Histological validation of cone-beam computed tomography versus laser fluorescence and conventional diagnostic methods for occlusal caries detection.

Abstract

The purpose of this study was to compare the validity of visual (VE), radiological (RE), cone beam computed tomography (CBCT), and laser fluorescence (LFE) examination methods for the detection of the occlusal noncavitated caries in permanent posterior teeth. Two examiners assessed 121 selected sites on the occlusal surfaces of 44 molar teeth by visual (International Caries Assessment and Detection System II [ICDAS]), radiographic (bite-wing projection) cone-beam computed tomography, and laser fluorescence (DIAGNOdent Pen) examination methods. After a 1-week interval, each measurement was repeated by two examiners. Then, the teeth were sectioned, and histological evaluation was performed, which serves as the gold standard. The lesion depths were classified and correlated with the methods evaluated for validation. The intra- and inter-examiner reliability (sensitivity, specificity) and reproducibility of all examination methods were calculated using a weighted Cohen's κ statistic. The correlation between the examination methods was determined using receiver operating characteristic (ROC) analysis indicating the area under the curve (AUC). CBCT exhibited excellent intra-examiner (0.76 for examiner 1, 0.78 for examiner 2) and fair to good inter-examiner (0.63 for the first, 0.64 for the second measurements) reproducibility. The intra-examiner reproducibility was excellent for the LFE method according to the weighted κ values of examiners 1 (0.90) and 2 (0.79). Among the combined methods, the highest AUC values (0.81-0.95) were obtained for the CBCT examination method performed by the two examiners at both the first and second measurements. Cone beam computed tomography showed better performance than other diagnostic methods.