A novel approach for evaluation of prostate deformation and associated dosimetric implications in IGRT of the prostate.
Essa Mayyas,Jinkoo Kim,Sanath Kumar,Chang Liu,Ning Wen,Benjamin Movsas,Mohamed A Elshaikh,Indrin J Chetty
DOI: https://doi.org/10.1118/1.4893196
IF: 4.506
2014-01-01
Medical Physics
Abstract:Purpose: Prostate deformation is assumed to be a secondary correction and is typically ignored in the planning target volume (PTV) margin calculations. This assumption needs to be tested, especially when planning margins are reduced with daily image-guidance. In this study, deformation characteristics of the prostate and seminal vesicles were determined, and the dosimetric impact on treatment plans with different PTV margins was investigated. Methods: Ten prostate cancer patients were retrospectively selected for the study, each with three fiducial markers implanted in the prostate. Two hundred CBCT images were registered to respective planning CT images using a B-spline-based deformable image registration (DIR) software. A manual bony anatomy-based match was first applied based on the alignment of the pelvic bones and fiducial landmarks. DIR was then performed. For each registration, deformation vector fields (DVFs) of the prostate and seminal vesicles (SVs) were quantified using deformation-volume histograms. In addition, prostate rotation was evaluated and compared with prostate deformation. For a patient demonstrating small and large prostate deformations, target coverage degradation was analyzed in each of three treatment plans with PTV margins of 10 mm (6 mm at the prostate/rectum interface), as well as 5, and 3 mm uniformly. Results: Deformation of the prostate was most significant in the anterior direction. Maximum prostate deformation of greater than 10, 5, and 3 mm occurred in 1%, 17%, and 76% of the cases, respectively. Based on DVF-histograms, DVF magnitudes greater than 5 and 3 mm occurred in 2% and 27% of the cases, respectively. Deformation of the SVs was most significant in the posterior direction, and it was greater than 5 and 3 mm in 7.5% and 44.9% of the cases, respectively. Prostate deformation was found to be poorly correlated with rotation. Fifty percent of the cases showed rotation with negligible deformation and 7% of the cases showed significant deformation with minimal rotation (<3 degrees). Average differences in the D-95 dose to the prostate + SVs between the planning CT and CBCT images was 0.4% +/- 0.5%, 3.0% +/- 2.8%, and 6.6% +/- 6.1%, respectively, for the plans with 10/6, 5, and 3 mm margins. For the case with both a large degree of prostate deformation (approximate to 10% of the prostate volume) and rotation (approximate to 8 degrees), D95 was reduced by 0.5% +/- 0.1%, 6.8% +/- 0.6%, and 20.9% +/- 1.6% for 10/6, 5, and 3 mm margin plans, respectively. For the case with large prostate deformation but negligible rotation (<1 degrees), D95 was reduced by 0.4 +/- 0.3, 3.9 +/- 1.0, and 11.5 +/- 2.5 for 10/6, 5, and 3 mm margin plans, respectively. Conclusions: Prostate deformation over a course of fractionated prostate radiotherapy may not be insignificant and may need to be accounted for in the planning margin design. A consequence of these results is that use of highly reduced planning margins must be viewed with caution. (C) 2014 American Association of Physicists in Medicine.