A virtual imaging framework for DCE-PACT estimation of tumor perfusion

Dynamic contrast-enhanced photoacoustic computed tomography (DCE-PACT) noninvasively measures tumor perfusion, a key biomarker of disease progression and response to treatment. However, DCE-PACT imaging using commercially available volumetric PACT imagers is extremely challenging due to the rotating gantry design, which limits the number of tomographic view angles that can be acquired at each imaging frame. Moreover, the high dimensionality of spatiotemporal imaging imposes substantial computational and memory burden. Nuclear norm (NN) regularization has widely used to promote low-rank structure. However, it introduces strong bias and may suppress localized dynamics. To overcome this, we propose the use of the Schatten-p norm as image prior to promote a low-rank structure while preserving localized features. By use of a high-fidelity framework for DCE-PACT virtual imaging studies of small animal models, we systematically assess the proposed method. Results illustrate that the Schatten-p image prior yields sharper images, more accurate arterial input function and tumor activity curve estimation, and tumor perfusion rates than NN.