Introdução e objetivo: A avaliação da gravidade da neoplasia em pacientes com câncer de próstata tem como propósito identificar e tratar somente os pacientes com doença clinicamente significativa. Os parâmetros clínicos e histopatológicos tradicionalmente utilizados na estratificação de risco classificam erroneamente uma parcela importante dos pacientes. O objetivo deste estudo foi avaliar o papel da ressonância nuclear multiparamétrica de próstata (RNMMP) e da biópsia transretal realizada com fusão de imagem e estimativa visual na estratificação precoce dos pacientes em vigilância ativa. Métodos: Foram incluídos prospectivamente pacientes com câncer de próstata de baixo risco candidatos a vigilância ativa: biópsia inicial convencional com no mínimo 12 fragmentos; escore de Gleason <= 6; PSA sérico <= 10,0; <= 3 fragmentos positivos; <= 50% de acometimento de cada fragmento; toque T1c ou T2a. Todos os pacientes foram submetidos à RNMMP e biópsia confirmatória: biópsia aleatória sistemática e biópsia guiada com fusão de imagem (ultrassom e RNMMP) e analisada por método cognitivo de estimativa visual. As regiões suspeitas para câncer foram definidas e classificadas utilizando-se a escala PI-RADS (Prostate Imaging Reporting and Data System). Definimos reclassificação na biópsia confirmatória como aparecimento de escore de Gleason >= 7, > 3 fragmentos positivos ou >= 50% de envolvimento de qualquer fragmento. A performance da RNMMP em prever os resultados da biópsia confirmatória foi estudada. Análise uni e multivariada mediante regressão logística avaliou a relação entre idade, PSA, densidade de PSA, número de fragmentos positivos na biópsia inicial e o escore da RNMMP e a chance de reclassificação na biópsia confirmatória. Resultados: Cento e cinco pacientes estiveram disponíveis para análise final. Quarenta e dois (40%) pacientes apresentaram PI-RADS 1,2 ou 3 e 63 (60%) PI-RADS 4 ou 5. No geral, 87 pacientes foram submetidos à biópsia guiada com fusão. A taxa de reclassificação entre pacientes com PI-RADS 1,2,3,4 e 5 foi de 0%, 23.1%, 9.1%, 74.5% e 100%, respectivamente. No geral, a sensibilidade, especificidade, valor preditivo positivo (VPP) e valor preditivos negativo (VPN) da RNMMP para predição de reclassificação foi de 92.5%, 76%, 81% e 90.5%, respectivamente. Daqueles reclassificados, 47 (44.8%) tinham escore de Gleason >= 7 e 11 apresentavam (10.4%) >= 3 fragmentos positivos e >= 50% de envolvimento de qualquer fragmento individualmente. Na análise multivariada, somente a densidade de PSA e a RNMMP permaneceram como fatores preditivos significativos para reclassificação (p < 0,05). Na tabulação cruzada a biópsia aleatória sistemática teria deixado de classificar corretamente 15 pacientes com câncer significativo detectados pela biópsia com fusão de imagem. Por outro lado, a biópsia aleatória sistemática detectou cinco casos de câncer significante que não teriam sido identificados pela biopsia de fusão de imagem isoladamente. Conclusões: A RNMMP é uma ferramenta importante na predição da taxa de reclassificação da gravidade da neoplasia de próstata em pacientes candidatos à vigilância ativa submetidos à biopsia confirmatória. A taxa de reclassificação na biopsia confirmatória é particularmente alta no grupo de pacientes com lesões PI-RADS grau 4-5. Apesar da utilidade da biópsia com fusão de imagem, recomenda-se a manutenção da prática de se retirar fragmentos sistemáticos aleatórios quando da realização da biópsia confirmatória para maximizar a detecção de neoplasia de comportamento agressivo / Introduction and objective: The goal of prostate cancer (PCa) risk stratification is to identify and treat only men with clinically significant disease. Clinical and pathologic parameters currently used to stratify PCa risk misclassify a significant amount of patients. The objective of this study was to evaluate the role of multiparametric magnetic resonance imaging of the prostate (mpMRI) and transrectal guided biopsy with visual estimation (TRUS-Bx) in early risk stratification of patients with prostate cancer on active surveillance. Methods: Study subjects were prospectively enrolled including patients with low risk, low-grade, localized PCa: Gleason <= 6, T1c-T2a, PSA <=10 ng/ml, no more than 3 positive cores and <= 50% involvement of single cores. They were followed with subsequent mp-MRI and confirmatory biopsy (CB): standard biopsy (SB) and visual estimation-guided TRUS-Bx. Cancer-suspicious regions (CSRs) were defined using Prostate Imaging Reporting And Data System (PIRADS) scores. Reclassification occurred if CB confirmed Gleason >= 7, > 3 positive fragments or >= 50% involvement of any core. The performance of mp- MRI on predicting CB results was assessed. Univariate and multivariate logistic regression were performed to study relationships between age, PSA, PSA density, number of positive cores in the initial biopsy and mpMRI grade on CB reclassification. Results: 105 patients were available for analysis. 42 (40%) patients had PIRADS 1,2 or 3 lesions and 63 (60%) had only grades 4 or 5 lesions. Overall, 87 patients underwent visual estimation TRUS-Bx. Reclassification among patients with PI-RADS 1,2,3,4 and 5 was 0%, 23.1%, 9.1%, 74.5% and 100%, respectively. Overall, mpMRI sensitivity, specificity, PPV and NPV for disease reclassification were 92.5%, 76%, 81% and 90.5%, respectively. Of men reclassified 47 (44.8%) were upstaged because of Gleason >= 7, 11 (10.4%) because of >= 3 positive fragments plus >= 50% involvement. On multivariate analysis, only PSA density and mpMRI remained significant for reclassification (p < 0,05). On cross-tabulation SB would have missed 15 significant cases detected by targeted biopsy. On the other hand SB detected 5 cases of significant cancer not detected by targeted biopsy alone. Conclusions: Multiparametric magnetic resonance imaging is a significant tool for predicting cancer severity reclassification on CBx among AS candidates. The reclassification rate on CBx is particularly high in the group of patients who have PI-RADS grades 4 or 5 lesions. Despite the usefulness of visual-guided biopsy, it still remains highly recommended to retrieve standard fragments during CBx in order to avoid missing significant tumors

