Performance of light-emitting diode device in detecting occlusal caries in the primary molars

This in vitro study aimed to compare the performance of a light-emitting diode (LED) device (Midwest Caries I.D.: MID), International Caries Detection and Assessment System (ICDAS) visual criteria, and fluorescence-based devices (DIAGNOdent: LF; DIAGNOdent pen: LFpen; and Quantitative Light-induced Fluorescence: QLF) in detecting occlusal caries in the primary molars. Eighty-eight primary molars with sound occlusal surfaces or carious lesions at different stages were assessed twice, with a 1-week interval in between, by one examiner using all three methods. Subsequently, the teeth were sectioned and lesion depth was verified using stereomicroscopy as a gold standard. Sensitivity, specificity, and accuracy were calculated at D₁ (all carious lesions—enamel and dentin) and D₃ (dentin lesions) thresholds. Correlation with histological analysis was evaluated using Spearman's rank correlation coefficients (rho). Weighted Kappa and intraclass-correlation (ICC) coefficients were calculated to assess intra-examiner reproducibility. At D₁ threshold, ICDAS and LFpen showed higher sensitivity than the other methods, whereas ICDAS, LF, and QLF showed higher specificity (p < 0.05), and MID showed lower accuracy. At D₃ threshold, ICDAS, LFpen, and QLF showed higher sensitivity than MID, whereas ICDAS, LF, and MID showed higher specificity (p < 0.05). All methods, except MID, showed statistically similar accuracy values (p < 0.05). Correlations with histopathological analysis varied from 0.15 (MID) to 0.57 (ICDAS). Intra-examiner reproducibility varied from 0.30 (MID) to 0.92 (ICDAS, LF, and QLF). The MID device exhibited a poor performance in detecting occlusal carious lesions in the primary molars, and ICDAS visual criteria exhibited greater accuracy than LF, LFpen, and QLF devices.