Dosimetric Effect of Respiratory Phase in MR-Guided SBRT for Locally Advanced Pancreatic Cancer

Abstract

<h3>Purpose/Objective(s)</h3> Dose escalation of stereotactic body radiation therapy (SBRT) for locally advanced pancreatic cancer (LAPC) has traditionally been limited by organs-at-risk (OARs) tolerance; however, MR-guided SBRT (MRgSBRT) is currently being evaluated for dose escalation in multi-institutional trials. This novel treatment method incorporates soft tissue gating and on-table adaptive planning to achieve optimal dosimetry. Moreover, respiratory gating has previously been employed to minimize dose to OARs for tumors that sit close to the diaphragm. It can be inferred that the superior position of the diaphragm in end exhale breath hold (EE) allows for greater abdominal organ separation than in deep inspiration breath hold (DI). Therefore, we hypothesized that patients (pts) treated during EE for LAPC would demonstrate favorable treatment planning metrics including increased target coverage and decreased monitor unit delivery compared to pts treated in DI. <h3>Materials/Methods</h3> Five pts with LAPC previously treated with MRgSBRT were included in this analysis. For each pt, at the time of simulation, a CT scan was performed for heterogeneity corrections followed by two planning MRI scans: one in end exhale (MR_EE) and one in deep inspiration (MR_DI). All pts were ultimately treated in DI. For this study, both MR_EE and MR_DI scans were anonymized for contouring of targets and OARs. CTV was defined as the gross tumor (GTV) plus elective nodal coverage of the superior mesenteric (SM) and celiac axis (CA) nodes. PTV was defined as a 3 mm expansion on the CTV. The prescription dose was 50 Gy in 5 fractions every other day. Each plan was normalized by taking one of the luminal GI OARs (stomach, duodenum, small bowel, or large bowel) to the maximum tolerance of 33Gy to 0.5cc. Treatment plans were created using 18-21 beams per plan with a constant number of available beams per patient. A two-tailed t-test of equal variance was used to compare CTV coverage, PTVopt coverage, and monitor units per fraction (MU/fx) between the plans created on the MR_EE and MR_DI scans. <b>Results:</b> <h3>Conclusion</h3> Dose-escalated MRgSBRT for LAPC results in adequate target coverage while also respecting OAR tolerances. However, we were not able to demonstrate any statistically significant difference between the two breath hold techniques (EE vs DI) for CTV coverage, PTVopt coverage or monitor units per fraction for our cohort. Therefore, we recommend treating using the most comfortable and reproducible breath hold per pt. Further work is needed to characterize other factors such as patient compliance and gating efficiency.