Calendar

Abstract:

Virtual imaging trials (VITs) are computational frameworks to evaluate medical imaging systems, techniques, and protocols by use of anatomically and physiologically realistic numerical phantoms and high-fidelity models of the image formation process. By constructing digital twins of patient populations and imaging systems, they provide a unique opportunity to accelerate the clinical translation of emerging imaging modalities, such as ultrasound and photoacoustic computed tomography. In the early stage of development of a new imaging modality, VITs are an economical and convenient route to systematically explore imaging system designs and image reconstruction methods in a clinically meaningful way, even before building a physical prototype or recruit live participants. In this talk, I will present how we can use VITs to advance dynamic contrast enhanced multispectral photoacoustic computed tomography (DCE-MPACT), a promising technique to monitor disease progression and response to therapy by estimating changes in tumor vascular perfusion. Dynamic, anatomically realistic numerical mouse and lesion phantoms are established along with a multiphysics high-fidelity model of the DCE-MPACT image formation process. Using these phantoms, we develop, assess, and refine new optimization and learning-based image reconstruction methods for DCE-MPACT, which are then successfully demonstrated in-vivo to estimate tumor perfusion rates in a small animal model. I will also discuss opportunities and challenges of artificial intelligence (AI) in medical imaging and the role of VITs, engineering principles, and clinically meaningful objective assessment of image quality to develop trustworthy AI methods to inform medical decision. I will conclude the talk presenting my vision for promoting VITs and AI as transformative tools to bring promising new medical technologies closer to the clinic—faster and with fewer ethical or financial barriers.

PRESENTER INFORMATION