Preview

PDF, English Download (30MB) | Terms of use

Abstract

The aim of this work is to develop a compressed sensing (CS) based optimization for in- verse treatment planning in radiation therapy. This approach is applied to the example of brachytherapy where fast optimization is essential during intra-operative treatment planning. In this thesis, a novel approach is presented that allows real-time intra-operative plan- ning for the first time. The standard inverse treatment problem is reformulated to resemble a CS problem. Highly specific solvers are developed and incorporated into a novel treatment planning system. By incorporating biological models and clinically important dosimetric criteria, the objective functions become more realistic. Being approximately two orders of magnitude faster than state-of-the-art methods, the CS based approach is proven to return the same or better quality in plans. The inherit sparsity approach in CS allows to decrease the amount of needles by up to 25% reducing the intervention time and the probability of side effects. In addition, the new objective functions further simplify the treatment planning. The novel CS based strategy and solvers can also be applied to other modalities in radi- ation therapy, e.g. intensity-modulated radiation therapy. The incorporation of sparse solution is a novel and promising paradigm for optimization in medical physics.