Magnetic Resonance Imaging (MRI) provides unique advantages such as superior soft tissue contrast, true multiplanar imaging, variable contrast mechanisms, measurement of temperature changes, perfusion and diffusion, and no ionizing radiation. Despite considerable research efforts in the field of interventional MRI, numerous challenges remain including restricted access to the patient, high acoustic noise and a shortage of MRI-safe devices. Novel methods and devices are presented in this thesis with the primary objective of enabling effective MRI-guided interventions, particularly abdominal needle and common catheter-based endovascular interventions. Firstly, a set of MRI-safe devices (guidewires, micro guidewires, catheters and micro catheters) were developed with passive or inductively coupling resonant markers for MRI visualisation. Secondly, a method was implemented for wireless tracking and dynamic guidance of instruments. Thirdly, a framework of technologies was developed for in-room display, wireless MRI remote control and multi-user communication along with a dedicated user interface and imaging protocol. These implementations were assessed in regards to MRI-safety, performance and usability and evaluated for MRI-guided liver biopsies, balloon angioplasty procedures and also for mechanical thrombolysis. Flow phantoms, Thiel soft-embalmed human cadavers with partially re-established perfusion and a porcine model were used for in vitro, ex vivo and in vivo validation, respectively. The results demonstrate that these interventions are experimentally feasible and practical when using the presented developments: automated device tracking and equipment designed for MRI-guided interventions streamlined procedural workflow. Specifically, it was shown that fast and accurate needle placements along complex trajectories were feasible using a wireless interactive display and control device with a dedicated user interface for interventions. Moreover, safe and efficacious balloon angioplasties of the iliac artery were practical using the described framework of technologies along with a dedicated MRI protocol. Finally, it was demonstrated that these developments could be adapted and applied to MRI-guided endovascular mechanical thrombolysis of the middle cerebral artery. The technologies, described in this thesis have been shown to overcome many of the present limitations and should therefore be useful for enabling MRI-guided interventions while not further constraining the operating physician in an already complex environment. Nevertheless, it is acknowledged that many crucial issues remain to be solved in the field of iMRI and in the context of the presented research. In particular further device optimisations, improvements of the tracking implementation along with further in vivo evaluations are required before moving towards clinical evaluation. This thesis sets the groundwork for moving ahead with the eventual clinical realisation of optimised MRI-guided interventions.

Numerous studies have demonstrated the potential efficacy of percutaneous image-guided interventions over open surgical interventions. The conventional image-guided procedures are limited by the freehand technique, requiring mental 3D registration and hand-eye coordination for needle placement. The outcomes of these procedures are associated with longer duration and increased patient discomfort with high radiation exposure. Previously, a static image overlay system was proposed for aiding needle interventions. Certain drawbacks associated with the static system limited the clinical translation. To overcome the ergonomic issues and longer calibration duration associated with static system, an adjustable image overlay system was proposed. The system consisted of monitor and semi-transparent mirror, attached together to an articulated mobile arm. The 90-degree mirror-monitor configuration was proposed to improve the physician access around the patient. MicronTracker was integrated for dynamic tracking of the patient and device. A novel method for auto-direct calibration of the virtual image overlay plane was proposed. Due to large mechanical structure, the precise movement was limited and consumed useful space in the procedure room. A mobile image overlay system with reduced system weight and smaller dimensions was proposed to eliminate the need for mechanical structure. A tablet computer and beamsplitter were used as the display device and mirror respectively. An image overlay visualization module of the 3D Slicer was developed to project the correct image slice upon the tablet device. The system weight was reduced to 1 kg and the image overlay plane tracking precision (0.11mm STD=0.05) was similar to the printed physical markers. The auto-calibration of the image overlay plane can be done in two simple steps, away from the patient table and without additional phantom. Based on the successful pre-clinical testing of the previous static system, the mobile image overlay system with reduced weight, increased tracking precision and easier maneuverability, can be possibly hand-held by the physician to explore the image volume over the patient and be used for a wide range of procedures. The mobile image overlay system shall be classified as Class II device as per FDA regulations, do not require extensive verification and validation efforts and further improves the commercialization opportunities. / Thesis (Master, Mechanical and Materials Engineering) -- Queen's University, 2014-06-26 18:51:03.958

