Towards improved treatment planning for head and neck microwave hyperthermia
MetadataShow full item record
This item's downloads: 240 (view details)
Hyperthermia is an emerging cancer treatment modality which involves applying heat to the malignant tumor. The heating can be delivered using electromagnetic energy, mostly in the radiofrequency or microwave range. Accurate patient-specific hyperthermia treatment planning is essential for effective and safe treatment, in particular for deep and loco-regional hyperthermia. An important aspect of hyperthermia treatment planning is the ability to focus microwave energy and heating into the tumour while reducing the occurrence of hotspots in surrounding healthy tissue. Typically, a multi-element antenna phase array hyperthermia system is used to focus the electromagnetic waves at the target region. This thesis presents methods for optimising the specific absorption rate distribution and resulting temperature distribution for head and neck cancer hyperthermia treatment. Several optimisation algorithms and objective functions have been evaluated to optimise the antenna amplitudes and phases of the hyperthermia systems. Evolutionary optimisation algorithms have been considered in this thesis and compared with a particle swarm optimisation method already in clinical use for the treatment of head and neck cancers. A differential evolution algorithm is proposed to improve target coverage. The differential evolution algorithm is shown to offer improved performance compared to the particle swarm optimisation algorithm. Most optimisation techniques reported in literature use static antenna settings throughout the treatment; however, in this thesis a dynamic approach is investigated. A time-multiplexed hyperthermia strategy is developed in order to better focus heating on the tumour while preserving predetermined areas in the healthy tissue. First, a multi-objective genetic algorithm is introduced, which generates multiple antenna settings which are applied sequentially. Thermal simulations are used to evaluate the performance of time-multiplexed steering. The results demonstrate the ability to enhance target heating while reducing hotspot temperatures. Finally, the time-multiplexing steering is evaluated against thermal tissue properties variation and is shown to be robust to temperature dependent thermal tissue properties.