Spatial Distortion in MRI-Guided Stereotactic Procedures: Evaluation in 1.5-, 3- and 7-Tesla MRI Scanners

Background: Magnetic resonance imaging (MRI) is replacing computed tomography (CT) as the main imaging modality for stereotactic transformations. MRI is prone to spatial distortion artifacts, which can lead to inaccuracy in stereotactic procedures. Objective: Modern MRI systems provide distortion correction algorithms that may ameliorate this problem. This study investigates the different options of distortion correction using standard 1.5-, 3- and 7-tesla MRI scanners. Methods: A phantom was mounted on a stereotactic frame. One CT scan and three MRI scans were performed. At all three field strengths, two 3-dimensional sequences, volumetric interpolated breath-hold examination (VIBE) and magnetization-prepared rapid acquisition with gradient echo, were acquired, and automatic distortion correction was performed. Global stereotactic transformation of all 13 datasets was performed and two stereotactic planning workflows (MRI only vs. CT/MR image fusion) were subsequently analysed. Results: Distortion correction on the 1.5- and 3-tesla scanners caused a considerable reduction in positional error. The effect was more pronounced when using the VIBE sequences. By using co-registration (CT/MR image fusion), even a lower positional error could be obtained. In ultra-high-field (7 T) MR imaging, distortion correction introduced even higher errors. However, the accuracy of non-corrected 7-tesla sequences was comparable to CT/MR image fusion 3-tesla imaging. Conclusion: MRI distortion correction algorithms can reduce positional errors by up to 60%. For stereotactic applications of utmost precision, we recommend a co-registration to an additional CT dataset.