Iterative Fluence Compensation and Spectral Unmixing for Spectroscopic Photoacoustic Imaging

An iterative fluence compensation and spectral un-mixing algorithm for spectroscopic photoacoustic imaging (SPA) is described. The algorithm focuses on solving the optical inverse problem. It employs optical prior knowledge of tissues, leverages a Monte Carlo simulator for fluence estimation, and assumes a linear mixed model for absorption, scattering, and anisotropy. After an initial guess of the tissue composition, the algorithm sequentially estimates the light fluence, solves for tissue concentrations in spectral unmixing, and updates the optical parameters iteratively until the estimated initial pressure converges to the measurement. The algorithm was validated in simulation, where ground truth data was synthesized from an in vivo study of prostate cancer on mice model. Performance of the algorithm and the scenario without fluence compensation was compared, and the convergence and the precision of the algorithm are reported.