This study introduces a radiation therapy treatment planning (RTP) workflow for prostate cancer patients that relies solely on magnetic resonance imaging (MRI) -based RTP images throughout the process. Scanner-specific intensity values of T1/T2*-weighted in-phase MR images and Hounsfield units (HUs) of computed tomography (CT) images in the male pelvis were analyzed. The results were utilized to generate a dual model conversion technique transforming the MRI data into pseudo-HUs by separate conversion models within and outside of bone segment. The technique was implemented by constructing pseudo-CT images for 15 prostate cancer patients. The feasibility of the obtained images for RTP was evaluated by comparisons against the standard CTs. Dose distributions were analyzed e.g. with volumetric-modulated arc therapy calculated by means of a Monte Carlo algorithm. The feasibility of MRI-based reference images for image guidance was investigated for patient position verification with X-ray -based localization images. The potential susceptibility-induced bone outline shift in MR images and the effect of different bone parts to the absorbed dose were quantified with a dedicated fresh bone phantom. The local soft- and bony tissue presentation uncertainties in the pseudo-CTs were on average 11 HUs and 99 HUs, respectively. The prostate dose level differences between those in pseudo-CTs and those in CTs ranged from -1.0% to 0.8%. The percentages of dose points in the body passing the 1 mm and 1% 2D gamma index criteria between the images ranged from 79% to 100%. The maximum dose distribution inconsistency behind the bones was 1.3% measured for a single 6 MV radiation field. The bone outlines in the MR images were illustrated correctly within a 1 mm-pixel size, but each 1 mm-sized systematic error in bone segmentation resulted in ~0.4% change to the prostate dose level in the pseudo-CTs. The SDs of differences between cone-beam CT (CBCT)-to-pseudo and CBCT-to-CT automatic registrations were ≤1.0 mm and ≤0.7°. The SDs of differences between pseudo-digitally reconstructed radiograph (DRR) and CT-DRR -based manual registrations with planar localization were ≤1.0 mm (kV) and ≤1.7 mm (MV). This work shows that it is possible to construct heterogeneous pseudo-CTs of the pelvis by transforming the intensity values of a single MR image into pseudo-HUs, and that by adopting these images for dose calculation and image guidance, the entire MRI-based RTP workflow for prostate cancer patients can be conducted according to required standards of modern RT.
|Translated title of the contribution||Magneettikuvaus -pohjainen sädehoito|
|Publication status||Published - 2015|
|MoE publication type||G5 Doctoral dissertation (article)|
- MRI-based RT
- treatment planning
- magnetic resonance imaging