Organ motion and inaccurate patient positioning may compromise radiation therapy outcome. With the aid of image guidance, it is possible to allow for a more accurate organ motion and motion control study, which could lead to the reduction of irradiated healthy tissues and possible dose escalation to the target volume to achieve better treatment results. The studies on the organ motion and image guidance were divided into the following four sections. The first, the interfractional setup uncertainties from day-to-day treatment and intrafractional internal organ motion within the daily treatment from five different anatomic sites were studied with Helical TomoTherapy unit. The pre-treatment mega voltage computed tomography (MVCT) provided the real-time tumor and organ shift coordinates, and can be used to improve the accuracy of patient positioning. The interfractional system errors and random errors were analyzed and the suggested margins for HN, brain, prostate, abdomen and lung were derived. The second, lung stereotactic body radiation therapy using the MIDCO BodyLoc whole body stereotactic localizer combined with TomoTherapy MVCT image guidance were investigated for the possible target and organ motion reduction. The comparison of 3D displacement with and without BodyLoc immobilization showed that, suppression of internal organ motion was improved by using BodyLoc in this study. The third, respiration related tumor motion was accurately studied with the four dimensional computed tomography (4DCT). Deformable registration between different breathing phases was performed to estimate the motion trajectory for lung tumor. Optimization is performed by minimizing the mean squared difference in intensity, and is implemented with a multi-resolution, gradient descent procedure. The fourth, lung tumor mobility and dosimetric benefits were compared with different PTV obtained from 3DCT and 4DCT. The results illustrated that the PTV3D not only included excess normal tissues but also might result in missed target tissue. The normal tissue complication probability (NTCP) from 4D plan was statistically significant smaller than 3D plan for both ipsilateral lung and heart.

Cervix cancer affects women of all ages and causes significant morbidity and mortality. Locally advanced disease is curable with radiotherapy (RT) in about 50% of patients, although often at the expense of serious side effects. In order to improve the therapeutic ratio of tumour control versus normal tissue toxicity, conformal intensity-modulated radiotherapy (IMRT) is being investigated. However, inter- and intra-fractional motion of cervix cancer can contribute to both geographical miss of the target and overdosing of surrounding normal tissues, particularly in the setting of conformal IMRT with steep dose gradients. Defining the target volume accurately and understanding the dose consequence of these complex intra-pelvic organ dynamics during external beam radiotherapy forms the essential foundations for future treatment optimization and adaptation. This in turn will lead to improvements in tumour control and disease-free survival while minimising treatment toxicity.

Cervix cancer affects women of all ages and causes significant morbidity and mortality. Locally advanced disease is curable with radiotherapy (RT) in about 50% of patients, although often at the expense of serious side effects. In order to improve the therapeutic ratio of tumour control versus normal tissue toxicity, conformal intensity-modulated radiotherapy (IMRT) is being investigated. However, inter- and intra-fractional motion of cervix cancer can contribute to both geographical miss of the target and overdosing of surrounding normal tissues, particularly in the setting of conformal IMRT with steep dose gradients. Defining the target volume accurately and understanding the dose consequence of these complex intra-pelvic organ dynamics during external beam radiotherapy forms the essential foundations for future treatment optimization and adaptation. This in turn will lead to improvements in tumour control and disease-free survival while minimising treatment toxicity.

