(OA06) Interfraction Variations in Stomach Volume Increase Dose to Stomach and Duodenum in Pancreas Stereotactic Body Radiotherapy and Is Mitigated by Daily Adaptive Magnetic Resonance Image Guided Radiation Therapy

Abstract

Magnetic Resonance Image Guided Radiation Therapy (MR-IGRT) permits clear daily visualization of the target and organs at risk (OARs) for daily dose adaptation. Using MR-IGRT with Cobalt-60 for locally advanced pancreatic cancer (LAPC) stereotactic body radiation therapy (SBRT), we noticed varied stomach volumes despite precautions not to eat prior to treatment. We hypothesized that larger stomach volumes would be associated with higher doses to OARs and that online adaptive re-planning would reduce OAR doses while maintaining prescription dose coverage. Ten LAPC patients received five fraction SBRT totaling 33-40Gy as limited by OAR constraints. Patients were instructed to avoid food and drink only sips of water for at least 3 hours prior to treatment. Treatment plans provided 90% coverage of the planning target volume (PTV) at prescription isodose (PID). While patients could have received adaptive or non-adaptive radiotherapy with MR-IGRT, all fifty fractions underwent simulated adaptation. First the original (as planned) doses were compared to the dose delivered only after 3D couch shifts (non-adaptive). An adaptive plan was created for each fraction mimicking the process for daily adaptive radiotherapy. Specifically, contours were adjusted to the anatomy visualized on the setup MRI for each fraction, and a “warm start” optimization run was performed to find a new optimizer solution with the day’s updated contours starting with the original optimizer objectives and control points. Adaptive plans were normalized to 90% PTV coverage at PID. We used linear regression and Bayesian approach to evaluate the association between changes in stomach volume and changes in radiation dose to visceral OARs for both non-adaptive and adaptive planning techniques. Only 8 non-adaptive MR-IGRT fractions maintained at least 90% PTV coverage compared to all 50 adaptive MR-IGRT fractions, and 23 non-adaptive fractions met all OAR objectives compared to 29 adaptive fractions. The original stomach volume (OSV) mean was 285.98cc, with a range from 125.6cc-855.8cc. The mean interfraction change from the OSV was +146.8cc, with a range from -403.76cc to +1271.13cc and a standard deviation of 245.57cc. The mean percentage change from the OSV was 53.2%, with a range from -47.18% to 334.32% and a standard deviation of 68.82%. Eight interfraction changes in stomach volume were greater than 500cc. For non-adaptive plans, Bayesian approach revealed a significant association between percentage increases in stomach volume and increased stomach V33 (p=.0041), stomach maximum dose (p=.0095) and duodenum V33 (p=.0208). Adaptive MR-IGRT showed no association between percentage increases in stomach volume and increased stomach V33 (p=.2763), stomach maximum dose (p=.3174) or duodenum V33 (p=.7736). This study demonstrates that for SBRT of LAPC, interfraction variations in stomach volume occur despite standard precautions and are associated with increased dose to the stomach and duodenum in non-adaptive plans. Adaptive planning limits dose to the stomach and duodenum and may permit safe SBRT dose escalation.