This thesis investigates the exact minimisation of treatment delivery time for Intensity Modulated Radiation Therapy (IMRT) for the treatment of cancer using Multileaf Collimators (MLC). Although patients are required to remain stationary during the delivery of IMRT, inevitably some patient movement will occur, particularly if treatment times are longer than necessary. Therefore minimising the treatment delivery time of IMRT may result in less patient movement, less inaccuracy in the dosage received and a potentially improved outcome for the patient. When IMRT is delivered using multileaf collimators in 'step and shoot' mode, it consists of a sequence of multileaf collimator configurations, or shape matrices; for each, time is needed to set up the configuration, and in addition the patient is exposed to radiation for a specified time, or beam-on time. The 'step and shoot leaf sequencing' problems for minimising treatment time considered in this thesis are the constant set-up time Total Treatment Time (TTT) problem and the Beam-on Time Constrained Minimum Cardinality (BTCMC) problem. The TTT problem minimises a weighted sum of total beam-on time and total number of shape matrices used, whereas the BTCMC problem lexicographically minimises the total beam-on time then the number of shape matrices used in a solution. The vast majority of approaches to these strongly NP-hard problems are heuristics; of the few exact approaches, the formulations either have excessive computation times or their solution methods do not easily incorporate multileaf collimator mechanical constraints (which are present in most currently used MLC systems). In this thesis, new exact mixed integer and integer programming formulations for solving the TTT and BTCMC problems are developed. The models and solution methods considered can be applied to the unconstrained and constrained versions of the problems, where 'constrained' refers to the modelling of additional MLC mechanical constraints. Within the context of integer programming formulations, new and existing methods for improving the computational efficiency of the models presented are investigated. Numerical results for all variations considered are provided. This thesis demonstrates that significant computational improvement can be achieved for the exact mixed integer and integer programming models investigated, via solution approaches based on an idea of systematically 'stepping-up' through the number of shape matrices used in a formulation, via additional constraints (particularly symmetry breaking constraints) and via the application of improved bounds on variables. This thesis also makes a contribution to the wider field of integer programming through the examination of an interesting substructure of an exact integer programming model. In summary, this thesis presents a thorough analysis of possible integer programming models for the strongly NP-hard 'step and shoot' leaf sequencing problems and investigates and applies methods for improving the computational efficiency of such formulations. In this way, this thesis contributes to the field of leaf sequencing for the application of Intensity Modulated Radiation Therapy using Multileaf Collimators.

En chirurgie mini-invasive guidée par l’image, une planification préopératoire précise des trajectoires des outils chirurgicaux est un facteur clé pour une intervention réussie. Cependant, une planification efficace est une tâche difficile, qui peut être considérablement améliorée en considérant différents facteurs contributifs tels que les déformations biomécaniques intra-opératoires, ou en introduisant de nouvelles techniques d'optimisation. Dans ce travail, nous nous concentrons sur deux aspects. Le premier aspect porte sur l'intégration de la déformation intra-opératoire dans le processus de planification de trajectoire. Nos méthodes combinent des techniques d'optimisation géométrique à base de simulations biomécaniques. Elles sont caractérisées par un certain niveau de généralité, et ont été expérimentées sur deux types d’interventions chirurgicales: les procédures percutanées pour l'ablation de tumeurs hépatiques, et la stimulation cérébrale profonde en neurochirurgie. Deuxièmement, nous étudions, mettons en œuvre, et comparons plusieurs approches d'optimisation en utilisant des méthodes qualitatives et quantitatives, et nous présentons une méthode efficace d'optimisation évolutionnaire multicritères à base de Pareto qui permet de trouver des solutions optimales qui ne sont pas accessibles par les méthodes existantes. / In image-guided minimally invasive surgery, a precise preoperative planning of the surgical tools trajectory is a key factor to a successful intervention. However, an efficient planning is a challenging task, which can be significantly improved when considering different contributing factors such as biomechanical intra-operative deformations, or novel optimization techniques. In this work, we focus on two aspects. The first aspect addresses integrating intra-operative deformation to the path planning process. Our methods combine geometric-based optimization techniques with physics-based simulations. They are characterized with a certain level of generality, and are experimented on two different surgical procedures: percutaneous procedures for hepatic tumor ablation, and in neurosurgery for Deep Brain Stimulation (DBS). Secondly, we investigate, implement, and compare many optimization approaches using qualitative and quantitative methods, and present an efficient evolutionary Pareto-based multi-criteria optimization method which can find optimal solutions that are not reachable via the current state of the art methods.

