Room: Davidson Ballroom A
Purpose: To assess the effectiveness of a dynamic adaptive system using CT-based three-dimensional image guidance in prostate cancer.
Methods: A retrospective analysis of five prostate cancer patients who received proton beam therapy (PBT) was performed. The in-room CT images were acquired once every three fractions. The original plan (OP) was generated using four fields in a passive scattering method. An extra margin to cover uncertainty in the range was added to the PTV. Bone registration (BR) was performed on each pretreatment daily CT image dataset. In this study, tumor registration (TR) was retrospectively performed. In addition, the system automatically adapted the range shifter (range adaptation method, RA) after the TR. Dose distributions for the three methods were calculated using the simplified Monte Carlo algorithm. The three dose distributions were compared to the dosimetric parameters of the OP. The following parameters were compared: D95%, D2%, HI, and mean dose for the CTV. The parameters including mean dose, D2%, V30%, V60%, V80%, and V100% for the OARs, the rectum and bladder, were also compared.
Results: The TR method achieved better CTV coverage than the BR method (average differences from the OP: CTV D95% = -1.2 Â± 2.1 %, 0.1 Â± 0.1 %, and 0.5 Â± 0.5 % for the BR, TR, and RA, respectively). For the OARs, the TR and RA methods achieved similar dosimetric parameters when compared with the OP. The maximum dose variation for the rectum and bladder using the BR were -30.5 % and 27.1 %, respectively.
Conclusion: Our results suggest that the TR method should be applied to achieve accurate dose delivery in PBT. In this study, the RA method showed comparable target dose coverage to the TR method, along with OARs dose-sparing effect because of the addition of an extra margin to cover uncertainty in the range.
Funding Support, Disclosures, and Conflict of Interest: This research was supported by Grant-in-Aid for JSPS Research Fellow.