(Tuesday, 7/31/2018) 1:45 PM - 3:45 PM
Room: Room 207
Purpose: To perform a dosimetric validation of synthetic computed tomography (sCT) images generated using whole-body magnetic resonance imaging (MRI) scans of children with abdominal tumors. Dose calculation accuracy was evaluated for volumetric modulated arc therapy (VMAT) and pencil-beam scanning (PBS).
Materials &
Methods: The proposed method relies on an automatic-atlas-based segmentation (bone, lungs and soft-tissue) followed by an MRI intensity conversion to Hounsfield units (HUs). Rigid registered CT and T2-weighted MRI scans in treatment position of 30 patients were used as input for the model. On the sCTs, the air cavities were assigned automatically as muscle HUs. VMAT plans, using a full-arc, and PBS plans, using 2/3 posterior irradiation fields, were robust optimized on the internal target volume (ITV) against a patient set-up uncertainty (5mm). To evaluate both sCT and plan robustness against patient anatomy differences such as the presence of random air cavities, ITV coverage differences between CT-sCT doses were calculated using the original sCTs and sCTs including the air cavities extracted from the CT (sCTair).
Results: Despite both CT and MRI were acquired on the same day (≤45 minutes), changes in the presence of air cavities between images were still detected. Mean ITV V95%/D98% differences between CT-sCT doses were 0.5%/0.8% (maximum 5.0%/3.6%) and 0.0%/0.3% (maximum 0.1%/0.7%) for VMAT and PBS, respectively. For the VMAT plans, 3 patients failed the ITV prescription (V95%>99%). Mean ITV V95%/D98% differences between CT-sCTair were 0.9%/-0.4% (maximum 5.0%/1.5%) and 0.0%/0.2% (maximum 0.1%/-1%) for VMAT and PBS, respectively.
Conclusions: Results show that accurate MRI-based photon and proton dose calculations are feasible when using whole-body sCTs of children. However, for an MRI-only photon workflow special attention might be needed during treatment due to the presence of random air cavities. For an MRI-only proton workflow, uncertainties due to random air pockets might be avoided by using posterior irradiation fields.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