Les gliomes sont des tumeurs cérébrales primitives représentant le deuxième cancer le plus fréquent chez l’enfant et la troisième cause de mortalité chez l’adulte jeune. Il a été démontré qu’une chirurgie d’exérèse tumorale permet d’augmenter la qualité et la durée de vie du patient, voir même dans certains cas, d’obtenir sa guérison. L’identification intraopératoire des résidus tumoraux permettrait au chirurgien de vérifier, in situ, la qualité de son geste d’exérèse. Une nouvelle modalité d’imagerie intraopératoire a été proposée comme solution pour détecter les tumeurs résiduelles. Elle se base sur l’utilisation d’une sonde nucléaire combinée à un système de localisation optique. Cette nouvelle modalité, appelée imagerie surfacique de positons (ISP), permet la génération d’images de la distribution surfacique d’un radiotraceur comme le 18 F − F DG d’une zone d’intérêt scannée. L’ISP n’étant actuellement pas utilisée en clinique, nous proposons pour la première fois une étude de faisabilité de son utilisation pour l’optimisation de la chirurgie des tumeurs cérébrales. Nous montrons les limites de l’utilisation potentielle de l’ISP dans un contexte neurochirurgical par des études expérimentales en considérant des facteurs intraopératoires pouvant influencer la qualité des images générées par le système. Les contributions présentées se concentrent sur trois axes. Dans un premier temps, nous souhaitons obtenir des images fonctionnelles d’ISP avec un temps de calcul faible. L’application de modèles d’acquisition aux mesures permettent d’améliorer la qualité des images, au détriment d’un temps de calcul élevé. Nous suggérons un nouveau modèle d’acquisition dédié au contexte intraopératoire, permettant l’amélioration de la résolution et du contraste des images pour un temps de calcul réduit. Un deuxième axe est dédié à l’étude de l’impact d’une acquisition intraopératoire sur les performances du système. Nous proposons l’estimation des paramètres intrinsèques de la sonde nucléaire utilisée et de l’influence de la hauteur du processus de scan sur leurs valeurs. Le dernier axe est consacré à la validation de modèles d’acquisition dans le contexte intraopératoire. Nous présentons une étude comparative des performances de modèles d’acquisition en considérant l’impact potentiel de la vitesse de scan durant l’acquisition. L’ensemble de ces travaux a contribué à l’étude de faisabilité de l’utilisation d’un système d’ISP en intraopératoire, proposée pour la première fois en neurochirurgie. / Gliomas are primitive brain tumors, which represent the second most frequent cancer among children and the third cause of death among young adult. It has been shown that resection surgery improves patient outcomes, leading to its cure for some cases. Intraoperative detection of residual tumor allows the surgeon to check the quality of its resection gesture. A new intraoperative imaging modality has been proposed as a solution to detect residual tumors. It relies on a nuclear probe associated with an optical localization system. This new modality, called positron surface imaging (PSI), generates images of the activity surface distribution of a 18 F based radiotracer of a scanned area of interest. Not used in clinical context yet, we proposed for the first time a feasibility study of its usage to optimize brain tumor surgery. We show limitations of the potential usage of PSI in neurosurgery through experimental studies, considering intraoperatives factors which may influence quality of generated images. Contributions in this study are presented on three axes. First, we want to generate PSI functional images with a low computational time. Acquisition models applied to measurements improve image quality at the cost of high computational time. We suggest a new acquisition model dedicated to intraoperative usage, allowing enhancement of spatial resolution and contrast of images for a low computational time. The second axis is dedicated to the study of the intraoperative acquisition impact on system performance. We propose to estimate the intrinsic parameters of nuclear probe and the study of scanning process on their values. Finally, we present the validation of acquisition models dedicated to intraoperative context. We show a comparative study of acquisition models performance considering potential impact of scanning speed process during acquisition. This work has contributed to the feasibility study of using an ISP system in the intraoperative context, proposed for the first time in neurosurgery.

Le but principal de ce projet est de préparer l’implantation clinique du système Clarity qui utilise une sonde ultrasonore pour visualiser l’anatomie interne du patient. Ce système est utilisé pour les cas de prostate et nécessite d’être adapté pour les cas de poumon. L’utilité de ce système est de suivre un substitut afin d’inférer la position d’une tumeur pulmonaire. L’hypothèse de cette étude est qu’un substitut interne serait mieux corrélé avec une tumeur pulmonaire que le seraient des marqueurs externes. Les sous-objectifs sont : 1) aborder l’adaptation du montage pour faire des acquisitions sur des patients ; 2) explorer la performance des algorithmes de détection de mouvements ainsi que des métriques de qualité d’image sur des images US et ciné IRM; 3) démontrer que la corrélation entre un substitut interne et une structure pulmonaire est plus grande que celle avec un substitut externe. Pour les acquisitions d’images US, la sonde est placée sur les volontaires et fixée à la table de traitement à l’aide d’un bras mécanique. Il a été démontré qu’une pression insuffisante peut causer une perte de signal dû à la forme curviligne de la sonde. La diminution de la moyenne des intensités de l’image et de l’écart-type confirme une perte de signal lors d’amplitudes respiratoires élevées justifiée par une perte de contact entre la sonde et la peau malgré la fixation de la sonde. Entre les algorithmes de corrélation croisée normalisée (NCC), d’erreur moyenne quadratique (RMS) et de flux optique, la méthode NCC semble la plus robuste pour suivre le substitut interne (structure dans le foie) dans les images IRM pour 5/9 volontaires sains ( = 0, 050). Cette méthode est utilisé présentement pour les cas de prostate. Le flux optique s’est montré plus efficace pour des cas spécifiques ce qui démontre l’intérêt d’adapter l’algorithme pour les cas de poumons. Enfin, il a été démontré sur les images IRM qu’un substitut interne au niveau du foie est plus efficace pour la majorité des volontaires (8/9) en comparaison avec un marqueur sur la peau placé dans la région abdominale. Le marqueur abdominal possède une meilleure corrélation qu’un marqueur thoracique (9/9) illustrant l’importance du positionnement d’un marqueur externe pour le suivi d’une tumeur pulmonaire. / The main objective of this thesis is to prepare the clinical implementation of the Clarity ultrasound system for indirect lung tumours tracking using a surrogate. It is currently used for motion management during prostate treatments and requires adaptation. Our hypothesis is that an internal marker would have a better correlation with the tumour’s position than an external surrogate. The sub-objectives are : 1) test different setups for the image acquisition on patients ; 2) explore the algorithms’ performance for motion detection as well as the image quality metrics on US and dynamic MRI images ; 3) evaluate the correlation between surrogates and a lung structure to determine which performs best. The ultrasound probe is fixed on the treatment couch for the acquisition on healthy volunteers using a mechanical arm. Low pressure on the patient’s skin results in a loss of signal due to the curvilinear shape of the probe. We observed a loss of contact between the probe and the volunteers’ skin due to ample movements causing a deterioration of the image quality. We tested three different motion detection algorithms on dynamic MRI images : normalized cross-correlation (NCC), root mean square error (RMS) and optical flow. The NCC algorithm is the most robust out of the three for 5/9 volunteers for the internal surrogate tracking ( < 0.050). In specific cases, the optical flow method performed better indicating an interest in developping a new algorithm for indirect lung tracking. Finally, the correlation between the surrogates and a lung structure were calculated using the MRI images. The internal surrogate inside the liver was proven more efficient for indirect lung tumour tracking for 8/9 volunteers. External markers give a greater prediction error. It has also been shown that the positioning of the external marker on the patient’s skin impacts the correlation. The abdominal marker is better than the thoracic one for all the volunteers.

