Room: Karl Dean Ballroom B1
Purpose: To quantify effects of ferromagnetic material of MR-guided radiotherapy (MRgRT) vault and ferromagnetic gantry components for a MR-guided cobalt versus linac design.
Methods: A clinical MRgRT system was upgraded from tri-cobalt-60 to a 6 MV inline linac within the same vault. Retrofitted MRgRT vault contains ferromagnetic rebar in concrete floor radially-surrounding gantry pit. Six gantry-mounted ferromagnetic steal shields (450 lbs) were installed 87 cm from isocenter on MR-linac to reduce Lorentz force on charge-carrying components. During cobalt-to-linac upgrade, unknown ferromagnetic vault sources (can lights and unistruts) were identified and removed within 5 Gauss, 10â€™ of isocenter. Metrics for magnetic field homogeneity and geometric distortion were quantified for MR-linac and MR-cobalt of the same magnet and vault acquired.
Results: Inhomogeneity (n=12 gantry angles, 45cm DSV) was reduced for MR-linac (max 16.27 ppm, mean 14.85Â±0.80 ppm) compared to MR-cobalt (max 19.41 ppm, mean 18.38Â±0.76 ppm, p<0.0001) for field-camera results. Using the spectral-peak method and post-gradient correction offsets, homogeneity within 24cm DSV was 2.41Â±0.78ppm(linac) and 1.01Â±0.09ppm(cobalt). Mean geometric distortion remained similar with 0.81Â±0.28mm(cobalt) and 0.80Â± 0.32mm(lianc) for 696 points 35cm DSV, p=0.175. Maximum MR-linac distortion was 1.9 mm in coronal plane, which was statistically different (p<0.0001, mean coronal distortion 1.04Â±0.25mm linac vs. 0.85Â±0.33mm cobalt). Y-sine tesseral harmonics of MR-linac were greatest in non-linearity with gantry rotation (range Â±2ppm) compared to other directions (range Â±0.5ppm). Both MR-linac and MR-cobalt remained within specification for magnetic field homogeneity (<25 ppm, 45cm DSV) and spatial distortion (<2mm, 35cm DSV).
Conclusion: Static ferrous-vault removal within 5 Gauss improved maximum magnetic inhomogeneity by 3 ppm, p<0.0001 for MR-linac compared to MR-cobalt, even with addition of ferromagnetic shields. A marginal increase in coronal spatial distortion was observed (1.04Â± 0.25mm linac vs. 0.85Â±0.33mm cobalt), likely due to variability of non-linear Y-harmonics during rotational interplay of MR-linac ferrous-shields and floor-rebar.
Funding Support, Disclosures, and Conflict of Interest: K. Mittauer reports personal fees from ViewRay Inc.; S. McMain is employed by ViewRay Inc.; J. Bayouth reports membership of Advisory Board of ViewRay, Inc.