Image reconstruction in photoacoustic tomography with variable speed of sound using a higher-order geometrical acoustics approximation

Previous research correcting for variable speed of sound in photoacoustic tomography (PAT) has used a generalized radon transform (GRT) model. In this model, the pressure is related to the optical absorption, in an acoustically inhomogeneous medium, through integration over non-spherical isochronous surfaces. This model assumes that the path taken by acoustic rays is linear and neglects amplitude perturbations to the measured pressure. We have derived a higher-order geometrical acoustics (GA) expression, which takes into account the first-order effect in the amplitude of the measured signal and higher-order perturbation to the travel times. The higher-order perturbation to travel time incorporates the effect of ray bending. Incorrect travel times can lead to image distortion and blurring. These corrections are expected to impact image quality and quantitative PAT. We have previously shown that travel-time corrections in ID suggest that perceivable differences in the isochronous surfaces can be seen when the second-order travel-time perturbations are taken into account with a 10% speed of sound variation. In this work, we develop iterative image reconstruction algorithms that incorporate this higher-order GA approximation assuming that the speed of sound map is known. We evaluate the effect of higher-order GA approximation on image quality and accuracy.