Acoustical and Optical Compensation for Spectral Unmixing of Sulfates using Ultrasound and Photoacoustic Tomography: In Silico and In Vitro Results

Abstract

Photoacoustic tomography (PAT) is a powerful adjunct to ultrasound tomography (UST), as it can provide functional and molecular information in addition to the structural information from UST. This means markers such as hypoxia and angiogenesis can be used in combination with acoustic sound speed (SS) and attenuation to aid in the diagnosis of tumors and other pathologies. However, one area of active research in photoacoustic imaging (PAI) is correcting for heterogeneous acoustic and optical properties. If left uncounted for, errors can be introduced into quantitative measurements such as oxygen saturation rendering them unsuitable for use by physicians. Therefore, a need exists to correct for these heterogeneities and to provide useful quantitative measurements. Here we have extended our previous work on model-based acoustic and optical compensation to improve spectral unmixing, to demonstrate the potential future improvements in in vivo oxygen saturation measurements. In silico and in vitro results were analyzed and showed that raw (uncompensated) PAT images measure the ratio of Cuso4/Total dye concentration with an error of up to 56% compared to the ground truth. Once optical compensation was applied Cuso4/Total Dye concentration was shown to have that the error was significantly reduced to ~4-5%. Marking a large improvement in the accuracy of linear unmixing and holding promise for this approach in improving results for future ex vivo and in vivo studies