Robustness of daily dose for each beam angle and accumulated dose for inter-fractional anatomical changes in passive carbon-ion radiotherapy for pancreatic cancer: bone matching versus tumor matching

Abstract Purpose We aimed to assess the robustness of accumulated dose distributions for inter-fractional changes in passive carbon-ion radiotherapy for pancreatic cancer. Methods Ninety-five daily CT image sets acquired after the treatment of eight patients with pancreatic cancer were used in this prospective study. Dose distributions with treatment beam fields were recalculated for bone matching (BM) and tumor matching (TM) positions on all daily CT images, the accumulated doses being calculated using deformable image registration methods. The prescribed dose was 55.2 Gy (relative biological effectiveness [RBE]) in 12 fractions. Dose volume parameters of V95 (%) for CTV and GTV, and D2cc (Gy(RBE)) for the stomach and duodenum were evaluated. Results The medians (range) of CTV V95 (%) were 91.9 (86.1–100.0), 80.5 (56.1–90.6), and 86.4 (72.5–96.5) for the Plan, accumulated with BM and TM, respectively; GTV values (%) were 98.0 (85.7–100.0), 93.3 (65.7–99.9), and 96.2 (84.8–100.0), respectively. There were significant differences between all combinations apart from the Plan and TM for both targets. The values of stomach D2cc (Gy(RBE)) were 36.0 (16.9–43.4), 36.7 (17.9–45.0), and 35.2 (16.8–43.5), respectively; duodenum values (Gy(RBE)) were 25.2 (21.3–40.3), 30.1 (23.3–48.6), and 28.3 (20.4–50.6), respectively. There was a significant difference between the Plan and BM for duodenum only. Conclusions TM is recommended over BM because it can achieve higher target dose coverage than BM. Nevertheless, it is not enough in some cases. Further technical improvements are necessary to improve the target dose coverage.