La chimioradiothérapie est le traitement des cancers de l’œsophage localement évolués et inopérables. Dans cette indication, la radiothérapie conformationnelle est utilisée couramment. Cependant, le pronostic de ces patients reste sombre. L’intérêt de la radiothérapie asservie à la respiration (RAR) a déjà été montré notamment dans le traitement des cancers pulmonaires, mammaires et hépatiques : elle permet de diminuer l’irradiation des tissus sains, et d’envisager une augmentation de dose au volume tumoral. Afin d’améliorer la prise en charge radiothérapique, nous proposons d’étudier la faisabilité de la RAR dans le traitement des cancers de l’œsophage. Nous étudierons la mobilité des cancers oesophagiens au cours de la respiration pour optimiser la définition des volumes cibles et notamment de la marge interne (IM). Nous analyserons la corrélation existant entre les mouvements tumoraux et les mouvements de la paroi thoracique afin de montrer que le mouvement des tumeurs oesophagiennes est induit par la respiration, pré-requis indispensable à l’utilisation des systèmes d’asservissement en respiration libre. Nous utiliserons différents outils d’analyse dosimétrique pour évaluer l’apport de la RAR dans le traitement des cancers de l’œsophage en comparant les plans dosimétriques effectués à différents temps respiratoires (fin d’expiration, fin d’inspiration et inspiration forcée) avec le plan dosimétrique effectué en respiration libre pour la même tumeur. Ceci nous permettra de quantifier le gain obtenu par la RAR et de déterminer la meilleure « fenêtre » de traitement au cours du cycle respiratoire en fonction des différents systèmes d’asservissement disponibles. Cette analyse dosimétrique sera complétée par un calcul de l’Equivalent de Dose Uniforme (EUD), dans sa forme linéaire quadratique, pour les différents volumes d’intérêt. Nous déterminerons au préalable ses conditions d’utilisation dans une étude théorique de dégradation des HDV / Chemoradiotherapy is now the standard treatment for locally advanced or inoperable esophageal carcinoma. In this indication, conformal radiotherapy is generally used. However, prognosis remains poor for these patients.Respiratory gating radiotherapy can decrease healthy tissus irradiation and allows escalation dose in lung, liver and breast cancer. In order to improve radiotherapy technique, we propose to study the feasibility of respiratory gating for esophageal cancer.We will study the respiratory motions of esophageal cancer to optimize target volume delineation, especially the internal margin (IM).We will test the correlation between tumour and chest wall displacements to prove that esophageal cancer motions are induced by respiration. This is essential before using free breathing respiratory gating systems.We will work out the dosimetric impact of respiratory gating using various dosimetric analysis parameters. We will compare dosimetric plans at end expiration, end inspiration and deep inspiration with dosimetric plan in free-breathing condition. This will allow us to establish the best respiratory phase to irradiate for each gating system.This dosimetric study will be completed with linear quadratic equivalent uniform dose (EUD) calculation for each volume of interest. Previously, we will do a theoretical study of histogram dose volume gradation to point up its use

La prostate est un organe pelvien qui joue un rôle important dans son environnement anatomique, notamment en assurant la sécrétion d’un liquide essentiel dans la composition du liquide séminal. Le cancer de la prostate représente la première cause mortalité chez l’homme à un âge avancé. Ce travail concerne le développement par la méthode des éléments finis d'un modèle numérique du mouvement des organes pelviens (prostate, vessie, rectum) et de leurs interactions. L’objectif est la réduction des marges d'irradiation du volume cible au cours d'une séance de radiothérapie, afin de ne pas altérer les organes sains avoisinants. Les marges choisies sont importantes compte tenu du fait que la prostate subi des déplacements importants lors des interactions permanentes avec les organes avoisinants. Un premier modèle est construit à partir de la géométrie des organes pelviens générée à partir d'images scanner acquises sur patients. Des lois hyper-élastiques sont adoptées pour modéliser le comportement mécanique du rectum et de la vessie et un comportement Hookéen est considéré pour la prostate. Les paramètres physiques du modèle sont déterminés à partir de la littérature, des données expérimentales et de nos propres mesures. Les conditions aux limites cinématiques et statiques (pressions de distension intra-vésicale et intra-rectale) sont définies à partir des observations anatomiques et reflètent la présence de l’entourage anatomique et les conditions de chargement. Des comparaisons entre les variations de forme et de position d'organes obtenues par simulation et les mesures obtenues par imagerie scanner (Keros et al. ; 2006) montrent des amplitudes de déplacements proches, avec des écarts variant entre 8% et 11%. Un modèle prenant en compte la variabilité des paramètres physiques inter et intra patients est envisagé en perspectives / The prostate plays an important biological role in the human body, such as secretion of some prostatic liquid essential in the semen composition. Prostate cancer is the first cause of mortality for men at an advanced age. The prostate motion due to the interactions with the surrounding anatomic entities is difficult to predict, hence important margins are usually adopted during X-ray irradiation, in order not to damage the surrounding healthy organs (bladder and rectum). The principal objective of this work is to set up a FE model of the motion and deformation of the human pelvic organs in order to reduce the margins. A first model is constructed from CT-scans of the human pelvic organs, allowing the generation of the organ geometrics. Hyperelastic modeling of the bladder and rectum behaviors were considered whereas a Hookean model was retained for the prostate. The model parameters are fixed by adopting literature data, experimental data (from CHU-Nancy) and experimental measurements achieved on pig. Boundary conditions are defined according either surrounding anatomy kinematic constraints or internal pressures that correspond to the bladder and rectum repletion’s. Simulated displacements show order of magnitudes of the prostate motion very close to measurements carried out by Keros et al. (2006) on a deceased person, with a relative error ranging from 8% to 11%. Those differences are essentially due to the variability in the physical parameters, pointing out the need for a statistical approach in order to take into account the material, geometrical and loading variability related to a panel of patients