Quantitative high-resolution photoacoustic spectroscopy by combining photoacoustic imaging with diffuse optical tomography

Abstract

The specificity of both molecular and functional photoacoustic (PA) images depends on the accuracy of the photoacoustic absorption spectroscopy. Because the PA signal is a product of both the optical absorption coefficient and the local light fluence, quantitative PA measurements of absorption require an accurate estimate of the optical fluence. Lightmodeling aided by diffuse optical tomography (DOT) methods can be used to provide the required fluence map and to reduce errors in traditional PA spectroscopic analysis. As a proof-ofconcept, we designed a phantom to demonstrate artifacts commonly found in photoacoustic tomography (PAT) and how fluence-related artifacts in PAT images can lead to misrepresentations of tissue properties. Specifically, we show that without accounting for fluence-related inhomogeneities in our phantom, errors in estimates of the absorption coefficient from a PAT image were as much as 33%. To correct for this problem, DOT was used to reconstruct spatial distributions of the absorption coefficients of the phantom, and along with the surface fluence distribution from the PAT system, we calculated the fluence everywhere in the phantom. This fluence map was used to correct PAT images of the phantom, reducing the error in the estimated absorption coefficient from the PAT image to less than 5%. Thus, we demonstrate experimentally that combining DOT with PAT can significantly reduce fluence-related errors in PAT images, as well as produce quantitatively accurate, highresolution images of the optical absorption coefficient.