| 总结: | <p>Radiation therapy plays a crucial role in the treatment of cancer. Major developments in the field, including 3D tomographic imaging, intensity modulation and novel particle modalities, have all contributed to iterative improvements in precise dose delivery. These advances have enabled highly conformal tumouricidal dose deposition with improved healthy tissue sparing. Proton therapy, in particular, has become an increasingly adopted modality due to its favourable depth-dose profile, allowing for much of the dose to be localised within the target region, with virtually no dose leakage to tissues beyond the target. However, physical dose does not tell the full story; radiobiological mechanisms and their intermodality differences must be well-understood, as these explain the processes by which the energy deposited in cells leads to cell death and, ultimately, to clinical outcomes.</p>
<p>Densely ionising particle tracks offer an enhanced cell killing efficiency over x-rays, which may be quantified through the relative biological effectiveness (RBE). The RBE is a multiplicative factor converting the proton dose to an x-ray dose equivalent that yields the same biological endpoint. In standard clinical practice a constant RBE of 1.1 is used, assuming protons are 10% more effective than x-rays for the same dose, though recently there has been an increased focus on more sophisticated, variable RBE models. Many decades of clinical experience with x-rays enables RBE-weighted proton dose planning under a familiar framework. The dose is shaped according to both absolute and volume-based tolerance doses. Although, in many modern treatment planning systems, dose-volume constraints (DVCs) are approximated as their exact formulation poses a non-convex optimisation problem.</p>
<p>This thesis proposes novel methods and recommendations for the inclusion of radiobiological factors in treatment planning through (1) a variable but pragmatic RBE model based on DNA double strand break induction, and (2) a flexible, projection-based inverse planning algorithm, suited to non-convex settings, that comprehensively addresses dose-volume effects through the exact modeling of DVCs. This author hypothesises that inclusion of these strategies will enable confident and effective treatment planning not only in terms of physical dose deposition but, crucially, in biological effect.</p>
|
|---|