The ERC Consolidator Grant ResolveStroke built upon our foundational work in Ultrasound Localization Microscopy (ULM), which initially demonstrated 2D super-resolution imaging of the rat brain using microbubble tracking with micrometric spatial and microsecond temporal resolution. The overarching goal of the ERC Consolidator project ResolveStroke was to advance ULM into a portable 3D scanner capable of rapid stroke diagnosis, specifically to distinguish between ischemic and hemorrhagic strokes and enable faster triage and thrombolytic treatment.
Through this project, we transformed ULM from a 2D technique, originally limited to healthy rats without a skull, into a 3D clinical imaging modality suitable for transcranial and abdominal applications.
We generalized ULM to 3D, enabling precise quantification of microbubble movement and amplitude, which allowed for detailed mapping of microvascular flow and structure. To address the computational bottleneck, we developed algorithms that reduced ULM processing times from 12 hours to just 3 minutes in 2D, a critical step toward real-time clinical feasibility. Additionally, we systematically identified the physical factors limiting ULM resolution, optimizing imaging parameters for both preclinical and clinical settings.
Preclinical studies were expanded to include 3D ULM imaging in animal models, where we developed biomarkers to differentiate between ischemic and hemorrhagic stroke types. A major technical achievement was the successful demonstration of 3D ULM through a sheep skull, which validated the feasibility of transcranial imaging in large animals and paved the way for human applications.
Clinical translation was a central focus of ResolveStroke. We performed 2D ULM imaging in moyamoya patients and completed 3D ULM clinical imaging in 19 ischemic stroke patients. A dedicated human stroke scanner was developed and deployed in a clinical study, with regulatory approval from ANSM.
To ensure reproducibility and foster collaboration, we distributed all 2D ULM codes, along with datasets and comparative metrics, to the global research community. To support broader adoption, we also published and distributed 3D ULM codes alongside a study demonstrating 3D ULM in awake mice.
The project also led to the launch of the ResolveStroke startup in 2022, co-founded with Vincent Hingot and Aritz Zamacola, which now employs 15 people and focuses on commercializing 3D transcranial contrast-enhanced ultrasound for cerebrovascular disease diagnosis. Further clinical applications were explored through two ERC Proof-of-Concept grants: StrokeMonitor (2022), which implemented ULM for detecting delayed cerebral ischemia in intensive care units, with a study completed on 13 patients and KidneyScope (2024), which will extend 3D ULM to glomerular imaging in chronic kidney disease and kidney transplants.
Additionally, we applied 2D ULM in neonates with strokes and arteriovenous malformations, in a collaboration with the team of F. Knieling in Erlangen.
Beyond neuroimaging, ResolveStroke expanded ULM’s applications to renal imaging, where we achieved the first visualization of glomeruli in human kidneys using clinical ultrasound scanners. This breakthrough allowed for the non-invasive assessment of glomerular density and microvascular architecture, offering a potential diagnostic tool for chronic kidney disease and kidney transplant monitoring.
Collectively, these advancements established ULM as a robust, clinically translatable imaging modality, setting the stage for its broader application in neurovascular and abdominal imaging.