The identification of eloquent areas is of utmost importance in the surgery of tumors located near speech-eloquent brain areas, since the classical concept of a constant localization was proven to be untrue and the spatial localization of these areas may show large interindividual differences. Some neurosurgical centers apply intraoperative electrophysiological methods that, however, necessitate the performance of surgery in the awake patient. This might be a severe burden both for the patient and the operating team in a procedure that lasts several hours; in addition, electrical stimulation may generate epileptic seizures. Alternatively, methods of functional brain imaging (e.g., PET, fMRI, MEG) may be applied, which allow individual localization of speech-eloquent areas. Matching of these image data with a conventional 3D-CT or MRI now allows the exact transfer of this information into the surgical field by neuronavigation. Whereas standards concerning electrophysiological stimulation techniques that could prevent a permanent postoperative worsening of language are available, until now it remains unclear whether the resection of regions shown to be active in functional brain imaging will cause a permanent postoperative deficit. / Die Identifikation sprachaktiver Areale ist von höchster Bedeutung bei der Operation von Tumoren in der Nähe des vermuteten Sprachzentrums, da das klassische Konzept einer konstanten Lokalisation des Sprachzentrums sich als unrichtig erwiesen hat und die räumliche Ausdehnung dieser Areale eine hohe interindividuelle Varianz aufweisen kann. Einige neurochirurgische Zentren benutzen deshalb intraoperativ elektrophysiologische Methoden, die jedoch eine Operation am wachen Patienten voraussetzen. Dies kann sowohl für den Patienten als auch das Operations-Team eine schwere Belastung bei diesem mehrstündigen Eingriff darstellen, zusätzlich können epileptische Anfälle durch die elektrische Stimulation generiert werden. Alternativ können Modalitäten des «functional brain imaging» (PET, fMRT, MEG usw.) eingesetzt werden, die die individuelle Lokalisation sprachaktiver Areale gestatten. Die Bildfusion dieser Daten mit einem konventionellen 3D-CT oder MRT erlaubt den exakten Transfer dieser Daten in den OP-Situs mittels Neuronavigation. Während Standards bei elektrophysiologischen Stimulationstechniken existieren, die eine permanente postoperative Verschlechterung der Sprachfunktion weitgehend verhindern, bleibt die Relevanz sprachaktiver Areale bei den neuesten bildgebenden Techniken bezüglich einer Operations-bedingten Verschlechterung der Sprachfunktion bisher noch unklar. / Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG-geförderten) Allianz- bzw. Nationallizenz frei zugänglich.

Es wird erwartet, dass die Integration der Magnetresonanztomografie (MRT) in die Protonentherapie die Treffgenauigkeit bei der Strahlentherapie für Krebserkrankungen deutlich verbessern wird. Besonders für Tumoren in beweglichen Organen des Thorax oder des Abdomens könnte die MRT-integrierte Protonentherapie (MRiPT) eine Synchronisierung der Bestrahlung mit der Tumorposition ermöglichen, was zu einer verminderten Normalgewebsdosis und weniger Nebenwirkungen führen könnte. Bis heute ist solch eine Integration jedoch aufgrund fehlender Studien zu potenziellen gegenseitigen Störeinflüssen dieser beiden Systeme nicht vollzogen worden. Diese Arbeit widmete sich zwei solcher Störeinflüsse, und zwar der Ablenkung des Protonenstrahls im Magnetfeld des MRT- Scanners, und umgekehrt, dem Einfluss der elekromagnetischen Felder der Protonentherapieanlage und des Protonenstrahls selbst auf die MRT-Bilder. Obwohl vorangegangene Studien den derzeitigen Konsens aufgezeigt haben, dass die Trajektorie eines abgebremsten Protonenstrahls im homogenen Phantom in einem transversalen Magnetfeld vorhersagbar ist, zeigte sich im quantitativen Vergleich der publizierten Modelle, der im ersten Teil dieser Arbeit vorgestellt wurde, dass die Vorhersagen dieser Modelle nur für eine begrenzte Anzahl von Kombinationen aus Magnetfeldstärke und Protonenenergie übereinstimmen. Die Schwächen bestehender analytischer Modelle wurden deshalb analysiert und quantifiziert. Kritische Annahmen und die mangelnde Anwendbarkeit auf realistische, d.h. inhomogene Magnetfeldstärken und Patientengeometrien wurden als Hauptprobleme identifiziert. Um diese zu überwinden, wurde ein neues semianalytisches Modell namens RAMDIM entwickelt. Es wurde gezeigt, dass dieses auf realistischere Fälle anwendbar und genauer ist als existierende analytische Modelle und dabei schneller als Monte-Carlo-basierte Teilchenspursimulationen. Es wird erwartet, dass dieses Modell in der MRiPT Anwendung findet zur schnellen und genauen Ablenkungsberechnung, zur Betrahlungsplanoptimierung und bei der MRT-geführten Strahlnachführung. In einem zweiten Schritt wurde die magnetfeldinduzierte Protonenstrahlablenkung in einem gewebeähnlichen Material durch Filmdosimetrie erstmalig