(Tuesday, 7/16/2019) 4:30 PM - 5:30 PM
Room: Exhibit Hall | Forum 2
Purpose: To evaluate the dosimetric impact of magnetic resonance imaging (MRI) distortion on photon and proton treatment plans.
Methods: A general model was used to simulate the 3D geometric distortion map of the MRI system based on vendor-provided sample distortion measurements for the MRI component of the Elekta Unity system. Distortions were applied to patient computed tomography (CT) images to synthesize the distorted CT for simulating MR-based treatment planning. New treatment plans were generated for the distorted CT using optimization objectives for original patient plans, and then applied to undistorted CT to recalculate the doses. Analytical anisotropic algorithm and pencil beam algorithm were used for photon and proton dose calculation, respectively. The dose grid resolution was 2.5 mm. Ten patients (5 photon IMRT and 5 proton) of different tumor sites (brain, pelvis, prostate, and abdomen) were selected from our patient database.
Results: For target coverage, MRI geometric distortion resulted in a difference in PTV or CTV Dmin up to 1% for photon plans and 3.9% for proton plans. For mean doses to organs at risk (OAR), the maximum deviation was 4.1% in the lens for photon plans and 3.1% in the optic nerve for proton plans. For maximum dose, the maximum deviation was 3.4% in maxilla for photon plans and 16.9% in the spinal cord (ΔD=335 cGy) for proton plans. For OARs of very small volume (< 1 cc) such as lens and cochlea, MRI distortion could change the volume up to 50%. For optic nerve and spinal cord, MRI distortion could change the volume by 25% and 6.82%, respectively.
Conclusion: MRI geometric distortion could lead to significant changes in contoured volumes as well as OAR and target doses for both photon and proton plans. It is essential to ensure that system distortions are frequently monitored and managed.
Keywords
MRI, Protons, Geometric Distortion
Taxonomy
IM/TH- MRI in Radiation Therapy: MRI for treatment planning
Contact Email
