High Intensity Focussed Ultrasound (HIFU) ablation is a promising technology for the non-invasive, targeted treatment of certain types of cancer. The technique functions by subjecting tumours to a cytotoxic level of intense, localised heating, while leaving the surrounding tissue unharmed. However, a number of limitations in the available HIFU treatment monitoring methods are currently hampering the effectiveness and clinical adoption of the therapy. This work aims to develop improved metrics of HIFU-induced biological damage that are specifically suited to monitoring and controlling HIFU ablation. Firstly, an optical method that enables straightforward quantification of thermal damage in protein-embedding hydrogels is developed. Secondly, hydrogels embedded with different cell lines are used to assess the performance of common temperature-based metrics of cell death across a range of HIFU-relevant conditions. Finally, a novel, passive acoustic detector designed for the real-time monitoring of HIFU-induced tissue damage is proposed. The detector is shown to predict lesioning with over 80% accuracy in regimes that are very likely to create lesions (60 J of acoustic energy or more), with an error rate of less than 6% for exposures that are too short to cause lesioning (up to 1 s long). The proposed detector could therefore provide a low-cost means of effectively monitoring clinical HIFU treatments passively and in real time.

Side effects of current chemotherapeutics limit their use in cancer therapy. Although many current drugs are highly toxic and potent, the effects they have on non-cancerous tissue are unbearable for patients. Targeting these drugs may provide a means to restrict their toxic effects to only cancer tissue while leaving healthy tissue unaffected. This approach requires that the drug is only available in cancer tissue, which has been achieved here by encapsulating drugs into liposomal nano-capsules which are capable of passively accumulating in cancerous tissue via the enhanced permeability and retention effect (EPR). In addition to localisation, a threshold dose must be achieved to deliver the desired toxic effect to the target tumour tissue. Previous strategies have relied on passive 'leaching' of the drug from liposomes, however this 'leaching' does not necessarily achieve the threshold dose required. In the present work, a new generation of liposomes has been developed whereby release is solely achieved in the presence of ultrasound triggered cavitation. Instigation of such cavitation events would normally require the target tissue be exposed to high and possibly damaging ultrasound pressures. To remove the need for these high pressures, cavitation nuclei have been developed to lower the cavitation threshold of surrounding media. To allow for improved co-localisation and treatment deeper into cancer tissue, cavitation nuclei were developed to be in the nanoscale size range. Two types of novel cavitation nuclei were produced, a rough surfaced carbon nanoparticle (CNP, ~180 nm) and smooth shaped polymeric nano-cup particle (NC, ~150, 470, or 770 nm). Both types of particle are solid nanoparticles with gas entrapped on their surface which was capable of cavitating in response to ultrasound without greatly affecting the particle itself. These particles are classified as cavicatalytic nanoparticles due to their ability to reduce the cavitation threshold of their surrounding media without being destroyed themselves. Finally, an entirely nanoscale release system was developed and tested in vitro and in vivo. The drug carrier (the liposome) and effector agent (the cavicatalytic nanoparticle) were used to demonstrate ultrasound triggered drug release, specifically in response to the generation of cavitation events. These cavitation events could be non-invasively monitored and characterised, adding to the potential clinical utility of the technologies developed and described here.

A la croisée des chemins entre médecine et physique des ultrasons, ce travail de thèse s’est intéressé à l’apport de solutions diagnostiques et de thérapeutiques novatrices dans le domaine de l’électrophysiologie cardiaque. Un système capable de délivrer des ultrasons focalisés dans le cœur par voie transoesophagienne sous guidage par ultrasons a été développé et testé in vivo chez 6 porcs. Les tirs HIFU ont été délivrés sur les oreillettes et les ventricules. Lors de l'autopsie, une analyse visuelle a démontré la présence de lésions thermiques dans les zones ciblées chez 3 animaux. Ces lésions ont été confirmées par analyse histologique (taille moyenne: 5,5 mm2 x 11 mm2). Aucune lésion thermique œsophagienne n'a été observée. Un animal a présenté une bradycardie due à un bloc auriculo-ventriculaire, ce qui a permis de confirmer une réelle interaction entre les tirs HIFU et le tissu nodal cardiaque. Nous avons cependant observé un manque de précision, principalement lié aux mouvements cardiaques ainsi qu’aux structures anatomiques situées entre les zones ciblées et le transducteur de thérapie. Ces difficultés ont été principalement reliées à l’anatomie du modèle porcin, loin de celle de l’être humain. La recherche d'un meilleur modèle a conduit à des tests d'imagerie concluants sur des babouins.Des expériences supplémentaires ont été conduites afin d'améliorer la cartographie des arythmies ventriculaires et le suivi de la formation de lésions pendant l'ablation.Des expériences ont été menées sur les ventricules gauches de quatre coeurs de