gemessen und mit Monte-Carlo-Simulationen verglichen. In einem transversalen Magnetfeld einer Flussdichte von 0,95 T wurde experimentell gezeigt, dass die laterale Versetzung des Bragg-Peaks für Protonenenergien zwischen 80 und 180 MeV in PMMA zwischen 1 und 10 mm liegt. Die Retraktion des Bragg-Peaks war ≤ 0,5 mm. Es wurde gezeigt, dass die gemessene Versetzung des Bragg-Peaks innerhalb von 0,8 mm mit Monte-Carlo-basierten Vorhersagen übereinstimmt. Diese Ergebnisse weisen darauf hin, dass die Protonenstrahlablenkung durch Monte-Carlo-Simulationen genau vorhersagbar ist und damit der Realisierbarkeit der MRiPT nicht im Wege steht. Im zweiten Teil dieser Arbeit wurde erstmalig ein MRT-Scanner in eine Protonenstrahlführung integriert. Hierfür wurde ein offener Niederfeld-MRT-Scanner am Ende einer statischen Forschungsstrahlführung einer Protonentherapieanlage platziert. Die durch das statische Magnetfeld des MRT-Scanners hervorgerufene Strahlablenkung wurde bei der Ausrichtung des MRT-Scanners berücksichtigt. Die sequenzabhängigen, veränderlichen Gradientenfelder hatten keinen messbaren Einfluss auf das transversale Strahlprofil hinter dem MRT-Scanner. Die Magnetfeldhomogenität des Scanners lag innerhalb der Herstellervorgaben und zeigte keinen relevanten Einfluss von Rotationen der Protonengantry im benachbarten Bestrahlungsraum. Eine magnetische Abschirmung war zum gleichzeitigen Betrieb des MRT-Scanners und der Protonentherapieanlage nicht notwendig. Dies beweist die Machbarkeit gleichzeitiger Bestrahlung und Bildgebung in einem ersten MRiPT Aufbau. Die MRT-Bildqualität des Aufbaus wurde darauffolgend anhand eines angepassten Standardprotokolls aus Spin-Echo- und Gradienten-Echo-Sequenzen quantifiziert und es wurde gezeigt, dass die Bildqualität sowohl ohne als auch mit gleichzeitiger Bestrahlung hinreichend ist. Alle bestimmten geometrischen Parameter stimmten mit den physikalischen Abmessungen des verwendeten Phantoms innerhalb eines Bildpixels überein. Wie es für Niederfeld-MRT-Scanner üblich ist, war das Signal-Rausch-Verhältnis (SNR) der MRT-Bilder gering, was im Vergleich zu den Standardkriterien zu einer geringen Bildhomogenität und zu einem hohen Geisterbildanteil im Bild führte. Außerdem wurde aufgrund von Unsicherheiten in der Hochfrequenzkalibrierung des MRT-Scanners eine starke Schwankung der vertikalen Phantomposition mit einem Interquartilabstand von bis zu 1,5 mm beobachtet. T2*-gewichtete Gradientenechosequenzen zeigten zudem aufgrund von Magnetfeldinho- mogenitäten relevante ortsabhängige Bildverzerrungen. Es wurde gezeigt, dass die meisten Bildqualitätsparameter mit und ohne gleichzeitige Betrahlung äquivalent sind. Es wurde jedoch ein signifikanter Betrahlungseinfluss in Form von einer vertikalen Bildverschiebung und einer Verminderung des SNR beobachtet, die durch eine Änderung im Magnetfeld des MRT-Scanners erklärt werden können, welche durch zu diesem Feld parallel ausgerichtete Komponenten im Fernfeld der Strahlführungsmagneten hervorgerufen wird. Während das verminderte SNR vermutlich irrelevant ist (Dif- ferenz im Median ≤ 1,5), ist die sequenzabhängige Bildverschiebung (Differenz im Median bis zu 0,7 mm) nicht immer vernachlässigbar. Diese Ergebisse zeigen, dass die MRT-Bilder durch gleichzeitige Bildgebung nicht schwerwiegend verfälscht werden, dass aber eine dedizierte Optimierung der Hochfrequenzkalibrierung und der MRT-Bildsequenzen notwendig ist. Im letzten Teil der Arbeit wurde gezeigt, dass ein stromabhängiger Einfluss des Protonenstrahls auf MRT-Bilder eines Wasserphantoms durch zwei verschiedene MRT-Sequenzen messbar gemacht und zur Reichweiteverifikation genutzt werden kann. Der Effekt war in verschiedenen Flüssigkeiten, jedoch nicht in viskosen und festen Materialen, nachweisbar und wurde auf Hitzekonvektion zurückgeführt. Es wird erwartet, dass diese Methode in der MRiPT für Konstanztests der Protonenreichweite bei der Maschinenqualitätssicherung nützlich sein wird. Zusammenfassend hat diese Arbeit die Genauigkeit der Vorhersage der Strahlablenkung quantifiziert und verbessert, sowie Potenzial und Realisierbarkeit einer gleichzeitigen MRT-Bildgebung und Protonenbestrahlung gezeigt. Die weitere Entwicklung eines ersten MRiPT-Prototyps ist demnach gerechtfertigt.:List of Figures v List of Tables vii 1 General Introduction 1 2 State of the Art: Proton Therapy and Magnetic Resonance Imaging 3 2.1 Proton Therapy 4 2.1.1 Physical Principle 4 2.1.2 Beam Delivery 7 2.1.3 Motion Management and the Role of Image Guidance 10 2.2 Magnetic Resonance Imaging 14 2.2.1 Physical Principle 14 2.2.2 Image Generation by Pulse Sequences 18 2.2.3 Image Quality 21 2.3 MR-Guided Radiotherapy 24 2.3.1 Offline MR Guidance 24 2.3.2 On-line MR Guidance 25 2.4 MR-Integrated Proton Therapy 28 2.4.1 Aims of this Thesis 32 3 Magnetic Field-Induced Beam Deflection and Bragg Peak Displacement 35 3.1 Analytical Description 36 3.1.1 Review of Analytical Models 36 3.1.2 New Model Formulation 41 3.1.3 Evaluation