement of teeth in a normal anatomical position

Fig. 2:

Cone-beam computed tomography (CBCT) image of the skull containing the teeth

Fig. 3:

Simple block diagram of the different steps of the methodology for better visualization

Fig. 4:

Receiver operating characteristic (ROC) curves of cross-sectional views. (c 5_0= thickness 1 filtration 0, c 5_1=thickness 1 filtration 1, c 5_2= thickness 1 filtration 2, c 10_0= thickness 3 filtration 0, c 10_1=thickness 3 filtration 1, c 10_2= thickness 3 filtration 2, c 20_0= thickness 5 filtration 0, c 20_1=thickness 5 filtration 1, c 20_2= thickness 5 filtration 2)

Fig. 5:

Receiver operating characteristic (ROC) curves of panoramic views. (pan 5_0= thickness 1 filtration 0, pan 5_1=thickness 1 filtration 1, pan 5_2= thickness 1 filtration 2, pan 10_0= thickness 3 filtration 0, pan 10_1=thickness 3 filtration 1, pan 10_2= thickness 3 filtration 2, pan 20_0= thickness 5 filtration 0, pan 20_1=thickness 5 filtration 1, pan 20_2= thickness 5 filtration 2)