porcs en mode travaillant. Le protocole visait à démontrer que différents modèles d'activation mécanique pouvaient être observés, que le ventricule soit en rythme sinusal, stimulé depuis l'épicarde ou l'endocarde. Des acquisitions d’imagerie de déformation électromécanique (EWI) ont été enregistrées sur les faces antérieures, latérales et postérieures du ventricule gauche. Les boucles ont été ensuite analysées à l’aveugle par deux lecteurs indépendants.Les interprétations des séquences EWI étaient correctes dans 89% des cas. Le taux de concordance globale entre les deux lecteurs était de 83%. Dans un ventricule stimulé, l'origine du front d'onde était focale et provenait de l'endocarde ou de l'épicarde stimulé. En rythme sinusal, le front d'onde était activé depuis tout l'endocarde, en direction de l'épicarde, à une vitesse de 1,7 ± 0,28 m.s-1. Les vitesses du front d'onde ont été mesurées respectivement lorsque l'endocarde ou l'épicarde étaient stimulés à une vitesse de 1,1 ± 0,35 m.s -1 et 1,3 ± 0,34 m.s-1 (p = NS). Nous avons aussi démontré sur des échantillons ex-vivo que l'imagerie trans oesophagienne par analyse des ondes de cisaillement (élastographie) pouvait cartographier l'étendue des lésions HIFU. Des tirs HIFU ont été réalisés à l'aide de la sonde trans oesophagienne sur des échantillons de blancs de poulet (n = 3), puis sur un modèle porcin ex vivo d'oreillette (gauche, n = 2) et de ventricule gauche (n = 1). L’élastographie a fourni des cartes de rigidité des tissus avant et après l'ablation. Les zones des lésions ont été obtenues par analyse et quantification des changements de couleur des tissus puis ont été comparées aux images par élastographie. Dans le blanc de poulet, la rigidité est passée en moyenne de 4.8±1.1 kPa à 20.5±10.0 kPa (ratio 5.0±3.2). Dans le ventricule gauche, la rigidité est passée en moyenne de 21.2±3.3kPa à 73.8±13.9kPa (ratio 3.7±1.2). Dans l’oreillette gauche, la rigidité est passée en moyenne de 12.2±4.3 kPa à 30.3±10.3 (ratio 3.2±2.0). En histologie, la taille des lésions variait de 0.1 à 1.5 cm2 dans la zone du plan d'imagerie. Les caractéristiques morphométriques étaient similaires entre histologie et élastographie / At the crossroads of medicine and physics, this work aimed to provide innovative diagnostic and therapeutic tools based on ultrasound, in the field of cardiac electrophysiology. A system capable of delivering HIFU into the heart by a transesophageal route using ultrasound (US) imaging guidance was developed and tested in vivo in six male pigs. HIFU exposures were performed on atria and ventricles. At the time of autopsy, visual inspection identified thermal lesions in the targeted areas in three of the animals. These lesions were confirmed by histologic analysis (mean size: 5.5 mm2 x 11mm2). No esophageal thermal injury was observed. One animal presented with bradycardia due to an atrio-ventricular block, which provides real-time confirmation of an interaction between HIFU and the electrical circuits of the heart. There was still a lack of accuracy, mainly related to cardiac motion, and to anatomical structures in between the targets and the transducer. It was mainly related to the in vivo model and its anatomy, far from the human’s. The search for a better model led to conclusive imaging tests on baboons. Additional experiments were conduced in order to improve the mapping of ventricular arrhythmias and the monitoring of lesion formation during ablation. First, experiments were conducted on left ventricles of four isolated working mode swine hearts. The protocol aimed at demonstrating that different patterns of mechanical activation could be observed whether the ventricle was in sinus rhythm, paced from the epicardium, or from the endocardium. Electromechanical wave imaging (EWI) acquisitions were recorded on the anterior, lateral, and posterior segments of the left ventricle. Loop records were blindly assigned to two readers. EWI sequences interpretations were correct in 89% of cases. The overall agreement rate between the two readers was 83%. When in a paced ventricle, the origin of the wave front was focal and originating from the endocardium or the epicardium. In sinus rhythm, wave front was global and activated within the entire endocardium towards the epicardium at a speed of 1.7±0.28 m.s-1. Wave front speeds were respectively measured when the endocardium or the epicardium were paced at a speed of 1.1 ± 0.35 m.s-1 vs 1.3±0.34 m.s-1 (p=NS). Lastly, we investigated the feasibility of a dual therapy and imaging approach with the same transoesophageal device. We demonstrated on ex-vivo samples that transoesophageal shear wave imaging (SWE) can map the extent of the HIFU lesions. HIFU ablation was performed with the transoesophageal probe on ex-vivo chicken breast samples (n=3), then atrium (left, n=2) and ventricle (left n=1, right n=1) of swine heart tissues. SWE provided stiffness maps of the tissues before and after ablation. Areas of the lesions were obtained by tissue color change with gross pathology and compared to SWE. Shear modulus of the ablated zones increased from 4.8±1.1 kPa to 20.5+/-10.0 kPa (ratio 5.0±3.2) in the chicken breast, from 12.2±4.3 kPa to 30.3±10.3 (ratio 3.2±2.0) in the atria and from 21.2±3.3kPa to 73.8±13.9kPa (ratio 3.7±1.2) in the ventricles. On gross pathology, the size of the lesions ranged from 0.1 to 1.5cm2 in the imaging plane area and morphometric characteristics were fitting with elasticity-estimated depths and widths of the lesions