Comparing different boost concepts and beam configurations for proton therapy of pancreatic cancer

Interfractional geometrical and anatomical variations impact the accuracy of proton therapy for pancreatic cancer. This study investigated field-in-field (FIF) and simultaneous integrated boost (SIB) concepts for scanned proton therapy treatment with different beam configurations.

Robustly optimized treatment plans for fifteen patients were generated using FIF and SIB techniques with two, three, and four beams. The prescribed dose in 20 fractions was 60 Gy(RBE) for the internal gross tumor volume (IGTV) and 46 Gy(RBE) for the internal clinical target volume. Verification computed tomography (vCT) scans was performed on treatment days 1, 7, and 16. Initial treatment plans were recalculated on the rigidly registered vCTs. V_100% and D_95% for targets and D_2cm³ for the stomach and duodenum were evaluated. Robustness evaluations (range uncertainty of 3.5 %) were performed to evaluate the stomach and duodenum dose-volume parameters.

For all techniques, IGTV V_100% and D_95% decreased significantly when recalculating the dose on vCTs (p < 0.001). The median IGTV V_100% and D_95% over all vCTs ranged from 74.2 % to 90.2 % and 58.8 Gy(RBE) to 59.4 Gy(RBE), respectively. The FIF with two and three beams, and SIB with two beams maintained the highest IGTV V_100% and D_95%. In robustness evaluations, the ΔD_2cm³ of stomach was highest in two beams plans, while the ΔD_2cm³ of duodenum was highest in four beams plans, for both concepts.

Target coverage decreased when recalculating on CTs at different time for both concepts. The FIF with three beams maintained the highest IGTV coverage while sparing normal organs the most.