Room: Davidson Ballroom A
Purpose: To develop and evaluate an automated planning method for near real-time adaptive intensity-modulated proton therapy in prostate cancer that accounts for daily density changes along the pencil-beam paths and shape changes of target and organs-at-risk.
Methods: Starting point is a treatment plan optimized on the planning CT. To account for density changes in a daily CT, the proposed method first adapts the energy of the pencil-beams to restore the original Bragg-peak location. Then, adding pencil-beams and optimizing all pencil-beam weights using a fast Reference-Point-Method results in a Pareto optimal plan for the daily anatomy, with similar trade-offs to those in the initial plan. The method was developed and evaluated using 8-10 daily CTs of 11 prostate cancer patients (88 CTs in total). The proposed method was compared to a previously published restoration method using simple pencil-beam weight optimization after the restoration step. The method was also benchmarked against a full, computation-intensive multi-criteria optimization method.
Results: Clinically acceptable target coverage (V95%â‰¥98% and V107%â‰¤2%) was obtained for 85/88 CTs compared with 45/88 CTs using the restoration method. Moreover, the medians of the rectum and bladder mean dose were reduced by the adaptation method by 6.2Gy and 4.7Gy, respectively. The largest improvements were observed for the rectum and bladder V45Gy, with median reductions of 10.3%-point and 10.8%-point, respectively. The near-real time adaptation took on average 2.9 minutes (range 2.3 - 3.3). In the absence of time constraints, a full multi-criteria optimization could further reduce the medians of the rectum and bladder mean dose by 6.7Gy and 9.1Gy, respectively. However, the proposed method did already obtain optimal results for the high-dose parameters.
Conclusion: The high acceptance rate, speed, and automation of the adaptive planning method demonstrates the feasibility of daily adapted proton therapy treatment plans that maximally conform the dose to the target.
Funding Support, Disclosures, and Conflict of Interest: This study was financially supported by ZonMw, the Netherlands Organization for Health Research and Development, grant number 104003012 and by Varian Medical Systems. Erasmus MC Cancer Institute also has research collaborations with Elekta AB, Stockholm, Sweden and Accuray Inc, Sunnyvale, USA.