The state-of-the-art radiotherapy delivery treatments available today require that the intended dose distributions delivered are verified by volumetric gel dosimetry. The development of tissue equivalent gel dosimeters that provide an integrated assessment of the dynamic treatments, are primarily imaged with Magnetic Resonance Imaging (MRI). This thesis describes the dual development and assessment of an optical tomographic scanner and normoxic gel dosimeter. MRI centers are currently limited in both time and resources in providing the routine imaging necessary for treatment verification. An alternative bench top imaging modality has been designed, built and developed to provide both complementary and comparable observations to MRI. It is hoped that this cost effective optical imaging system could alleviate this technological reliance. The optical tomography scanner is evaluated from a series of investigations into the capabilities and limitations of optical tomographic imaging used in conjunction with gel dosimetry. Previously, the manufacture of gel dosimeters required anoxic environments in which the presence of oxygen in the dosimeter is limited. This requirement limited the production of gel dosimeters to chemistry laboratories that possessed the required technical expertise. MRI and optical imaging have been used to investigate the properties, dose response and the batch-to-batch reproducibility of a normoxic MAGIC gel dosimeter. The results obtained are encouraging having shown successfully reconstructed optical images obtained from various dose distributions delivered to the MAGIC gel dosimeter.

High-intensity focused ultrasound (HIFU) is rapidly emerging as a viable alterna- tive to conventional therapies in the treatment of deep-seated, solid tumours. In contrast to surgical methods, extracorporeal HIFU transducers non-invasively tar- get pathogenic tissue deep beneath the skin, inducing thermal necrosis of a volume of tissue typically coincident with the ultrasound focus. More recently, cavitation activity has been observed to enhance focal heating, whilst providing a unique op- portunity for real-time treatment monitoring. Unfortunately, the stochastic nature of cavitation makes it difficult to initiate and sustain the level of cavitation activity required for enhanced heating, and to confine the spatial extent of cavitation to the focal volume. The overall aim of this thesis is to design and implement a real-time, closed- loop controller for sustaining thermally relevant cavitation within the HIFU focal region. This is intended to improve the speed and reproducibility of tissue ablation, whilst providing clinicians with real-time feedback as to the extent and location of the ablated region. A quantitative relationship between the level of cavitation activity and asso- ciated temperature rise is first sought experimentally, by investigating cavitation- enhanced heating in two different tissue-mimicking materials (TMM) that yield dif- ferent levels of cavitation for the same HIFU exposure conditions. It is found that a minimum level of inertial cavitation activity is required for cavitation-enhanced heating to dominate the heating process, which is achieved in the first material but not the second. However, the introduction of exogenous, artificial nuclei to the second material is seen to augment cavitation levels to the extent that cavitation- enhanced heating becomes dominant. Subsequently, HIFU experimentation is extended to non-perfused, ex vivo bovine liver, into which a variety of cavitation nuclei are introduced to augment cav- itation levels, and hence heating. Commercially available lipid-shelled microbub- bles are contrasted with custom-made sonosensitive nanoparticles for their ability to seed cavitation events, culminating in an empirical relationship between iner- tial cavitation and heating that is common to both types of exogenous nuclei, and which agrees with the in vitro results. Moreover, the abnormally large lesions pro- duced are found to correlate with a broad spatial distribution of inertial cavitation events, as seen on two-dimensional passive acoustic maps. Based on these encouraging results, a novel negative-feedback, real-time con- trol system is implemented to sustain inertial cavitation within the focal region for extended periods of time. The controller is designed to be both asymmetric and adaptive, deploying different feedback gains to adjust the peak rarefactional focal pressure (PRFP), depending on whether cavitation activity is above or below the level required for cavitation-enhanced heating. With active cavitation control in vitro, the associated focal temperature elevation is maintained at a cytotoxic level for 20 seconds using less than half the energy input required in the absence of cavi- tation control. In order to test the applicability of the novel controller to a near-physiological environment, HIFU exposures are eventually performed in a unique normothermic perfused liver model that accounts for both heat advection and nuclei replenish- ment. Following preliminary experimentation, the controller is modified to account for the inherent variability in the cavitation threshold of perfused tissue, whilst the cavitation demand is also increased to account for heat advection. Following these modifications, use of the controller is found to enable greatly improved re- producibility of HIFU-induced lesions compared to those achieved without cavita- tion control, with a lesion size that is directly related to the cavitation demand. A cost-effective method for enabling caviation-enhanced, cavitation-controlled and cavitation-monitored HIFU therapy has thus been developed, which enables suc- cessful tissue ablation at acoustic energies lower than in current clinical use.

