Performances d’un dispositif d’imagerie ultrasonore pour le suivi des mouvements prostatiques en radiothérapie

摘要

The emergence of hypofractionated treatments implies an increase of precision on the dose delivery. Organs motions between and during the irradiation fraction are a limiting factor for achieving a high quality treatment of mobile tumor localizations. In particular, hypofractionation was demonstrated as a promising strategy for prostate cancer treatment, whereas a high digestive toxicity cannot be excluded.During this thesis, motions of the target volume were quantified during radiotherapy delivered for treating a prostate cancer, on the prostatic gland or after a prostatectomy. An innovative imaging modality was used, based on two different ultrasound (US) probes: a transabdominal (TA) probe and a transperineal (TP) probe. Contrary to US systems previously commercialized, motions between simulation and treatment sessions were quantified with a monomodal registration US/US. Moreover, intrafraction motions of the target volume were recorded with the TP probe, since images can be continuously acquired with this probe. A methodology for evaluating the performances of the TA and TP probes were developed for the correction of interfractions motions, in clinical conditions. US data were compared with the current reference modality, the Cone Beam Computed Tomography (CBCT). Moreover, the uncertainties on the target localization due to the variability of the probe pressure and to inter-operator variability of the manual US/US registration were quantified for the first time. Poor agreements of the interfraction motions measurements were found between US and CBCT modalities, using the TA probe. Furthermore, a high variability of the TA probe pressure and of the inter-operator registration was observed. Using the TP probe, a corrective method of the US reference image localization was developed and validated, leading to a high agreement with the CBCT modality.Intrafraction motions were observed with the TP probe, and were highly patient and session dependent. A dosimetric study was proposed to evaluate the impact of the intrafraction motion on the dose delivery. The effects were patient dependent and more important when the treatment was simulated using a hypofractionated strategy.