Gizem Temiz^{1,2}, Fernando Pérez-García^{1,2}, Catherine Jenny^{3}, Stéphane Lehéricy^{1,4}, Marguerite Cuttat ^{3}, Didier Dormont^{4}, Damien Galanaud^{1,4}, Chales Valery^{5}, Carine Karachi^{1,5}, Romain Valabregue ^{2}, Sara Fernandez-Vidal^{1,2}, Nadya Pyatigorskaya^{4}, and Eric Bardinet^{1,2}

Accurate MRI-based targeting is a critical issue for stereotactic surgery. Therefore, geometric distortions need to be evaluated for any pre-operative MR imaging protocol. In this study, we investigated MRI protocols used in Deep Brain Stimulation and Gamma Knife radiosurgery, and focused on the influence of 5 factors on the geometric distortions, at 1.5T and 3T, for 3D T1-weighted and 3D FLAIR images. We found that in order to minimize geometric distortions in stereotactic imaging operator training, careful centering in the MR scanner and systematic activation of constructor’s distortion correction filter are essentials.

Stereotactic surgery is a well-established treatment approach for both Deep Brain Stimulation (DBS) and Gamma Knife (GK) radiosurgery. For these procedures, accurate MRI-based targeting is a critical issue with a need of producing distortion-free images. Thus, evaluation of geometric distortions in clinical MRI systems is crucial. Geometric distortions can be induced by the magnet itself, by a particular sequence or by the patient. These distortions depend on a number of factors, many of which have been studied in the literature [1]. Nevertheless, for stereotactic imaging protocols, some questions remained to be explored. The aim of our study was to analyze a group of sources of geometric distortion in order to propose guidelines for good MRI practices for pre-operative image acquisitions.

We focused on 3D images used for targeting at 1.5T and 3T (FSPGR on a 1.5T GE Optima ; MP2RAGE, FLAIR, and SPACE on a 3T Siemens SKYRA ), using a geometric phantom (GRID 3D, MODUS QA: rectilinear grid of 2002 points). For FSPGR, MP2RAGE, FLAIR and SPACE acquisitions, resolutions were 0.94 x 0.94 x 1.2 mm^{3}, 1 x 1 x 1 mm^{3}, 0.48 x 0.48 x 1 mm^{3}, 0.94 x 0.94 x 1 mm^{3 }respectively. We assessed the influence of five factors: 1) operator-induced distortions by studying inter- and intra-operator variations, 2) position of the magnet isocenter (laser position) with respect to the phantom, 3) distortion-correction filter available in the MR systems (and coupled effect of the laser positioning and correction filter), 4) frame-induced distortions, 5) susceptibility-based geometric distortions by acquiring pairs of images (inverting the phase-encoding direction).

An in-house Python-based module was implemented in 3D Slicer [2] for the automatic assessment of geometric distortions. All phantom images were preprocessed by cropping, denoising [3] and resampling to 0.5-mm3 voxels. Detection of the intersections was performed by normalized cross-correlation of the image with a 3D-cross kernel. Residual tilt was removed. The distortion field was computed as the differences between theoretical and measured points.

This study was partially supported by GE Healthcare.

The authors acknowledge the financial support provided by the Fondation pour la Recherche Médicale (FRM) (Project:DIC20161236441) and by the "Investissements d’avenir" ANR-10-IAIHU-06.

[1] Weygand et al. , International Journal of Radiation Oncology, 95.4 1304-1316, 2016

[2] Fedorov, Andriy et al. Magnetic resonance imaging, 30.9, 1323–1341, 2012

[3] Mirebeau, Jean-Marie et al. arXiv: 1503.00992., 2015