Determining depths of incipient caries from OCT imaging

We are developing optical coherence tomography (OCT) in combination with polarized Raman spectroscopy for detection and longitudinal monitoring of incipient caries. OBJECTIVES: To determine whether OCT imaging can be used to non-destructively determine depths of incipient caries lesions. METHODS: Triplicate OCT images (850 nm) were obtained from extracted human premolar/molar teeth containing natural incipient caries (n=8 teeth; n=9 approximal lesions). Optical attenuation coefficients ( μ_t) were calculated from compounded A-scans of OCT images. The A-scan profile corresponding to the minimum μ_t was used to determine lesion depth by calculating the number of pixels with OCT back-scattered signal intensity above a threshold value of 500 counts. OCT-derived depth values were correlated with lesion depths measured from polarized light microscopy (PLM) of ~100 μm thick histological tooth sections (n=15 slices). RESULTS: OCT images of incipient lesions displayed highest light back-scattering intensity at the enamel surface with moderate light back-scattering into enamel. PLM also revealed typical triangular shaped subsurface lesions. Lesion depths measured using PLM ranged from 434-1145 μm, with OCT-derived lesion depths ranging from 495-1050 μm. Using Pearson's product-moment correlation analysis, OCT-derived depth values were linearly correlated to PLM depth values at r=0.91 (coefficient of determination r²=0.83) with the null hypothesis of a chance correlation (r=0) rejected at p<0.05. CONCLUSIONS: PLM can be used to determine lesion depths but requires tooth cutting ex vivo to reveal the subsurface lesion. In contrast, OCT imaging can be performed on unsectioned tooth samples in vivo. In this pilot study, the strong correlation of OCT-derived lesion depths with histological depth analyses suggests that OCT imaging can be used to reliably determine the lesion depth in a non-destructive manner. This method has potential for estimating incipient caries depth and guiding clinical treatment decisions. Funded by the Canadian Institutes of Health Research and US National Institutes of Health (R01DE017889).