Développement de la microscopie de localisation ultrasonore (ULM) volumétrique et transcrânienne adaptée pour l'humain

Abstract

A stroke is a disruption of blood circulation in the brain. This rupture can have different origins, and brain imaging is necessary to determine the cause. This imaging must be fast, easy to access, and sufficiently precise to allow diagnosis. Ultrasound is currently not recommended for diagnosis because of its inability to image the brain's blood vessels with sufficient resolution. Nevertheless, ultrasound could be an ideal candidate for stroke diagnosis due to its cost and portability. Ultrasound localization microscopy (ULM) allows imaging of the microcirculation of blood vessels by transcending the diffraction limit. This method consists of localizing and tracking microbubbles injected into the blood. A recent study using a sequence carried out with a multiplexed matrix probe showed that volumetric ULM made it possible to differentiate the cause of stroke in rats. However, imaging the human brain differs from that of the rat brain because of its size and the thickness of the skull. The objective of this thesis is to adapt the volumetric ULM sequence that has been used in rats to image the human brain. The first part deals with the modifications of the sequence to image human brain vessels while maintaining a sufficient volume rate and signal-to-noise ratio for the ULM algorithms. A sequence has thus been proposed to allow this volumetric ULM imaging through the skull. Then, in a second part, this sequence was tested in a preclinical setting on sheep. These experiments on sheep, in agreement with an ethical protocol, consisted of imaging the brain through the skull using the ULM sequence and comparing the results with MRI imaging. In addition, the sheep's skull was also studied to quantify the difference between the model and the human. Finally, in the last part, corrections of the skull's aberrations were pursued to allow an improvement of the transcranial 3D ULM.

Ähnliche Arbeiten