In this thesis, novel Case-Based Reasoning (CBR) methods were developed to be included in CBRDP (Case-Based Reasoning Dose Planner) -an adaptive decision support system for radiotherapy dose planning. CBR is an artificial intelligence methodology which solves new problems by retrieving solutions to previously solved similar problems stored in a case base. The focus of this research is on dose planning for prostate cancer patients. The records of patients successfully treated in the Nottingham University Hospitals NHS Trust, City Hospital Campus, UK, were stored in a case base and were exploited using case-based reasoning for future decision making. After each successful run of the system, a group based Simulated Annealing (SA) algorithm automatically searches for an optimal/near optimal combination of feature weights to be used in the future retrieval process of CBR. A number of research issues associated with the prostate cancer dose planning problem and the use of CBR are addressed including: (a) trade-off between the benefit of delivering a higher dose of radiation to cancer cells and the risk to damage surrounding organs, (b) deciding when and how much to violate the limitations of dose limits imposed to surrounding organs, (c) fusion of knowledge and experience gained over time in treating patients similar to the new one, (d) incorporation of the 5 years Progression Free Probability and success rate in the decision making process and (e) hybridisation of CBR with a novel group based simulated annealing algorithm to update knowledge/experience gained in treating patients over time. The efficiency of the proposed system was validated using real data sets collected from the Nottingham University Hospitals. Experiments based on a leave-one-out strategy demonstrated that for most of the patients, the dose plans generated by our approach are coherent with the dose plans prescribed by an experienced oncologist or even better. This system may play a vital role to assist the oncologist in making a better decision in less time; it incorporates the success rate of previously treated similar patients in the dose planning for a new patient and it can also be used in teaching and training processes. In addition, the developed method is generic in nature and can be used to solve similar non-linear real world complex problems.

Bibliography: p. 217-234. / xvii, 234 p. : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / A method of calculating beam data using sector-integration of Monte Carlo-generated pencil beam kernels has been proposed. This technique can be used both for circular and irregular fields, however it is more useful in calculating dosimetry parameters for irregular fields where the shape of treatment fields is determined during planning. / Thesis (Ph.D.)--University of Adelaide, Dept. of Physics and Mathematical Physics, 1996

