Investigating the Geometric Distortion in 3 Tesla MRI Images Based on the Phantom Studies

Abstract

Purpose: Geometric errors in images called image distortion are one of the main problems in magnetic resonance imaging, including 3D imaging, measuring blood flow velocity, functional imaging and treatment planning in radiation therapy. The geometric distortion in MRI images is due to the non-uniformity of the magnetic field and nonlinearity of gradients. In this study, the accuracy and the repeatability of the images were evaluated respectively by phantom measuring and repeating the measurements in the phantom and then we correct these geometric errors.
Methods: The magnetic resonance imaging of the phantom was performed on the 3 Tesla Siemens Prisma Model to measure the geometric distortion using the network pattern. The spin echo protocol was repeated three times with T₁, T₂ and PD weightings to measure the repeatability of image distortion. Image distortion was evaluated by measuring the distance between the edges using a MATLAB program. Furthermore, non-uniformity of the magnetic field and the nonlinearity of the gradients were examined using appropriate phantoms. 
Result: The average error obtained in the 25 cm field of view was 1 pixel in both directions x and y (each pixel was 1.024 mm). Given the images by phantom, the device gradient was linear. Furthermore, considering the B₁ and B₀ fields’ measurements, the B₀ of the device was 0.3125 ppm over a 24 cm DSV (diameter of spherical volume).
Conclusion: Since the brain coil displacement was 1 pixel, the device could be used in 3DMRI, velocity MRI, FMRI and RTTP.