Multispectral Imaging of Enamel Demineralization

Objectives: Most new lesions are found in the pits and fissures of occlusal tooth surfaces. However, conventional techniques, namely dental x-rays, do not provide high sensitivity for detecting early occlusal lesions. Furthermore, studies have shown that visible- and fluorescence-based caries detection systems suffer false positives from staining. Previous work has demonstrated that reflectance imaging at 1450-nm yields high contrast between sound and lesion structure without the interference from staining. However, it is not known whether longer NIR wavelengths demonstrate higher lesion contrast and if lesion severity can be determined from multiple wavelengths. Our objective is to determine which NIR wavelengths for reflectance imaging yields the highest lesion contrast and if its multispectral analysis can infer lesion depth and severity.
Methods: For this study, (n = 25) human molars were collected and sterilized. Acid-resistant varnish was applied over occlusal surfaces excluding a 1.5 x 1.5-mm window. This window was exposed for varying degrees of demineralization: 24, 48, 72, 96, and 120 hours; with a surface softened lesion model at a pH of 4.5. Tooth samples were imaged at selected wavelengths from 405 – 1950-nm. Contrast measurements were calculated using windows of demineralization and sound tissue.
Results: Highest contrast was observed at wavelengths beyond 1460-nm. The contrast was significantly higher (P< 0.05) at 1950-nm than other wavelengths.
Conclusions: This study demonstrates that NIR reflectance at 1950-nm can provide high contrast images of demineralization on tooth surfaces for early detection of occlusal dental lesions, and that contrast ratio increases with lesion depth at 1950-nm.