Comparison of Various Imaging Modes for Photoacoustic Tomography

Abstract

Photoacoustic tomography (PAT) has shown great potential in many medical applications. According to the geometry of electromagnetic irradiation and acoustic signal detection, PAT can be categorized into three imaging modes, the forward mode, the backward mode and the sideward mode. To deeply understand their theorems and suitable applications, simulation studies are carried out to compare the precision and noise robustness of reconstructions of these modes. The absorbed energy density of the irradiated tissue is modeled by the Monte Carlo method. Then generated photoacoustic signals are simulated by the finite-difference time-domain method. Four representative algorithms are employed to reconstruct the image from simulated signals. It is found that in the sideward mode, the reconstructed image around the center of the scanning circle has better precision and stronger robustness to noise compared with the marginal region. In both forward and backward modes, the region relatively close to transducers can be reconstructed more clearly, but is more sensitive to noises. The sideward mode has better precision but slower reconstruction speed compared with forward and backward modes. Moreover, the choice of the time-domain or frequency-domain reconstruction algorithm is shown to have an impact on the noise pattern of the reconstructed image. Therefore, the sideward mode PAT is suitable for the precise imaging of protuberant human tissues, and the optimum imaging region (OIR) is around the center of the scanning circle. The forward/ backward mode PAT is suitable for the fast imaging of plane human parts, and the OIR is at an appropriate distance from the scanning line.