Sviluppo di strumenti computazionali per il miglioramento dell'efficienza della radioterapia

Modern intensity-modulated radiotherapy (IMRT) techniques including fixed-gantry IMRT, VMAT and Helical Tomotherapy allow for the delivery of treatments with high levels of target dose conformity while sparing the surrounding healthy tissues at the same time, ultimately reducing the risk for side effects. On the one hand, implementing IMRT treatments is a sophisticated process based upon the need to solve an optimization problem known as inverse-planning aimed at finding the optimal beam modulation for the achievement of the prescribed dose objectives. Typical inverse-planning problems involve the optimization of cost-functions of hundreds to thousands of variables. Therefore, the efficiency of this process is highly constrained by the capability of the used algorithms to explore the cost function's landscape efficiently when looking for an optimal solution. On the other hand, the use of intensity-modulated beams possibility introduces factors that might affect the dosimetric and delivery accuracy of the resulting plan either because of inaccuracies in the dose calculation occurring in the treatment planning system (TPS) or errors during treatment delivery. In literature, the collection of such factors is referred to as plan complexity. Thereby, prior to treatment delivery to the patient, each plan is delivered onto a system of dose detectors to check for its dosimetric and delivery accuracy. However, this process which is known as patient-specific quality assurance (PSQA), highly impacts the overall efficiency of the radiotherapy workflow since it requires considerable time and the involvement of several resources. It turns out that, with the aim of improving the overall efficiency of the radiotherapy workflow, the development and use of new tools are worth being investigated. The topic of this thesis falls within this context. On the one hand, new methods based on complexity metrics are proposed as possible support tools for the characterization of the dosimetric and delivery accuracy of radiotherapy treatment plans. A novel software package called UCoMX is presented, which is the first tool for the extraction of complexity metrics from IMRT/VMAT and Helical Tomotherpay plans made freely available. Moreover, three original investigations aimed at validating the use of complexity indicators within the PSQA process are presented, showing that they would allow for a 50% reduction of the PSQA workload at the center where this research has been carried out. In the near future, the use of such tools could become an effective alternative to standard measurement-based PSQA procedures. On the other hand, on an independent research line, the use of quantum computers for the optimization of the fluence maps of fixedgantry IMRT plans is proposed. Quantum computers are under intense development by several research groups and companies worldwide and might lead to a revolution in the field of computation. Therefore, their use in the inverse-planning process might lead to unprecedented benefits in terms of efficiency. In this work, two original investigations that explored the feasibility of using quantum computers in radiotheray are proposed through the definition of novel strategies for the reformulation of the inverse-planning problem in a form compatible with the features of quantum hardware. The ultimate aim of the present doctoral project was to put the basis for the future adoption of the aforementioned tools in clinics, by investigating both their potentiality and limitations.