Abstract
PURPOSE: To investigate the impact of interfractional anatomical changes and setup correction methods on dose distributions in pancreatic cancer patients under breath-hold conditions. METHODS: Three intensity-modulated proton therapy (IMPT) plans with different beam arrangements and one volumetric-modulated arc therapy (VMAT) plan prescribing 54 Gy in 30 fractions were created for 10 patients who underwent three additional CT scans performed at an interval of 1-2 weeks. The additional CT sets were rigidly registered to the simulation CT set using both bone-matching (BM) and organ-matching (OM) methods in each patient. Recalculated dose distributions and dose-volume indices on the additional CT sets using either the BM or the OM method were compared with the simulation values. RESULTS: Differences in the gross tumor volume D(98%) value from the simulation sets ranged from -0.8 to -5.9% on average. In addition, the variations were larger with OM compared with BM for two IMPT plans. Meanwhile, differences in the D(98%) value in the region isotropically enlarged by 5 mm from the gross tumor volume were significantly improved with OM on two IMPT plans and the VMAT plan. Among the organs at risk, the dose-volume indices were significantly improved with OM only in the duodenum on all plans. CONCLUSION: Organ-matching may be a better setup correction technique than BM for both photon therapy and IMPT plans. However, in some beam arrangements of IMPT, the dose distribution may be somewhat worse using OM, due to interfractional anatomical variation. Therefore, it is important to choose beam angles that are less likely to be influenced by changes in the gastrointestinal gas volume, especially in IMPT plans.