La curiethérapie est une technique ancienne utilisant des sources scellées de faible ou moyenne énergie, toujours pertinente aux plans thérapeutique et économique et toujours en évolution (e.g. curiethérapie à Haut Débit de Dose (HDD)). Cette modalité de traitement permet de délivrer une forte dose d'irradiation dans un volume-cible limité, et permet de minimiser le risque de cancer radio-induit en préservant les Organes à Risques (OAR). Cependant, elle génère des gradients de dose élevés rendant la dosimétrie in vivo délicate à mettre en oeuvre. Les écarts constatés entre doses délivrées et prescrites sont ainsi fréquemment supérieurs à l'écart maximal toléré par la réglementation (± 5%) en usage pour la radiothérapie externe conventionnelle et rendue obligatoire en France par décret depuis 2011. Ce travail de thèse s'est déroulé dans le cadre du projet ANR-TECSAN INTRADOSE et exploite les acquis technologiques antérieurs démontrés à l'issue du projet Européen MAESTRO et du projet ANR-TECSAN CODOFER, en particulier une instrumentation RL/OSL (Radioluminescence - Optically Stimulated Luminescence) multivoies réalisée et validée au plan préclinique dans le cadre du projet MAESTRO. Le projet INTRADOSE a pour objectif de démontrer la faisabilité d'une Dosimétrie In Vivo (DIV) intracavitaire par cathéter dosimétriques à fibres optiques et cristaux d'alumine Al2O3:C dans le but d'améliorer la sécurité des patients traités par Curiethérapie HDD. Ce nouveau type de détecteur permet de mesurer une distribution de doses (mulitpoints) proche des OAR, il présente un petit diamètre (<ou= 3 mm) destiné à un usage intracavitaire (e.g. insertion dans l'urètre), transparent, stable sous radiations et réutilisable après lecture des doses et stérilisation. Au cours de mon travail de thèse, nous avons développé ce nouveau cathéter dosimétrique basé sur la Luminescence Stimulée Optiquement utilisant les propriétés du cristal d'alumine. Différents tests ont permis d'évaluer sa faisabilité et sa compatibilité avec l'application médicale. Dans un second temps, des validations métrologiques et précliniques ont été menées dans le but de valider la réponse du détecteur dans le cadre d'une application de curiethérapie HDD de prostate. Ces différentes mesures ont été également confrontées à des simulations Monte-Carlo. Financé par l'Agence Nationale pour la Recherche - Technologies pour la Santé (ANR-TECSAN), ce projet associe le CEA LIST [Laboratoire de Mesures Optiques (LMO) et Laboratoire National Henri Becquerel (LNHB)] et le Centre de Lutte Contre le Cancer (CLCC) Léon Bérard (CLB), Centre Conseil et référent en Radiophysique / The brachytherapy is an old technique using sealed radioactive sources of low or average energy. This technique is still therapeutically and economically relevant today and always evolving (e.g. High Dose Rate (HDR) brachytherapy). This treatment enables to deliver a high dose of irradiation in a limited tumoral volume and enables to minimize the risk of radiation-induced cancer as preserving the Organs at Risks (OAR). However, this technique generates high dose gradients, which makes in vivo dosimetry difficult to implement. Hence, the deviations observed between doses delivered and prescribed are often up to the maximal deviation tolerated by the nuclear safety regulations (± 5%) in conformational radiotherapy. Those regulations have been made mandatory in France since 2011. This thesis has been done within the framework of the ANR-TECSAN INTRADOSE project and is based on the past technological benefits demonstrated during the MAESTRO European project and the ANR-TECSAN CODOFER project, in particular a RL/OSL multichannel instrumentation (Radioluminescence - Optically Stimulated Luminescence) made and validated in preclinical evaluation during the MAESTRO project. The purpose of the INTRADOSE project is to demonstrate the feasibility of the intracavitary In Vivo Dosimetry (IVD) by dosimetric catheter using optical fibers and alumina crystals Al2O3:C with the aim of improving the safety of patients treated by HDR brachytherapy. This new probe enables to measure a dose distribution (several points) close to the OAR, it offers a little diameter (<or= 3 mm) designed for an intracavitary use (e.g. to insert in the urethra), it is transparent, radiation stable and reusable after dose reading and sterilization. During this study, we have first developed this new dosimetric sensor based on the OSL using the properties of the alumina crystal. Several tests have been done in order to evaluate the feasibility and the compatibility with a medical application. Then, metrological and preclinical validations have been done in order to validate the response of the sensor for a prostatic application using HDR brachytherapy. These different measurements have also been compared to Monte-Carlo simulations. Funded by the Agence Nationale pour la Recherche - Technologies pour la Santé (ANR-TECSAN), this project combines the CEA LIST [Laboratoire de Mesures Optiques (LMO) and the Laboratoire National Henri Becquerel (LNHB)] and the Centre de Lutte Contre le Cancer (CLCC) Léon Bérard (CLB), advice center and advisor in radiophysics