of Analytical and Numerical Models 44 3.1.4 Discussion 51 3.2 Monte Carlo Simulation and Experimental Verification 54 3.2.1 Verification Setup 54 3.2.2 Monte Carlo Simulation 56 3.2.3 Experimental Verification 60 3.2.4 Discussion 61 3.3 Summary 63 4 Integrated In-Beam MR System: Proof of Concept 65 4.1 Integration of a Low-Field MR Scanner and a Static Research Beamline 65 4.1.1 Proton Therapy System 66 4.1.2 MR Scanner 66 4.1.3 Potential Sources of Interference 67 4.1.4 Integration of Both Systems 68 4.2 Beam and Image Quality in the Integrated Setup 70 4.2.1 Beam Profile 70 4.2.2 MR Magnetic Field Homogeneity 72 4.2.3 MR Image Quality - Qualitative In Vivo and Ex Vivo Test 74 4.2.4 MR Image Quality - Quantitative Phantom Tests 77 4.3 Feasibility of MRI-based Range Verification 86 4.3.1 MR Sequences 86 4.3.2 Proton Beam Parameters 88 4.3.3 Target Material Dependence 91 4.3.4 Discussion 92 4.4 Summary 96 5 Discussion and Future Perspectives 99 6 Summary/Zusammenfassung 105 6.1 Summary 105 6.2 Zusammenfassung 108 Bibliography I Supplementary Information XXIX A Beam Deflection: Experimental Measurements XXIX A.1 Setup XXIX A.2 Film Handling and Evaluation XXX A.3 Uncertainty Estimation XXX B Beam Deflection: Monte Carlo Simulations XXXIII B.1 Magnetic Field Model XXXIII B.2 Uncertainty Estimation XXXIV C Integrated MRiPT Setup XXXVI C.1 Magnetic Field Map XXXVI C.2 Sequence Parameters XXXVI C.3 Image Quality Parameters XLII C.4 Range Verification Sequences XLII / The integration of magnetic resonance imaging (MRI) into proton therapy is expected to strongly increase the targeting accuracy in radiation therapy for cancerous diseases. Especially for tumours situated in mobile organs in the thorax and abdomen, MR-integrated proton therapy (MRiPT) could enable the synchronisation of irradiation to the tumour position, resulting in less dose to normal tissue and reduced side effects. However, such an integration has been hindered so far by a lack of scientific studies on the potential mutual interference between the two components. This thesis was dedicated to two of these sources of interference, namely the deflection of the proton beam by the magnetic field of the MR scanner and, vice versa, alterations of the MR image induced by the electromagnetic fields of the proton therapy facility and by the beam itself. Although previous work has indicated that there is general consensus that the trajectory of a slowing down proton beam in a homogeneous phantom inside a transverse magnetic field is predictable, a quantitative comparison of the published methods, as presented in the first part of this thesis, has shown that predictions of different models only agree for certain proton beam energies and magnetic flux densities. Therefore, shortcomings of previously published analytical methods have been analysed and quantified. The inclusion of critical assumptions and the lack of applicability to realistic, i.e. non-uniform, magnetic flux densities and patient anatomies have been identified as main problems. To overcome these deficiencies, a new semi-analytical model called RAMDIM has been developed. It was shown that this model is both applicable to more realistic setups and less assumptive than existing analytical approaches, and faster than Monte Carlo based particle tracking simulations. This model is expected to be useful in MRiPT for fast and accurate deflection estimations, treatment plan optimisation, and MR-guided beam tracking. In a second step, the magnetic field-induced proton beam deflection has been measured for the first time in a tissue-mimicking medium by film dosimetry and has been compared against Monte Carlo simulations. In a transverse magnetic field of 0.95 T, it was experimentally shown that the lateral Bragg peak displacement ranges between 1 mm and 10 mm for proton energies between 80 and 180 MeV in PMMA. Range retraction was found to be ≤ 0.5 mm. The measured Bragg peak displacement was shown to agree within 0.8 mm with Monte Carlo simulations. These results indicate that proton beam deflection in a homogeneous medium is accurately predictable for intermediate proton beam energies and magnetic flux densities by Monte Carlo simulations and therefore not impeding the feasibility of MRiPT. In the second part of this thesis, an MR scanner has been integrated into a proton beam line for the first time. For this purpose, an open low-field MR scanner has been placed at the end of a fixed horizontal proton research beam line in a proton therapy facility. The beam deflection induced by the static magnetic field of the scanner was taken into account for alignment of the beam and the FOV of the scanner. The pulse sequence-dependent dynamic gradient fields did not measurably affect the transverse beam profile behind the MR scanner. The MR magnetic field homogeneity was within the vendor’s specifications and not relevantly influenced by the rotation of the proton gantry in the neighbouring treatment room. No