Changes in cerebral blood flow are associated with stroke, aneurysms, vascular cognitive impairment, neurodegenerative diseases and other pathologies. Brain angiograms, typically performed via computed tomography or magnetic resonance imaging, are limited to millimetre-scale resolution and are insensitive to blood-flow dynamics. Here, we show that ultrafast ultrasound localization microscopy of intravenously injected microbubbles enables transcranial imaging of deep vasculature in the adult human brain at microscopic resolution and the quantification of haemodynamic parameters. Adaptive speckle tracking to correct for micrometric brain-motion artefacts and for ultrasonic-wave aberrations induced during transcranial propagation allowed us to map the vascular network of tangled arteries, to functionally characterize blood-flow dynamics at a resolution of up to 25 μm, and to detect blood vortices in a small deep-seated aneurysm in a patient. Ultrafast ultrasound localization microscopy may facilitate the understanding of brain haemodynamics and of how vascular abnormalities in the brain are related to neurological pathologies.

Medical ultrasound is a widely used diagnostic imaging technique for tissues and blood vessels. However, its spatial resolution is limited to a sub-millimeter scale. Ultrasound Localization Microscopy was recently introduced to overcome this limit and relies on subwavelength localization and tracking of microbubbles injected in the blood circulation. Yet, as microbubbles follow blood flow, long acquisition time are required to detect them in the smallest vessels, leading to long reconstruction of the microvasculature. The objective of this work is to understand how blood flow limits acquisition time. We studied the reconstruction of a coronal slice of a rat’s brain during a continuous microbubble injection close to clinical concentrations. After acquiring 192000 frames over 4 minutes, we find that the biggest vessels can be reconstructed in seconds but that it would take tens of minutes to map the entire capillary network. Moreover, the appropriate characterization of flow profiles based on microbubble velocity within vessels is bound by even more stringent temporal limitations. As we use simple blood flow models to characterize its impact on reconstruction time, we foresee that these results and methods can be adapted to determine adequate microbubble injections and acquisition times in clinical and preclinical practice.

Because it drives the compromise between resolution and penetration, the diffraction limit has long represented an unreachable summit to conquer in ultrasound imaging. Within a few years after the introduction of optical localization microscopy, we proposed its acoustic alter ego that exploits the micrometric localization of microbubble contrast agents to reconstruct the finest vessels in the body in-depth. Various groups now working on the subject are optimizing the localization precision, microbubble separation, acquisition time, tracking, and velocimetry to improve the capacity of ultrasound localization microscopy (ULM) to detect and distinguish vessels much smaller than the wavelength. It has since been used in vivo in the brain, the kidney, and tumors. In the clinic, ULM is bound to improve drastically our vision of the microvasculature, which could revolutionize the diagnosis of cancer, arteriosclerosis, stroke, and diabetes.

Ultrasound Localization Microscopy (ULM) is an ultrasound imaging technique that relies on the acoustic response of sub-wavelength ultrasound scatterers to map the microcirculation with an order of magnitude increase in resolution as compared to conventional ultrasound imaging.Initially demonstrated in vitro, ULM has matured and sees implementation in vivo for vascular imaging of organs or tumors in both animal models and humans. The performance of localization algorithms greatly defines the quality of vascular mapping. Here, we compiled and implemented a collection of ultrasound localization algorithms and devised three in silico and in vivo datasets to compare their performance through 11 metrics. We also present two novel algorithms designed to increase speed and overall performance. By providing a comprehensive open package to perform ULM that includes localization algorithms, the datasets used, and the evaluation metrics, we aim to equip researchers with a tool to identify the optimal localization algorithm for their 2 application, benchmark their own software and enhance the overall quality of their ULM images while uncovering the algorithms' own limits. MainThe circulatory system carries the essential nutrients of life to cells in the body. It forms a 100,000 kilometer-long network composed of centimeter-wide arteries down to capillaries that are a few micrometers in diameter at most. The study of the vascular system is essential for both the diagnosis and treatment of cardiovascular diseases, cancer, diabetes, stroke, or organ dysfunction.Due to its diversity of scale, imaging the vasculature is a daunting task and few techniques are capable of measuring micro-hemodynamics deep in the human body.Ultrasound imaging is extensively used in medical practice as a non-invasive tool that provides soft-tissue diagnosis, prognosis, or guides interventions. Using the Doppler effect, it can also measure blood flow in real-time. With the advent of plane wave techniques 1 , ultrasound has reached frame rates up to 20 kHz making it possible to observe and measure fast occurring changes 1 such as functional changes in the brain 2 , as well as increase sensitivity to blood flow that allows more accurate filtering 3,4 . To increase blood's contrast, micrometric gas microbubbles can be intravenously injected in vivo. Contrast-Enhanced Ultrasound (CEUS) is mostly used for perfusion studies and cardiac imaging 5 .Because conventional ultrasound imaging, Doppler, and contrast-enhanced ultrasound all rely on the propagation of sound waves, ultrasound imaging is largely limited in resolution by diffraction.Recently, Ultrasound Localization Microscopy (ULM) has broken that limit by isolating a small number of microbubbles as subwavelength sources in each image and localizing them with micrometric precision [6][7][8] . Similar to PALM (PhotoActivated Localization Microscopy), it uses

