Monte Carlo OCT simulations investigating the effect of multiple scattered light on OCT-based attenuation coefficient quantification.

The attenuation coefficient µ t of biological tissue could serve as an indicator of structural and functional changes related to the onset or progression of disease. Current state-of-the-art depth-resolved OCT-based µ t quantification methods rely on the assumption that the detected signal is dominated by single-scattered light, limiting the applicability of the model primarily to weakly scattering samples or superficial imaging depths. To assess the effect of multiple scattered light on the accuracy of the single-scattering model and attenuation coefficient estimation for different scattering coefficients and focus positions, we employ a hybrid particle-wave Monte Carlo simulation model to acquire focus series depth profiles for several Intralipid 20% dilutions. Additionally, we perform focus series simulations with illumination beams of half and twice the standard beam radius for a single sample to investigate the effect of the beam radius on the accuracy of attenuation coefficient estimation and of the single-scattering model. For each depth profile we calculate the fraction of single-scattered light contributing to the OCT signal. At each depth we fit a confocal function to the focus series depth profiles to determine the Rayleigh range z R . We estimate the depth-resolved µ t for each depth profile by employing Vermeer's method after confocal function compensation based on the beam parameters. For weakly scattering samples (scattering coefficient µ s < 1 mm-1) we find that the single-scattered light fraction is high over the entire 5 mm depth range, while the measured confocal functions and µ t match the single-scattering model predictions. For the more strongly scattering samples (µ s ≥ 1 mm-1) we find that the single-scattered light fraction decreases rapidly with depth and the fitted confocal functions and estimated µ t start to deviate from the single-scattering model predictions. Finally, we find a beam parameter dependent confocal shape in the single-scattered light fraction curves which is caused by confocal rejection of multiple scattered light by the single-mode fibre detector. Therefore for accurate attenuation coefficient estimation based on a single-scattering model, a narrow but well-collimated beam is optimal for maximal rejection of multiple scattered light.