Photoacoustic computed tomography in biological tissues: algorithms and breast imaging

Abstract

Photoacoustic computed tomography (PAT) has great potential for application in the biomedical field. It best combines the high contrast of electromagnetic absorption and the high resolution of ultrasonic waves in biological tissues. In Chapter II, we present time-domain reconstruction algorithms for PAT. First, a formal reconstruction formula for arbitrary measurement geometry is presented. Then, we derive a universal and exact back-projection formula for three commonly used measurement geometries, including spherical, planar and cylindrical surfaces. We also find this back-projection formula can be extended to arbitrary measurement surfaces under certain conditions. A method to implement the back-projection algorithm is also given. Finally, numerical simulations are performed to demonstrate the performance of the back-projection formula. In Chapter III, we present a theoretical analysis of the spatial resolution of PAT for the first time. The three common geometries as well as other general cases are investigated. The point-spread functions (PSF's) related to the bandwidth and the sensing aperture of the detector are derived. Both the full-width-at-half-maximum of the PSF and the Rayleigh criterion are used to define the spatial resolution. In Chapter IV, we first present a theoretical analysis of spatial sampling in the PA measurement for three common geometries. Then, based on the sampling theorem, we propose an optimal sampling strategy for the PA measurement. Optimal spatial sampling periods for different geometries are derived. The aliasing effects on the PAT images are also discussed. Finally, we conduct numerical simulations to test the proposed optimal sampling strategy and also to demonstrate how the aliasing related to spatially discrete sampling affects the PAT image. In Chapter V, we first describe a prototype of the RF-induced PAT imaging system that we have built. Then, we present experiments of phantom samples as well as a preliminary study of breast imaging for cancer detection.