The aim of the study was to combine an X-ray micro-computed tomography (μCT), enhanced with convolutional neural network (CNN) assisted voxel classification and volume segmentation, with photoluminescence (PL) and micro-Raman spectroscopy (μ-RS) for tooth structural integrity evaluation at the microcrack (MC) site of the extracted human teeth. Four maxillary premolars with visible enamel MCs were first examined utilizing an X-ray μCT and segmented with CNN to identify enamel, dentin, and cracks. Secondly, buccal and palatal teeth surfaces with MCs and sound areas were used to obtain fluorescence spectra illuminated with laser exposure wavelengths of 325 nm (CW) and 266 nm (0.5 ns pulsed), spot diameter ~ 80 μm. Thirdly, chemical composition inside the crack and the difference from the sound area were determined utilizing μ-RS method with a 785 nm laser (CW), spot diameter ∼ 3 μm. The proposed approach, which sequentially integrates X-ray μCT in combination with CNN assisted segmentation, PL, and μ-RS, revealed variations in the material composition along the crack line compared to the sound enamel. This includes alterations in the hydroxyapatite crystals’ quantity and/or quality at the sites of cracks versus uncracked enamel, suggesting a potential compromise in the structural integrity of the tooth in the areas affected by MCs.
|