Microbubbles Ultrasound Super-Localization Imaging – MUSLI

Beyond their resonance and nonlinearity, microbubbles can also be pushed or destroyed by a sufficiently large ultrasound pulse transmission and undergo drastic changes in their scattering characteristics within milliseconds, as shown by ultrafast optical and acoustical imaging. This phenomenon is stochastic since each microbubble, being of different size and surrounded by a different environment, will respond to ultrasound differently. At appropriate transmit acoustical pressure, it could thus be possible to activate and detect bubbles one by one. Consequently, microbubbles could act as acoustically-activated sources which could then be exploited for ultrasound localization microscopy beyond the diffraction limit leading to sono-activated ultrasound localization microscopy, an acoustic analog of FPALM optical microscopy. However, to perform superlocalization the frame rate required to observe these fast events cannot be attained by conventional focused ultrasound (framerates < 100 Hz) as it becomes here mandatory to detect transient events at the millisecond scale over large regions of interest.

The following figure summarize our results in-vitro. The resolution was defined based on the Rayleigh criteria. In conventional focused imaging, two channels separated by 200 um could not be distinguished. However, the profile of the density of super-localized contrast agents demonstrated that these channels, which were separated by wavelength/4.5, could clearly be separated.
Beyond profiles, a full 3D mapping of the microbubble events were derived from the echoes acquired by the transducer matrix (bottom left). Using conventional focused imaging, the network of channels was impossible to observe since the channels were much smaller and less distant than the wavelength of the acoustic wave (2nd figure bottom). However, when distinct events derived by sono-activated ultrasound localization microscopy were plotted in space, the microfluidic system could be distinguished with an accuracy of wavelength/11. A map of the entire microfluidic system was created, which showed that microbubbles did not flow in all channels equally.

Now that the proof of concept in vitro has been performed, we will take every step that separate us from the clinical application of superlocalization. The first is to describe the theoretical limits of the method. Indeed, at present, there is no physical description of the resolution we could achieve. We should then solve the problem of motion sensitivity to perform the first superlocalization imaging in vivo on a chicken embryo. These studies will allow us to prepare for super-resolution imaging of the vascular system in the rat brain. Since the ultrasonic plane wave method and the microbubbles are clinically approved, then we could move quickly to the first studies in humans.

The first article entitled «Sono Ultrasound-Activated Localization Microscopy« was submitted to Applied Physics Letters in September 2013. It demonstrates the principle of superlocalization in-vitro. This paper is the first publication on the subject, and we want to make two other publications in the next year on, 1) theoretical and experimental limitations of the method 2) in vivo applications of the method.
The patent «Super-Resolution Ultrasonic« is still being examined in the USA.

Ultrasound imaging is performed by emitting acoustic pulses and analyzing the echoes of the structures encountered. The backscattered echoes are complex and result from the interference between all scatterers within a resolution cell. Because of this superposition, it becomes impossible to distinguish the structures inside this pixel. Like other wave processes, the maximum axial and lateral resolution is of the order of half the wavelength (lambda / 2). When the ultrasound is performed at 5 MHz, for example, the position of the target cannot be resolved within 150 micrometers. Due to this fundamental limitation of the physics of waves, we can get limited information on the microvasculature, which is often the key to understand and diagnose cancer and cardiovascular disease.

Microbubbles ultrasound super-localization imaging (MUSLI) would significantly improve the resolution of ultrasound imaging. Indeed, we propose to only activate a limited number of scatterers in order to distinguish the wavefront from a single target.This selective activation occurs due to the gradual destruction of microbubbles under the effect of ultrasound. The wave front is no longer a superposition of the echoes of several targets, its shape is precisely defined by the position of a punctual scatterer. Our hypothesis, in part verified by preliminary experiments, is that this position can be established within a few micrometers, far beyond the diffraction limit. If the scatterer is a single microbubble injected into the blood, its localization could create a map of the entire vasculature with a resolution of a few micrometers deep within tissue, representing a gain of a least a 100-fold in resolution.

The development of Microbubble ultrasound super-localization imaging (MUSLI) covers basic physics to in-vivo applications to disease. The first task is to describe the resolution limits of the technique, through in-vitro experimentations with ultrasound scanners and model microvessels. The passage to 3D is the second phase of the project. Since punctual events have to be intercepted in their entirety, large numbers of probes have to be used and highly parallel systems have to be developed. Before going to in-vivo experiment, motion compensation has to be added to the system to position precisely the disruption of the microbubbles. Finally, the last two tasks will be performed in small animal. Task 3 will consist in mapping the vasculature of healthy organs such as the kidney and the brain of a rodent. The goal is to add several branches on the vascular tree from other existing techniques. This map would then be compared to the imaging of the same organs with micro-CT and confocal microscopy, which have a high-resolution, but cannot be done in-vivo. Finally, task 4 will consist in confirming that the mapping of microvasculature in a tumor can provide additional information for diagnosis and drug monitoring. Angiogenesis will be measured with microscopy and compared to the MUSLI images.