Quantitative evaluation of mucosal vascular contrast in narrow band imaging using Monte Carlo modeling

Narrow-band imaging (NBI) is a spectrally-selective reflectance imaging technique for enhanced visualization of superficial vasculature. Prior clinical studies have indicated NBI's potential for detection of vasculature abnormalities associated with gastrointestinal mucosal neoplasia. While the basic mechanisms behind the increased vessel contrast-hemoglobin absorption and tissue scattering-are known, a quantitative understanding of the effect of tissue and device parameters has not been achieved. In this investigation, we developed and implemented a numerical model of light propagation that simulates NBI reflectance distributions. This was accomplished by incorporating mucosal tissue layers and vessel-like structures in a voxel-based Monte Carlo algorithm. Epithelial and mucosal layers as well as blood vessels were defined using wavelength-specific optical properties. The model was implemented to calculate reflectance distributions and vessel contrast values as a function of vessel depth (0.05 to 0.50 mm) and diameter (0.01 to 0.10 mm). These relationships were determined for NBI wavelengths of 410 nm and 540 nm, as well as broadband illumination common to standard endoscopic imaging. The effects of illumination bandwidth on vessel contrast were also simulated. Our results provide a quantitative analysis of the effect of absorption and scattering on vessel contrast. Additional insights and potential approaches for improving NBI system contrast are discussed.