Toxicity of the respiratory system is quite common after radiotherapy in thoracic tumours. The quantification of lung tissue response to irradiation is important in designing treatments associated with a minimum of complications and maximum tumor control. This work aims to estimate volumes V13, V20 and V30 as an index of radiation pneumonitis occurrence, to evaluate the predictive strength of the relative seriality, Lyman-Kutcher-Burman(LKB) and parallel normal tissue complication probability (NTCP) models regarding the incidence of radiation pneumonitis in a group of patients following lung cancer radiotherapy when lung perceived as paired and single organ respectively and also software development for the determination of the best estimates of the models’ parameters based on maximum likelihood method. The study was based on 46 patients and for each patient, lung dose-volume histograms (DVHs) and the clinical treatment outcome was available. From the 46 patients treated, 28 of them were scored as having radiation induced pneumonitis, with RTOG criteria grade ≥2. Firstly lungs were evaluated as a paired organ. Analyzing this material we failed to associate volume V13, V20 and V30 with radiation pneumonitis occurrence (χ2-test: probability of agreement between observed and predicted results using the 0.05 significance level). By applying ANOVA of the NTCP models examined in the overall group considering lungs as paired organs the LKB with Martel et al parameter set gave the best results, whereas when lungs perceived as individual organ (unhealthy lung volume-PTV) the best model was appeared to be LKB with Burman et al parameter set. However, in this relatively small group of lung cancer patients NTCP models didn’t show excessive correlation with the clinical outcome. Nevertheless, when total lung volume irradiated and total dose received were taken into account as factors of radiation pneumonitis prediction, correlation was almost duplicated for both perception of lungs. In order to achieve the best fitting of models to the clinical outcome for the specific patient group, maximum likelihood analysis was applied via software development using mle programming language, to find those parameters that maximize the likelihood function. When lungs perceived as single organ, the best fitting of models to the clinical outcome for relative seriality were D50 = 22Gy, γ= 2, s=0.031, LKB model D50 = 23Gy, m=0.18, n=1 and for parallel model, D50 = 20Gy, m=0.2, n=0.6. Maximum likelihood analysis was not applied for paired lung assumption as constraints did not allow us to properly fit the models. / -

- / -

- / -

La radiothérapie est une ressource incontestée dans l'arsenal de traitement du cancer. De par sa spécificité, elle induit des risques pour les patients et pour les professionnels de santé qui imposent qu'elle soit opérée dans un environnement et un cadre maîtrisé et contrôlé. Cependant, de par la multiplicité des intervenants et la complexité de l'organisation médicale mise en place autour du patient, il s'est produit ces dernières années des incidents et des accidents dans le processus de traitement dont les conséquences sont parfois dramatiques. Pour limiter voire supprimer ces évènements, il est indispensable d'aborder la problématique du traitement par radiothérapie sous l'angle de l'analyse des risques de systèmes sociotechniques en intégrant la complexité technique/humain/organisationnelle du processus dans une vue "transverse". La contribution apportée par ce travail se présente sous la forme d’une démarche basée sur des modèles qui permettent d’évaluer, à priori et pour un établissement donné, un indicateur de la sécurité offerte aux patients lors de leur traitement par radiothérapie / Radiotherapy is an uncontested resource in the arsenal of cancer treatment. This technique implies risks for both patients and radiotherapy professionals, which requires it to be carried out in a controlled and monitored environment. However, because of the multiplicity of actors, the complexity of the healthcare organization set up around the patient and the complexity of medical devices, in recent years some incidents and accidents in treatment process appeared, with sometimes tragic consequences. To reduce or eliminate these events, it is essential to address the problem of radiation therapy in terms of risk analysis of socio-technical systems, by including technical / human / organizational complexity in a "transverse" view. The main contribution of this work consists in an approach based on models that assess a priori, for a given radiotherapy process, an indicator of the security provided to patients during their radiotherapy treatment

