Measuring geometric accuracy in magnetic resonance imaging with 3D-printed phantom and nonrigid image registration

Abstract

Objective We aimed to develop a vendor-neutral and interaction-free quality assurance protocol for measuring geometric accuracy of head and brain magnetic resonance (MR) images. We investigated the usability of nonrigid image registration in the analysis and looked for the optimal registration parameters. Materials and methods We constructed a 3D-printed phantom and imaged it with 12 MR scanners using clinical sequences. We registered a geometric-ground-truth computed tomography (CT) acquisition to the MR images using an open-source nonrigid-registration-toolbox with varying parameters. We applied the transforms to a set of control points in the CT image and compared their locations to the corresponding visually verified reference points in the MR images. Results With optimized registration parameters, the mean difference (and standard deviation) of control point locations when compared to the reference method was (0.17 ± 0.02) mm for the 12 studied scanners. The maximum displacements varied from 0.50 to 1.35 mm or 0.89 to 2.30 mm, with vendors' distortion correction on or off, respectively. Discussion Using nonrigid CT-MR registration can provide a robust and relatively test-object-agnostic method for estimating the intra-and inter-scanner variations of the geometric distortions.

Augmenting Performance: A Systematic Review of Optical See-Through Head-Mounted Displays in Surgery

“…The second most frequently used data type was MRI, either as a standalone data type [ 62 , 72 , 78 ] or in combination with CT [ 49 , 52 , 53 ]. Magnetic resonance imaging has better sensitivity than CT when imaging soft tissue structures; however, it is less geometrically precise [ 79 ] and has less signal intensity when imaging bony structures. The registration of MRI and CT data is advantageous in situations where geometrically precise bone imaging and soft tissue information are used in tandem.…”

Head-Mounted Display-Based Augmented Reality for Image-Guided Media Delivery to the Heart: A Preliminary Investigation of Perceptual Accuracy

“…Comparing the best-case assessment of TRE, with 0.3-0.9 mm of error, contributed, based on marker tracking alone, to the TRE of 1.1 mm RMS measured in our repeatability study (T R ), we can infer that the error, due to other sources such as display calibration and vergence-accommodation, is in the order of 0.2-0.8 mm. With registration of imaging data from preoperative imaging to intraoperative surgical space, we expect additional error on the order of 0.2 mm [43] due to the inherent resolution limitations of MRI. Therefore, the theoretical best-case TRE of our combined guidance system is 1.35 mm RMS.…”

Characterizing geometric distortions of 3D sequences in clinical head MRI

“…The grid was placed in a drum (17 cm in height and 20 cm in diameter) that was filled with mineral oil. Further details of the phantom are presented in [ 34 ].…”

“…In addition, the phantom was scanned with CT (SOMATOM Definition Edge, Siemens Healthineers, Erlangen, Germany) with 120 kV tube voltage, 191 mA tube current, and 0.5 mm slice thickness, and reconstructed with 0.4 mm × 0.4 mm pixel size, 512 × 512 image matrix, and general-purpose soft-tissue J45s-kernel. This CT scan was the same as used in [ 34 ], but we performed a QA CT scan with similar settings before the MRI acquisitions to affirm that the phantom had remained unchanged.…”

“…In an earlier work, our group has presented MRI scanner QA methods based on diagnostic 3D FLAIR (FLuid-Attenuated Inversion Recovery) acquisitions [ 33 ], yet the geometric fidelity of the images was not assessed. In addition, we have previously studied the geometric accuracy of MPRAGE (Magnetization Prepared—RApid Gradient Echo) acquisitions on 12 MRI scanners with a 3D-printed phantom and a nonrigid-registration-based analysis [ 34 ], but no test-subject acquisitions were utilized. The geometric distortion in MRI have been widely studied, however, the geometric distortions are rarely evaluated in anatomical images for QA purposes despite their potential additional distortion sources compared with phantom studies.…”