Ultrasound Localization Microscopy can map blood vessels with a resolution much smaller than the wavelength by localizing microbubbles. Current implementations of the technique are limited to 2-D planes or small fields of view in 3D. These suffer from minute-long acquisitions, out-of-plane microbubbles and tissue motion. In this study, we exploit the recent development of 4D ultrafast ultrasound imaging to insonify an isotropic volume up to 20000 times per second and perform localization microscopy in the three dimensions. Specifically, a 32x32 elements, 9-MHz matrix-array probe connected to a 1024-channel programmable ultrasound scanner was used to achieve sub-wavelength volumetric imaging of both the structure and vector flow of a complex 3D structure (a main canal branching out into two side canals). To cope with the large volumes and the need to localize the bubbles in the three dimensions, novel algorithms were developed based on deconvolution of the beamformed microbubble signal. For tracking, individual particles were paired following a Munkres assignment method and velocimetry was done following a Lagrangian approach. ULM was able to clearly represent the 3D shape of the structure with a sharp delineation of canal edges (as small as 230 µm) and separate them with a spacing as low as 52µm. The compounded volume rate of 500Hz was sufficient to describe velocities in the [2.5-150] mm/s range and to reduce the maximum acquisition time to 12s. This study demonstrates the feasibility of in vitro 3D ultrafast ultrasound localization microscopy and opens up the way towards in vivo volumetric ULM.

No abstract

In the field of ischemic cerebral injury, precise characterization of neurovascular hemodynamic is required to select candidates for reperfusion treatments. It is thus admitted that advanced imaging-based approaches would be able to better diagnose and prognose those patients and would contribute to better clinical care. Current imaging modalities like MRI allow a precise diagnostic of cerebral injury but suffer from limited availability and transportability. The recently developed ultrafast ultrasound could be a powerful tool to perform emergency imaging and long term follow-up of cerebral perfusion, which could, in combination with MRI, improve imaging solutions for neuroradiologists. Methods: In this study, in a model of in situ thromboembolic stroke in mice, we compared a control group of non-treated mice (N=10) with a group receiving the gold standard pharmacological stroke therapy (N=9). We combined the established tool of magnetic resonance imaging (7T MRI) with two innovative ultrafast ultrasound methods, ultrafast Doppler and Ultrasound Localization Microscopy, to image the cerebral blood volumes at early and late times after stroke onset and compare with the formation of ischemic lesions . Results: Our study shows that ultrafast ultrasound can be used through the mouse skull to monitor cerebral perfusion during ischemic stroke. In our data, the monitoring of the reperfusion following thrombolytic within the first 2 h post stroke onset matches ischemic lesions measured 24 h. Moreover, similar results can be made with Ultrasound Localization Microscopy which could make it applicable to human patients in the future. Conclusion: We thus provide the proof of concept that in a mouse model of thromboembolic stroke with an intact skull, early ultrafast ultrasound can be indicative of responses to treatment and cerebral tissue fates following stroke. It brings new tools to study ischemic stroke in preclinical models and is the first step prior translation to the clinical settings.

The resolution of an imaging system is usually determined by the width of its point spread function and is measured using the Rayleigh criterion. For most system, it is in the order of the imaging wavelength. However, super resolution techniques such as localization microscopy in optical and ultrasound imaging can resolve features an order of magnitude finer than the wavelength. The classical description of spatial resolution no longer applies and new methods need to be developed.In optical localization microscopy, the Fourier Ring Correlation has showed to be an effective and practical way to estimate spatial resolution for Single Molecule Localization Microscopy data. In this work, we wish to investigate how this tool can provide a direct and universal estimation of spatial resolution in Ultrasound Localization Microscopy. Moreover, we discuss the concept of spatial sampling in Ultrasound Localization Microscopy and demonstrate how the Nyquist criterion for sampling drives the spatial/temporal resolution tradeoff.We measured spatial resolution on five different datasets over rodent's brain, kidney and tumor finding values between 11 m and 34 m for precision of localization between 11 m and 15 m. Eventually, we discuss from those in vivo datasets how spatial resolution in Ultrasound Localization Microscopy depends on both the localization precision and the total number of detected microbubbles.This study aims to offer a practical and theoretical framework for image resolution in Ultrasound Localization Microscopy.