Investigation of a quantitative photoacoustic tomography fitting procedure on multiple targets in reflection geometry with diffuse optical measurement assistance

Abstract

Traditional Photoacoustic tomography provides the distribution of absorbed optical energy densities which are the products of the optical absorption coefficients and fluences. However, the absorption coefficient is the only functional parameter that is related to disease diagnosis, such as cancer. In this paper, we report the experimental investigation of an improved fitting procedure which can quantitatively characterize optical absorption coefficients of multiple targets. The original fitting procedure was proposed by us and used for a single target. The fitting procedure included a complete photoacoustic forward model, which incorporated an analytical model of light transport and a model of acoustic propagation. Using the target information from the PAT images and the background information from diffuse optical measurements (DOM), the fitting method minimizes the photoacoustic measurements and forward model data and recovers the target absorption coefficient quantitatively. The fitting errors in the absorption coefficients can reach 20% to 100% if the original fitting procedure is directly used on multiple targets. In our improved fitting method, the ratio between the photoacoustic intensities is introduced and served as extra input to the fitting procedure. As a result, the total number of unknown parameters is reduced and fitting accuracy is improved. The hybrid system used in the experiment combines a 64-channel photoacoustic system with a frequency-domain diffused optical system. The experiment was performed in the reflection geometry suitable for breast imaging. Phantom experiments include the combination of high contrast and low contrast targets with absorption coefficients ranging from 0.07 to 0.28 cm^-1 and with different spatial separations. The phantoms were inserted into a chicken breast tissue. The fitting errors of multiple targets were reduced to less than 20% for both high and low contrast targets. These results illustrate the potential application of this quantitative DOM-assisted photoacoustic fitting procedure to image and diagnose breast cancer having multiple and complex tumor distribution.