ESTRO 38 Abstract book

S210 ESTRO 38

1 University Medical Center Utrecht, Department of Radiation Oncology, Utrecht, The Netherlands ; 2 Netherlands Cancer Insitute, Department of Radiation Oncology, Amsterdam, The Netherlands Purpose or Objective We have collected and analyzed an extensive 3D cine-MR dataset to study the visualization of intrafraction motion of the prostate and surrounding organs in MR-guided RT. Here we report on a remarkable relationship of prostate intrafraction motion with initial bladder volume. Material and Methods Fifty-nine patients had repeated weekly cine-MR imaging sessions over five weeks in a multicenter Medical Ethics Review Board approved study. Each session was scheduled shortly after a radiotherapy fraction. 30 patients were instructed to drink 400 mL prior to irradiation as well as scanning, while 29 patients were instructed to drink 200 mL only prior to irradiation. A cine-MR session consisted of 55 sequentially obtained 3D datasets (‘dynamics’), using a balanced 3D gradient echo sequence. Each dynamic was acquired over a 11 second period, with the complete cine- MR session covering a ten minute period, similar to the duration of RT fraction delivery. The bladder volume (BV) in each dynamic was obtained with an in-house developed bladder segmentation algorithm. In total, 238 sessions were available for analysis. Prostate intrafraction motion results were obtained with an in-house developed automatic soft-tissue tracking algorithm that was validated with an independent marker tracking algorithm (success rate > 98%, ESTRO 37). We assigned cine-MR sessions to either a group with a BV in the first dynamic below the overall mean BV (Group 1) or to a group with a BV above the group mean (Group 2). We analyzed differences in prostate intrafraction behavior between these two groups. Results Deviations between the automatic segmentation of the bladder and manual delineations were typically < 10 mL. The mean bladder volume ± SD was 0.17 ± 0.07 L, 0.12 ± 0.03 L and 0.24± 0.05 L in respectively all sessions, group 1 (135 sessions) and group 2 (103 sessions). BV increase during the ten minute sessions was small (0.014 ± 0.013 L). The distributions of intrafraction translations in AP as well as CC directions were significantly different between groups 1 and 2 for all time points > 2 minutes (p < 0.05, Kolmogorov-Smirnov test). Particularly the magnitude of motions in the AP direction was markedly larger in group 2 (Fig. 1). The distributions of 3D displacements were equally different between the two groups, as illustrated by Fig. 2. The frequency of displacements > 5 mm almost doubled in group 2 relative to group 1 near the end of the sessions. Conclusion We have found a significant and clinically meaningful increase of prostate intrafraction motion in patients with an initial high bladder volume. Our working hypothesis is that this is due to increased discomfort for larger BV. The presented results suggest that caution should be applied with too strict drinking instructions for prostate cancer patients as full bladder conditions may deteriorate targeting accuracy even after daily on-line corrections.

OC-0409 Comparison of different strategies to derive time-resolved volumetric MRI in MRI-guided radiotherapy C. Paganelli 1 , S. Portoso 1 , N. Garau 1 , G. Meschini 1 , R. Via 2 , P. Keall 3 , M. Riboldi 4 , G. Baroni 1 1 Politecnico di Milano, Dipartimento di Elettronica- Informazione e Bioingegneria, Milano, Italy; 2 Paul Scherrer Institute, Center for Proton Therapy, Villigen, Switzerland; 3 University of Sydney, ACRF Image X Institute- Sydney Medical School, Sydney, Australia ; 4 Ludwig-Maximilians-Universität München, Department of Medical Physics, Munich, Germany Purpose or Objective MRI-Linacs enable 2D cine-MRI centred in the tumour for motion detection in radiotherapy. However, they lack 3D information due to temporal-spatial limits. To derive time-resolved 3D information, different approaches have been proposed in the literature, but a rigorous comparison between these strategies has not yet been performed. The goal of this study is to quantitatively investigate published strategies to derive time-resolved volumetric MRI in MRI-

guided radiotherapy. Material and Methods

Five published strategies to derive time-resolved 3DMRI were selected: propagation (PROP) 1 , out-of-plane motion