External Beam Radiation Therapy (EBRT) is a common treatment option for patients with prostate cancer. When treating the prostate with EBRT, a geometric volume (PTV margin) is added around the prostate to account for uncertainties in treatment planning and delivery. Current methods for estimating PTV margins rely on the analysis of population-based inter- and intra-fraction motion data. These methods do not consider the patient-to-patient differences in demographic or clinical presentation of patient specific factors (PSFs), such as age, weight, body-mass index, health and performance status, prostate-specific antigen levels, Gleason scores, presence of bowel problems, or other health conditions. The purpose of this thesis is to investigate the feasibility using regression-based predictive algorithms to predict the extent of prostate motion for the purpose of personalizing the PTV margin using PSFs as inputs. Benchmarking simulations of Linear, Ridge, LASSO, SVR, kNN, and MLP algorithms were performed by simulating prostate intra-fraction motion and realistic variations in PSFs. Sample sizes ranged from n=20 to 800, with varying levels of noise into the motion data (0-10mm). Leave-one-out cross validation was used to train and validate algorithm performance. The results suggest that algorithm performance improves significantly within the first 50 – 100 patients, and this rate of improvement is independent of noise in prostate motion. The Ridge regression algorithm predicted intra-fraction motion to the lowest mean absolute error in simulated motion, performing especially well in small datasets. To evaluate the clinical utility of this approach, pre- and post-treatment prostate motion data, treatment time data, and rectal distension data was recorded in 21 patients, along with a variety of PSFs. In the analysis of patient data, the LASSO algorithm out-performed the Ridge algorithm, predicting the mean and standard deviation of an individual prostate cancer patient’s intra-fraction motion to within 0.8mm and 0.4 mm mean absolute error, respectively. However, prostate motion predictions did not correlate with PSFs, possibly due to the small sample size. This work demonstrates the feasibility of using regression-based algorithms for predicting prostate motion, and hence the opportunity to personalize PTV margins in prostate cancer patients. / Graduate / 2018-08-22

En proton thérapie, la technique de balayage, permet de traiter efficacement le patient vis à vis de l'irradiation de la tumeur et la protection des tissus sains. Ces bénéfices dosimétriques peuvent cependant être grandement dégradés par les mouvements intra-fraction. Par conséquent, l'étude de méthodes d'atténuation ou d'adaptation est nécessaire. C'est pour cela, que nous avons développé un logiciel ”open-source” de calcul et d'évaluation de dose en 4D, MSPT (Motion Simulator for Proton Therapy), pour la technique de balayage. Son but est de mettre en avant l'impact des mouvements intra-fraction en calculant la répartition de dose dans le patient. En outre, l'utilisation de MSPT nous a permis de mettre au point et de proposer une nouvelle méthode d'atténuation du mouvement basée sur l'ajustement du poids du faisceau quand celui-ci balaye la tumeur. En photon thérapie, un enjeu principal pour les traitements délivrés à l'aide de collimateurs multi-lames (MLC) consiste à trouver un ensemble de configurations du MLC permettant d'irradier correctement la tumeur. L'efficacité d'un tel ensemble se mesure par le total beam-on time et le total setup time. Dans notre étude, nous nous intéressons à la minimisation de ces critères, d'un point de vue algorithmique, pour de nouvelles technologies de MLC: le MLC rotatif et le MLC à double couche. De plus, nous proposons un algorithme d'approximation pour trouver un ensemble de configurations minimisant le total beam-on time pour le MLC rotatif / In proton therapy, the delivery method named spot scanning, can provide a particularly efficient treatment in terms of tumor coverage and healthy tissues protection. The dosimetric benefits of proton therapy may be greatly degraded due to intra-fraction motions. Hence, the study of mitigation or adaptive methods is necessary. For this purpose, we developed an open-source 4D dose computation and evaluation software, MSPT (Motion Simulator for Proton Therapy), for the spot-scanning delivery technique. It aims at highlighting the impact of intra-fraction motions during a treatment delivery by computing the dose distribution in the moving patient. In addition, the use of MSPT allowed us to develop and propose a new motion mitigation strategy based on the adjustment of the beam's weight when the proton beam is scanning across the tumor. In photon therapy, a main concern for deliveries using a multileaf collimator (MLC) relies on finding a series of MLC configurations to deliver properly the treatment. The efficiency of such series is measured by the total beam-on time and the total setup time. In our work, we study the minimization of these efficiency criteria from an algorithmic point of view, for new variants of MLCs: the rotating MLC and the dual-layer MLC. In addition, we propose an approximation algorithm to find a series of configurations that minimizes the total beam-on time for the rotating MLC