Open Collections

Information fusion for prostate brachytherapy planning

University of British Columbia

Low-dose-rate prostate brachytherapy is a minimally invasive treatment approach for localized prostate cancer. It takes place in one session by permanent implantation of several small radio-active seeds inside and adjacent to the prostate. The current procedure at the majority of institutions requires planning of seed locations prior to implantation from transrectal ultrasound (TRUS) images acquired weeks in advance. The planning is based on a set of contours representing the clinical target volume (CTV). Seeds are manually placed with respect to a planning target volume (PTV), which is an anisotropic dilation of the CTV, followed by dosimetry analysis. The main objective of the plan is to meet clinical guidelines in terms of recommended dosimetry by covering the entire PTV with the placement of seeds. The current planning process is manual, hence highly subjective, and can potentially contribute to the rate and type of treatment related morbidity. The goal of this thesis is to reduce subjectivity in prostate brachytherapy planning. To this end, we developed and evaluated several frameworks to automate various components of the current prostate brachytherapy planning process. This involved development of techniques with which target volume labels can be automatically delineated from TRUS images. A seed arrangement planning approach was developed by distributing seeds with respect to priors and optimizing the arrangement according to the clinical guidelines. The design of the proposed frameworks involved the introduction and assessment of data fusion techniques that aim to extract joint information in retrospective clinical plans, containing the TRUS volume, the CTV, the PTV and the seed arrangement. We evaluated the proposed techniques using data obtained in a cohort of 590 brachytherapy treatment cases from the Vancouver Cancer Centre, and compare the automation results with the clinical gold-standards and previously delivered plans. Our results demonstrate that data fusion techniques have the potential to enable automatic planning of prostate brachytherapy.

Text

2016-06-20

Attribution 4.0 International

http://hdl.handle.net/2429/58305

Electrical and Computer Engineering

University of British Columbia

UBCV

http://creativecommons.org/licenses/by/4.0/