magnetic field compensation system was required for simultaneous operation of the MR scanner and the proton therapy system. These results proof that simultaneous irradiation and imaging is feasible in an in-beam MR setup. The MR image quality of the in-beam MR scanner was then quantified by an adapted standard protocol comprising spin and gradient echo imaging and shown to be acceptable both with and without simultaneous proton beam irradiation. All geometrical parameters agreed with the mechanical dimensions of the used phantom within one pixel width. As common for low-field MR scanners, the signal-to-noise ratio (SNR) of the MR images was low, which resulted in a low image uniformity and a high ghosting ratio in comparison to the standardised test criteria. Furthermore, a strong fluctuation of the vertical phantom position due to uncertainties in the pre-scan frequency calibration was observed, with an interquartile range of up to 1.5 mm. T2*-weighted gradient echo images showed relevant nonuniform deformations due to magnetic field inhomogeneities. Most image quality parameters were shown to be equivalent with and without simultaneous proton beam irradiation. However, a significant influence of simultaneous irradiation was observed as a shift of the vertical phantom position and a decrease in the SNR, both of which can be explained by a change in the B0 field of the MR scanner induced by components of the fringe field of the beam line magnets directed parallel to B0 . While the decrease in SNR is not expected to be relevant (median differences were within 1.5 ), the sequence-dependent phantom shift (median differences of up to 0.7 mm) can become non-negligible. These results show that the MR images are not severely distorted by simultaneous irradiation, but a dedicated optimisation of the pre-scan RF calibration and the MR sequences is required for MRiPT. Lastly, a current-dependent influence of the proton beam on the MR image was shown to be measurable in water in two different MR sequences, which allowed for range verification measurements. The effect was observed in different liquids but not in highly viscose and solid materials, and most probably induced by heat convection. This method is expected to be useful in MRiPT for consistency tests of the proton range during machine-specific quality assurance. In conclusion, this work has improved and quantified the accuracy of beam deflection predictions and shown the feasibility and potential of in-beam MR imaging, justifying further research towards a first MRiPT prototype.:List of Figures v List of Tables vii 1 General Introduction 1 2 State of the Art: Proton Therapy and Magnetic Resonance Imaging 3 2.1 Proton Therapy 4 2.1.1 Physical Principle 4 2.1.2 Beam Delivery 7 2.1.3 Motion Management and the Role of Image Guidance 10 2.2 Magnetic Resonance Imaging 14 2.2.1 Physical Principle 14 2.2.2 Image Generation by Pulse Sequences 18 2.2.3 Image Quality 21 2.3 MR-Guided Radiotherapy 24 2.3.1 Offline MR Guidance 24 2.3.2 On-line MR Guidance 25 2.4 MR-Integrated Proton Therapy 28 2.4.1 Aims of this Thesis 32 3 Magnetic Field-Induced Beam Deflection and Bragg Peak Displacement 35 3.1 Analytical Description 36 3.1.1 Review of Analytical Models 36 3.1.2 New Model Formulation 41 3.1.3 Evaluation of Analytical and Numerical Models 44 3.1.4 Discussion 51 3.2 Monte Carlo Simulation and Experimental Verification 54 3.2.1 Verification Setup 54 3.2.2 Monte Carlo Simulation 56 3.2.3 Experimental Verification 60 3.2.4 Discussion 61 3.3 Summary 63 4 Integrated In-Beam MR System: Proof of Concept 65 4.1 Integration of a Low-Field MR Scanner and a Static Research Beamline 65 4.1.1 Proton Therapy System 66 4.1.2 MR Scanner 66 4.1.3 Potential Sources of Interference 67 4.1.4 Integration of Both Systems 68 4.2 Beam and Image Quality in the Integrated Setup 70 4.2.1 Beam Profile 70 4.2.2 MR Magnetic Field Homogeneity 72 4.2.3 MR Image Quality - Qualitative In Vivo and Ex Vivo Test 74 4.2.4 MR Image Quality - Quantitative Phantom Tests 77 4.3 Feasibility of MRI-based Range Verification 86 4.3.1 MR Sequences 86 4.3.2 Proton Beam Parameters 88 4.3.3 Target Material Dependence 91 4.3.4 Discussion 92 4.4 Summary 96 5 Discussion and Future Perspectives 99 6 Summary/Zusammenfassung 105 6.1 Summary 105 6.2 Zusammenfassung 108 Bibliography I Supplementary Information XXIX A Beam Deflection: Experimental Measurements XXIX A.1 Setup XXIX A.2 Film Handling and Evaluation XXX A.3 Uncertainty Estimation XXX B Beam Deflection: Monte Carlo Simulations XXXIII B.1 Magnetic Field Model XXXIII B.2 Uncertainty Estimation XXXIV C Integrated MRiPT Setup XXXVI C.1 Magnetic Field Map XXXVI C.2 Sequence Parameters XXXVI C.3 Image Quality Parameters XLII C.4 Range Verification Sequences XLII

京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第18139号 / 医博第3859号 / 新制||医||1002(附属図書館) / 30997 / 京都大学大学院医学研究科医学専攻 / (主査)教授 武藤 学, 教授 武田 俊一, 教授 富樫 かおり / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM