Room: Exhibit Hall | Forum 2
Purpose: MRI-Linacs allow online plan adaption based on daily MR images. Electron densities need to be assigned to these images for dose calculation. Two possible approaches are bulk assignment, in which the electron densities are assigned manually, or deform the simulation CT to the daily MRI. The goal of this work is to compare the dose distribution generated with these two methods.
Methods: Ten pancreas treatment plans originally calculated with CT-based electron densities were recalculated using a bulk electron density assignment. The air electron density was assigned to the gas inside the patient abdomen, bone electron density was assigned to the vertebrae and water was assigned to the remaining tissue inside the skin contour. The following DVH parameters were compared: GTV mean, PTV mean, PTV_V95 and D0.03cc for some organs at risk (OARs). The full DVH and isodose distributions were also reviewed.
Results: The DVHs appear nearly identical for the OARs. For the tail regions of the OAR DVHs, the differences are clinically acceptable (average <1.3%, maximum <5%). For the PTV_V95 the maximum difference was <1%. Larger discrepancies were seen for the GTV mean (average <0.9%, maximum <2.6%) and PTV mean (average <0.8%, maximum <2.5%) due to changes in hot spot sizes and location.
Conclusion: For abdominal plans (dominant presence of soft tissue with no lungs or pelvic bones), the bulk density can be used without sacrificing the calculation accuracy of the PTV coverage or OAR sparing. Bulk density removes the need for a CT simulation and avoids errors related to electron densities deformation accuracy. As the CT deformation fails to adjust the air pockets from one day to the next, water and air overrides are needed to correct it. Air overrides are also needed daily for bulk assignment, but it leads to a faster adaptive workflow.