Preview

PDF, English Download (3MB) | Terms of use

Abstract

In this thesis, a number of new image-based techniques for the management of intrafractional motion during radiation therapy are presented. Intra-fractional motion describes all kinds of anatomy changes - most prominently respiration - that occur during a single treatment session. Spatially confining the radiation dose to the tumour tissue and thus sparing surrounding healthy tissue is assumed to be crucial for a successful treatment with limited side effects. Unfortunately, the delivery of dose distributions that are sharply confined to the tumour is greatly complicated by patient motion. If not accounted for, this motion will lead to a smearing out of the original dose distribution and will facilitate the redistribution of dose from tumour to healthy tissue. Possible technical solutions for this issue include the interruption of the radiation delivery if the tumour leaves a predefined spatial ‘window’, and the reshaping of the treatment field ‘on-the-fly’ to follow the tumour. Regardless of which delivery techniques is selected, the patient motion needs to be reliably detected in real-time to allow for an adaptation of the treatment delivery. First, we present experimental results for a novel x-ray imaging system that is attached to the treatment delivery device and enables us to continuously monitor the tumour motion during treatment delivery with sub-mm accuracy, a latency better than 90 ms, and a 7 Hz update rate. Second, we present a Monte Carlo simulation for an improved amorphous-silicon flat-panel detector that reduced treatment beam filtration by 60% and long-range MV-scatter by 80%. We conclude this thesis by presenting results of an experimental demonstration of a novel dose-saving actively-triggered 4d cone-beam computed tomography device.