Comparative analysis of optical and numerical models for reflectance and color prediction of monolithic dental resin composites with varying thicknesses

Abstract

Objective: To assess the prediction accuracy of recent optical and numerical models for the spectral reflectance and color of monolithic samples of dental materials with different thicknesses. Methods: Samples of dental resin composites of Aura Easy Flow (Ae1, Ae3 and Ae4 shades) and Estelite Universal Flow Super Low (A1, A2, A3, A3.5, A4 and A5 shades) with thicknesses between 0.3 and 1.8 mm, as well as Estelite Universal Flow Medium (A2, A3, OA2 and OA3 shades) with thicknesses between 0.4 and 2.0 mm, were used. Spectral reflectance and transmittance factors of all samples were measured using a X-Rite Color i7 spectrophotometer. Four analytical optical models (2 two-flux models and 2 four-flux models) and two numerical models (PCA-based and L*a*b*-based) were implemented to predict spectral reflectance of all samples and then convert them into CIE-L*a*b* color coordinates (D65 illuminant, 2°Observer). The CIEDE2000 total color difference formula (ΔE00) between predicted and measured colors, and the corresponding 50:50% acceptability and perceptibility thresholds (AT00 and PT00) were used for performance assessment. Results: The best performing optical model was the four-flux model RTE-4F-RT, with an average ΔE00 = 0.72 over all samples, 94.87% of the differences below AT00 and 65.38% below PT00. The best performing numerical model was L*a*b*-PCHIP (interpolation mode), with an average ΔE00 = 0.48, and 100% and 79.69% of the differences below AT00 and PT00, respectively. Significance: Both optical and numerical models offer comparable color prediction accuracy, offering flexibility in model choice. These results help guide decision-making on prediction methods by clarifying their strengths and limitations.