Spatiotemporal optimering av behandlingsplaner för Gammakniven
Year: 2025
Type: dissertation
Abstract: The Leksell Gamma Knife performs highly accurate stereotactic radiosurgery for brain tumors. Gamma Knife treatments have traditionally been performaed in single sessions, but for treating larger targets, a fractionation approach could be useful, due to the better cell repair capacity between fractions in healthy tissue in tumors. The purpose of this work is to investigate non-uniform fractionation of Gamma Knife treatments, with foxus on the mathematical optimization of the treatment planning. The motivation for allowing the fractions to be independent is the poteintially reduced daamage to healthy tissue in comparison with uniformly fractionated treatments. With non-uniformly fractionated treatments, we have to take the spatio-temporal effects into account in the optimization, which raises the need for using biolagically based measures such as biologically effective dose (BED), rather than physical dose. The correlation between these two quantities is non-linear, but by linearizing the BED, the treatment planning optimization problem can be formulated as a linear program (LP), for which we have familiar solving methods. Two different LP approaches are proposed, with partly different penalization objects. The implemented models take as input the sector and collimator settings from already optimized single-session plans, and are thus meant to serve as a post-processing step in the treatment planning. Both models manage to reduce the mean BED in organs at risk (OARs) for some of the patient cases tested, compared to univormly fractionated reference plans with the same BED target coverage. For one of the models, also the maximum OAR BED is reduced. However, the performance of the models depends on the properties of the specific tumors and OARs.
Author: Forsberg, Erik
Language: English
Subfield: Radiation
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7134101305
