Indirect MRI planning for prostate external beam radiotherapy without the use of implanted ducials


Objectives: The aim of this study is to assess the feasibility of the integration of MRI into the prostate radiotherapy planning pathway. This includes quantifi cation of the inter-clinician and inter-modality variation in the size and position of the delineated prostate on CT and MR, evaluation of CT-MRI image registration methods and investigation into the potential reduction in bladder and rectum dose from planning on the MRI delineated prostate. Materials and Methods: Three patients with histologically proven cancer of the prostate underwent pre-treatment CT and MRI imaging. The prostate was delineated on each scan by four clinicians. The inter-observer variation for the whole prostate, prostatic base and apex was quanti ed using a variety of metrics. Three methods of image registration were tested to fuse the CT and MRI images, and evaluated to find the method that minimised the inter-modality variation at the base of the prostate. Treatment plans were prepared for each delineation of the prostate to quantify the possible dose reduction to the bladder and rectum through use of the MR prostate delineation. Results: A signi cant di erence in the prostate volume as delineated on CT or MRI was found (mean volume 30.9cm3 and 24.2cm3 respectively, p=0.0002). The mean reduction in volume from CT to MRI was 26%. The inter-observer variation was signi cantly smaller at the base of the prostate than the apex for both modalities. Auto-registration using a volume of interest around the base of the prostate minimised the inter-modality variation at the prostatic base, and this was used to fuse the CT and MRI images for treatment planning. Statistically signi cant volume reduction was achieved at all DVH points for the OARs (p<0.05) including clinically signifi cant volume reductions for the rectum V30, V40, V50, V60 and V70. Conclusion: The inter-observer variation was found to be signifi cantly smaller at the prostatic base than the apex indicating that CT-MRI coregistration is most reliable at the base of the gland. Due to a smaller delineated prostate on T2 weighted MRI, incorporating MRI imaging into the radiotherapy planning pathway by registering CT and MRI images leads to signi cant dose sparing to the rectum.