Noncontact imaging of absorption and scattering in layered tissue using spatially modulated structured light

Abstract

Sinusoidal patterns of spatially modulated near-infrared (650 nm) structured light are used to interrogate multilayer phantoms and tissue. Diffuse reflectance is imaged at multiple spatial frequencies from 0–0.3 mm−1. ac and dc components of the image are fit to a two layer model formulated from the diffusion approximation to the Boltzman transport equation. The two-layer model depends on optical properties (absorption, μa, and reduced scattering, μs′) in each layer and on top layer thickness (d). Layered tissue phantoms with variable optical properties in each layer (μa=0.006–0.034 mm−1 and μs′=0.89–1.45 mm−1) were constructed to test the accuracy of the model. Constraining top layer thickness to within 25% of the correct value in a four-parameter fit results in recovery of upper layer optical properties with average accuracies of ±2% for top layer μs′ and ±17% for top layer μa. Bottom layer μa can then be recovered to an average accuracy of ±25% with two parameter fits. Average accuracies of top and bottom layer absorption can further be improved to 12% and 18%, respectively, by fitting for each alone. Bottom layer scattering and top layer thickness do not vary significantly from initial guesses because of poor sensitivity to these parameters in frequency dependent reflectance data. Measurements of in vivo volar forearm optical properties at 650 nm produced spatially varying skin (d=2 mm) optical property maps that range from 0.025–0.045 and 1.7–2 mm−1 for upper layer μa and μs′ and 0.005–0.015 and 0.5–3 mm−1 for lower layer μa and μs′, respectively. These preliminary results suggest that spatial modulation of the source provides sufficient depth sensitivity to allow noncontact mapping and quantification of layered tissue optical properties using a wide-field, noncontact approach.