Absolute quantification of I-131 activity in tumors and normal tissue is essential for internal dose estimates and is one of the greatest challenges in contemporary Nuclear Medicine. Therefore, for obtaining more accurate results, the most suitable and clinical applicable method, for both patients and scientists, for the quantification of I-131 has to be found. In the theoretical part of this project, the two methods for the quantitation are explained and analyzed as well as the factors that affect the obtained clinical image followed by the corrections that must be applied for gaining a more discrete image. For the planar technique, two methods for scatter and septal penetration correction were used applying two different matrix sizes specifically, 256x256 and 512x512 matrix sizes. On the other hand, for the SPECT technique, a 64x64 matrix size was used and a comparison between an auto-ROI and a same pattern of ROI segmentation that was produced manually for the purpose of this project, was performed. Furthermore, for patient specific dosimetry the total absorbed dose for the thyroid gland was calculated. Figures and tables for planar and SPECT technique were produced for each matrix size along with the final curve of the absorbed dose of the patient. For the planar technique it was observed that the method proposed by Anne Larsson, according to statistical analysis, gives results with lower statistical errors than the method proposed by Macey. In addition, 512x512 matrix size for the planar technique leads to lower statistical errors when compared with the results from using the 256x256 matrix size. Moreover, for the SPECT technique, the auto-sum that the software of the camera provides gives exactly the same results compared with the manual-sum. Finally, the total thyroid absorbed dose was 45,0472 Gy a result very close to the ICRU proposed dose value. / Η απόλυτη ποσοτικοποίηση της ενεργότητας του I-131 σε καρκινικούς και σε φυσιολογικούς ιστούς είναι απαραίτητη για την εκτίμηση της εσωτερικής δόσης και αποτελεί μία από τις μεγαλύτερες προκλήσεις της σύγχρονης Πυρηνικής Ιατρικής. Για τον λόγο αυτό, για την απόκτηση ακριβέστερων αποτελεσμάτων, θα πρέπει να βρεθεί η καταλληλότερη μέθοδος ποσοτικοποίησης, που να είναι εύχρηστη τόσο για τον ασθενή όσο και για το επιστημονικό προσωπικό των νοσοκομείων. Στο θεωρητικό μέρος της παρούσας εργασίας, αναλύονται οι δύο μέθοδοι ποσοτικοποίησης καθώς και οι παράγοντες που επηρεάζουν την ποιότητα της εικόνας ακολουθούμενοι από τους διορθωτικούς παράγοντες για την απόκτηση μιας ευκρινέστερης κλινικής εικόνας. Όσον αφορά την Planar τεχνική, περιγράφονται δύο μέθοδοι για διόρθωση της σκεδαζόμενης ακτινοβολίας καθώς επίσης και του septal penetration στην τελική εικόνα, χρησιμοποιώντας δύο διαφορετικά μεγέθη μήτρας, συγκεκριμένα, 256x256 και 512x512. Για την τεχνική SPECT, το μέγεθος μήτρας που χρησιμοποιήθηκε ήταν 64x64 και στην συνέχεια έγινε σύγκριση της περιοχής ενδιαφέροντος (ROI) που έδωσε το λογισμικό της κάμερας με την περιοχή ενδιαφέροντος που σχεδιάστηκε χειροκίνητα για την συγκεκριμένη μελέτη. Επίσης, υπολογίστηκε η συνολική απορροφούμενη δόση στον θυρεοειδή αδένα ενός ασθενούς. Κατασκευάστηκαν διαγράμματα τόσο για την planar όσο και για την SPECT τεχνική και για τα δύο μεγέθη μήτρας καθώς και η καμπύλη απορροφούμενης δόσης του ασθενούς. Παρατηρήθηκε ότι στην planar τεχνική, η μέθοδος που προτείνεται από την Anne Larsson, δίνει καλύτερα αποτελέσματα και μικρότερα σφάλματα συγκρινόμενη με την μέθοδο που προτείνεται από τον Macey. Επιπλέον, παρατηρήθηκε ότι το 512x512 μέγεθος μήτρας δίνει πιο ευκρινή εικόνα απ’ ότι το 256x256 μέγεθος μήτρας. Επίσης, με την SPECT τεχνική, παρατηρείται ότι χρησιμοποιώντας το αυτόματο άθροισμα των κρούσεων που δίνεται από το λογισμικό της κάμερας προκύπτουν τα ίδια ακριβώς αποτελέσματα με την χρήση του αθροίσματος των κρούσεων όταν αυτό γίνεται χειροκίνητα. Τέλος, η τελική απορροφούμενη δόση του ασθενούς υπολογίστηκε 45,0472 Gy, ένα αποτέλεσμα που δεν αποκλείει από τις τιμές που προτείνονται από το ICRU.