Room: Room 207
Purpose: To propose a method using CBCT to monitor beam-specific water equivalent path length (WEPL) variations in pediatric proton therapy.
Methods: Phantom-based measurements and a retrospective analysis of 10 pediatric patients receiving proton therapy were performed. CT images of an anthropomorphic phantom representing a treatment planning scan were deformed to CBCT images of the same phantom with bolus added to simulate body circumference changes. Deformation accuracy was independently verified with commercial software using a set of metrics recommended by TG-132. The deformed-CT (synthetic-CT) was compared to CT of the phantom with bolus (ground truth) to evaluate for differences in WEPL and target coverage. Ten children with pelvic tumors were selected that underwent repeat-CT for replanning due to significant inter-fractional anatomy changes. On the days when repeat-CT was performed, the original planning CT was deformed to the same-day CBCT to generate a synthetic-CT to evaluate its accuracy as a surrogate to repeat-CT.
Results: The phantom study showed an excellent agreement with only 1.0mm ± 0.2mm difference in WEPL between the ground truth-CT and synthetic-CT. The CTV D95 was 98.4% when calculated for the ground truth-CT and synthetic-CT. Analysis of the pediatric pelvic tumor cases showed similar deviations in WEPL between the paired initial-CT and repeat-CT and the pair of initial-CT and the synthetic-CT, with a 75th percentile of 5.1mm and 4.8mm, respectively. The results were based on 20 proton fields. Coverage degradation due to proton path length changes reduced target coverage (CTV Dmin) by a mean value of 12.1% and 11.3%, respectively, when calculated on repeat-CT and synthetic-CT.
Conclusion: The proposed method was found to be accurate for clinical monitoring of proton range variations and may be used in adaptive therapy to eliminate unnecessary CT studies in pediatric proton therapy patients.
Authors:
Ozgur Ates, Li Zhao, Chia-Ho Hua, Jinsoo Uh, Fakhriddin Pirlepesov, Weiguang Yao, Matthew Krasin, and Thomas Merchant
Not Applicable / None Entered.
Not Applicable / None Entered.