PROJECT SUMMARY/ABSTRACT Tumor vascular networks are highly abnormal and complex, vary significantly by tumor type and on a patient- by-patient basis. More importantly, patient-specific tumor vascular properties are known to regulate cancer progression and treatment response in patients with primary or metastatic liver cancers. Non-invasive rapid diagnostic methods to characterize unique tumor vascular properties in liver malignancies, and to provide clinical decision-support, are currently not available. Our goal is to introduce computational approaches to characterize complex attributes of vascular networks by considering the interconnected nature of voxels in 3D dynamic contrast-enhanced ultrasound (3D DCE-US), and to validate these as biomarkers for diagnostics or treatment monitoring. We have already contributed to pioneering 3D DCE-US liver imaging in the clinic and have demonstrated its ability to minimize 2D-based sampling errors and improved prediction of treatment response longitudinally. However, current quantification approaches are designed for 2D imaging and do not take into account vascular heterogeneity and the contrast flow field. In addition, conventional parameters that are typically extracted from averaged intensities in large regions of interest fail to account for spatial variations of perfusion common in complex tumor tissues. The exclusive volumetric nature of our data along with the intravascular nature of ultrasound contrast agents, has allowed us to demonstrate that additional information is encoded in spatial flow maps beyond convention, and that this information is sensitive to treatment response and representative of the underlying true vascular network. Thus, our hypothesis is that characterizing the interconnected nature of voxels in 3D DCE-US can capture new information from volumetric tissues for cancer applications, and beyond cancer. Therefore, we propose a set of specific aims designed to test this hypothesis by: i) further developing and ii) extensively validating new perfusion measurements to characterize different tumor vascular properties and detect subtle microvascular changes following therapy, and iii) clinically validating their clinical utility for diagnostics and treatment monitoring in real patients imaged with 3D DCE-US. Our proposal takes advantage of the recent clinical introduction of commercial ultrasound contrast agents and 3D imaging probes to advance non-invasive bedside liver imaging for capturing complex flow beyond convention.