Motion model ultrasound localisation microscopy: Preclinical findings and clinical challenges

Abstract

Superresolution ultrasound represents a major breakthrough for noninvasive tissue characterization. It mainly comprises ultrafast and motion model driven approaches. This talk will focus on motion model Ultrasound Localisation Microscopy (mULM), an approach working at frame rates of most clinical ultrasound devices. It realizes vascular tracks by subwavelength localization of microbubbles and estimating their probability to have moved from one voxel to another. mULM reliably depicts vascular pattern of different tumors and provides functional vascular characteristics that cannot be obtained by other imaging modalities. We will also report on our experiences in clinically translating mULM and applying it to track breast cancer response to neoadjuvant chemotherapy. Currently, we are exploring mULM for other indications, i.e., chronic kidney diseases preclinically and clinically. Here, it became obvious that microbubble dose and injection speed must be carefully adapted to each application to ensure that vascular architectures are sufficiently displayed and to avoid microbubble overload in the images, which complicates correct assignments of microbubbles to tracks. Motion compensation is another challenge for clinical translation, potentially demanding 3D transducers. Thus, we are highlighting the significant potential of superresolution ultrasound and pinpointing challenges that need to be addressed to make it a robust and reliable clinical tool.