Artefact Correction in MRI Temperature Mapping

Abstract

The thesis describes techniques that can be used to reduce or eliminate artefacts arising due to magnetic field disturbances and/or motion in temperature maps obtained by the Proton Resonance Frequency Shift (PRFS) method of MRI temperature mapping. The first chapter describes a hardware modification, which enables PRFS temperature mapping during percutaneous transluminal angioplasty performed using a hot balloon. It is proposed to separate in time media injection and heating by first inflating a balloon with a media at an initial temperature, and then by heating the media up using laser light. The separation is shown to eliminate flow, motion and susceptibility-redistribution artefacts and to enable real-time MRI temperature mapping using the PRFS technique. The second chapter describes a post-processing technique, which eliminates temperature artefacts caused by breathing. A set of baseline images characterizing Respiratory Induced Resonance Offset (RIRO) at a variety of respiratory cycle instants is acquired before the thermal treatment starts. During the treatment, the temperature evolution is found from two successive images. Then, the calculated temperature changes are corrected for the additional contribution caused by RIRO using the pre-treatment baseline images acquired at the identical instances of the respiratory cycle. The third chapter describes an acquisition and post-processing technique, which utilizes the fat signal to improve the immunity of the PRFS method of MRI temperature mapping against magnetic field disturbances. The described method is based on the utilization of several gradient and spin echoes acquired within one repetition interval with water and fat-selective scans. The fourth chapter describes an approach to post-processing PRFS temperature maps acquired during High Intensity Focused Ultrasound (HIFU) treatments, which allows the user to extract from the maps some additional information that can be used to monitor the treatment in a more safe and reliable way as well as to adjust the treatment in case it deviates from the desired scenario. A set of mathematical operations (including discrete analogues of time derivative, directional derivative, gradient and Laplacian) is applied to temperature maps to reveal the HIFU beam waist, focus, heat deposition site and their evolution during sonications. The location of the maximum temperature spot is identified and traced as well as its displacements are visualized and characterized. The resulting images reveal the information, which is usually hidden on the traditional temperature maps.