KEY POINTS
- This retrospective study included 74 patients treated with magnetic resonance-guided online adaptive stereotactic body radiotherapy to the pancreas between 2018 and 2024; 70 had primary pancreatic cancer.
- All patients received standardized instructions to avoid food and fluids for 2 hours before simulation and each fraction; 73 patients received 50 Gy in 5 fractions, and one received 40 Gy in 4 fractions.
- Only 42% had reproducible stomach volumes, while 43% had intermediate variability and 15% had high variability; variability was significantly greater in the latter groups (p<0.00001).
- Median stomach volume at simulation was lower in patients with reproducible anatomy than in the variable and unsuccessful groups (183 vs 255 and 238 cm³; p<0.001). A 229 cm³ threshold identified 67% of reproducible cases and 87% of non-reproducible cases.
- Total stomach volume correlated poorly with stomach volume within 3 cm of the planning target volume (r=0.28; R²=0.08). Dosimetric effects and clinical toxicity were not evaluated.
CLINICAL TAKEAWAY
Uniform two-hour fasting should not be assumed to provide reproducible gastric anatomy during pancreatic stereotactic body radiotherapy. Simulation stomach volume may help identify patients requiring closer image guidance or individualized preparation, but the retrospective design, unverified compliance, and absence of dosimetric or clinical endpoints make these findings hypothesis-generating.