A novel model to predict tooth bleaching efficacy using autofluorescence of the tooth

Objectives: We aimed to confirm whether autofluorescence emitted from teeth can predict tooth bleaching efficacy and establish a novel model combining natural color parameters and tooth autofluorescence data to improve the predictability of tooth bleaching. Methods: A total of 61 tooth specimens were prepared from extracted human molars/premolars and immersed in 35% hydrogen peroxide for 1 h for tooth bleaching. The changes in laser-induced fluorescence (∆LIF) were assessed using Raman spectrometry. Tooth color and autofluorescence data were obtained using quantitative light-induced fluorescence (QLF) technology. Pearson correlation analyses were used to confirm the relationship between ∆LIF and autofluorescence. Intraclass correlation coefficients (ICC) were calculated to compare the conventional and new prediction models. Decision tree analysis was performed to evaluate clinical applicability. Results: The yellowness-to-blueness value from fluorescence imaging showed a moderate correlation with ∆LIF (r= –0.409, p = 0.001). The degree of agreement between the actual efficacy and that predicted by our novel model was high (ICC=0.933, p = 0.002). Decision tree analysis suggested that tooth autofluorescence could be a key factor in prediction of tooth bleaching outcomes. Conclusions: Our findings showed that autofluorescence detected from QLF images may be used to predict tooth bleaching efficacy. Our proposed model appeared to improve the predictability of tooth bleaching.