Phase-Changing Nanodroplets
Nanodroplets are a new class of ultrasound contrast agents designed to push imaging beyond its traditional limits. Unlike microbubbles, which are already in the gaseous state, nanodroplets are tiny liquid particles that can be vaporized on demand by focused ultrasound energy.
This ability to switch between liquid and gas makes them uniquely versatile: in their stable liquid form, they can circulate through the smallest capillaries; when activated, they expand into microbubbles that strongly scatter ultrasound waves, producing a localized signal exactly where and when it is needed.
Phase-changing nanodroplets open the door to targeted, functional imaging — for example, assessing blood flow, monitoring tissue perfusion, or even mapping the permeability of the gut or blood–brain barrier. Their controllable nature also makes them candidates for precise, image-guided drug delivery.
By combining nanodroplets with advanced beamforming and super-resolution imaging, we can explore physiology at scales and levels of control that were previously inaccessible — helping to reveal how disease alters microvascular function and how treatment might restore it.
Super-Resolution Ultrasound and the Heart
The heart’s microvasculature — the network of the smallest vessels that nourish cardiac tissue — has traditionally been invisible to conventional ultrasound. Super-resolution ultrasound overcomes this limit by tracking tiny contrast microbubbles as they move through the bloodstream, allowing us to visualize vessels only tens of micrometers wide.
By revealing how blood flows through these microvessels, super-resolution imaging provides new insight into how oxygen and nutrients reach heart muscle, how disease alters those pathways, and how therapies might restore them. Unlike MRI or CT, it is portable, radiation-free, and can be repeated safely, making it a promising tool for monitoring cardiac health at the bedside or in the clinic.
My work explores how this technology can move from laboratory demonstration to a practical, non-invasive method for studying and diagnosing cardiovascular disease — turning what was once invisible into actionable knowledge about the living heart.
Blood Flow Imaging and Atherosclerosis
Atherosclerosis — the buildup of plaque inside arteries — develops over years, guided not only by cholesterol and inflammation but also by the forces of blood flow acting on vessel walls. Regions exposed to disturbed or low shear stress are especially prone to plaque formation.
Ultrasound blood flow imaging allows us to measure these hemodynamic forces directly, non-invasively, and in real time. By combining high-frame-rate plane wave imaging with advanced flow reconstruction methods, we can map both the speed and direction of blood near the vessel wall. These measurements help identify areas where vascular stress may trigger inflammation or early disease — long before symptoms appear.
This approach offers a new perspective on cardiovascular health: understanding not just where disease occurs, but why. Through improved flow visualization, we move closer to predicting and preventing atherosclerosis rather than only treating its consequences.
3D Imaging of the Optic Nerve Sheath
The optic nerve sheath — the thin membrane surrounding the optic nerve — expands when pressure inside the skull rises. Measuring its diameter provides a valuable, non-invasive indicator of intracranial pressure, a critical parameter in conditions such as brain injury, stroke, or high-altitude cerebral edema.
Traditional ultrasound methods capture the optic nerve sheath in two dimensions, which can introduce measurement variability and limit diagnostic accuracy. To overcome this, I developed 3D freehand ultrasound imaging of the optic nerve sheath, allowing its full geometry to be reconstructed rather than estimated from a single slice.
This approach improves both precision and reproducibility by capturing subtle variations in sheath shape and volume. It also enables dynamic studies of how the sheath responds to changes in pressure, respiration, or environmental conditions — for example, during hypoxia, cold exposure, or altitude studies.
By advancing from 2D to true 3D visualization, this work moves us closer to reliable, bedside monitoring of brain pressure — safely, quickly, and without the need for invasive procedures.
Singular Value Decomposition in Blood Flow Imaging
Ultrasound data often contains a mixture of signals — from moving blood cells, vibrating tissue, and electronic noise — all overlapping in time and space. To reveal the subtle flow information hidden within, we use a mathematical technique called Singular Value Decomposition (SVD).
SVD separates the recorded ultrasound data into components based on how strongly and coherently they vary over time. Slow, correlated signals such as tissue motion can then be distinguished from the fast, random fluctuations of flowing blood. By filtering and reconstructing only the components linked to blood flow, we obtain clearer, more accurate vascular images without the need for contrast agents.
This approach forms the backbone of many advanced ultrasound methods, including vector flow imaging and super-resolution microscopy, where precision in separating flow and background is essential. In my work, I also study the potential artifacts that SVD filtering can introduce — knowledge that helps ensure these powerful algorithms are used transparently and reliably in clinical imaging.
Acoustic Wave Sparsely Activated Localization Microscopy (AWSALM)
AWSALM is a super-resolution ultrasound technique designed to image the smallest blood vessels deep within tissue — far beyond the limits of conventional ultrasound resolution. It works by sparsely activating tiny contrast agents, such as nanodroplets or microbubbles, using carefully timed acoustic waves. Each activation produces a brief, localized signal that can be precisely tracked and combined to reconstruct microvascular networks at a microscopic scale.
This sparse activation strategy dramatically reduces background noise and acquisition time compared to traditional super-resolution methods, while maintaining high spatial accuracy. It also allows selective imaging of specific vascular regions, making it well suited for studying organ function or disease progression.
To push the method further, we developed Fast-AWSALM, an optimized sequence and processing pipeline that achieves the same sub-diffraction resolution at higher frame rates. This improvement brings super-resolution ultrasound closer to real-time imaging, enabling dynamic visualization of blood flow and perfusion changes as they happen.
Together, AWSALM and Fast-AWSALM represent important steps toward practical, high-speed, high-resolution ultrasound — combining precision microscopy with the depth, safety, and accessibility of traditional ultrasound imaging.
The ULTRA-SR Challenge
The ULTRA-SR Challenge was the first international benchmarking initiative for super-resolution ultrasound imaging, created to evaluate and compare localization and tracking algorithms across the research community. Super-resolution ultrasound relies on detecting and reconstructing the precise positions of thousands of microscopic contrast agents, but until recently, there was no standardized way to assess the accuracy or robustness of different algorithms.
As one of the organizers of the ULTRA-SR Challenge at the IEEE International Ultrasonics Symposium 2022 in Venice, I helped develop a simulation and evaluation platform that generated realistic ultrasound data, incorporating contrast agent physics and flow dynamics. The challenge brought together 24 research groups from 10 institutions worldwide, each contributing their own reconstruction approaches to a shared dataset.
The results, published in IEEE Transactions on Medical Imaging, established a foundation for transparent, reproducible comparison in the field — accelerating progress toward clinically viable super-resolution imaging. Beyond the competition itself, ULTRA-SR fostered collaboration, open data sharing, and collective improvement across the ultrasound research community.