Purpose
With the emergence of MR-guided radiotherapy, intra-fractional motion tracking has the potential to decrease geometric uncertainty and reduce planning margins, but several considerations still require examination. This study evaluated dosimetric uncertainty caused by gastrointestinal air cavities arising from the electron return effect and lack of electron density information for pancreatic cancer treatments.
Methods
Ten patients treated with SBRT on a 0.35T MR-guided Co-60 teletherapy unit were retrospectively assessed. Initial plans were calculated using electron densities assigned through deformable registration between CT and MR simulations. Air cavities inside the gastrointestinal tract within 5cm superior/inferior to the target were contoured on seven images: CT simulation, MR simulation, and each fraction’s pre-treatment MRI. Three patients showing large variations underwent further analysis in which cavity electron density was bulky assigned for plan re-calculation to evaluate dose perturbations.
Results
Considerable cavity volume changes were observed between CT and MR simulations (132.6+/-234.4cc), despite being acquired consecutively on the same day. The average air cavity volume for all images was 192+/-173cc. DICE similarity indices ranged from 0-0.8 (0.22+/-0.18), indicating large spatial variation of the air cavities. Spatial dose perturbations were observed at air-tissue interfaces, however dosimetric impact was found to be minimal in most instances. Significant dosimetric discrepancies were observed in the coverage and hot spots of the target for one patient with increases up to 7% in the D98 and 15.5% in the D2 of the GTV.
Conclusion
Results show there are measureable perturbations of the dose distribution due to gastrointestinal air cavity variations for pancreatic SBRT patients. The largest changes were observed in target coverage and dose homogeneity. If electron density is assigned through deformable registration between CT and MR simulations, results suggest that DIR alone will lead to dosimetric errors when significant differences exist between the air cavities of the two image volumes.