Analysis of different image-registration algorithms for Fourier decomposition MRI in functional lung imaging

Abstract

Background Motion correction is mandatory for the functional Fourier decomposition magnetic resonance imaging (FD-MRI) of the lungs. Therefore, it is important to evaluate the quality of various image-registration algorithms for pulmonary FD-MRI and to determine their impact on FD-MRI outcome. Purpose To evaluate different image-registration algorithms for FD-MRI in functional lung imaging. Material and Methods Fifteen healthy volunteers were examined in a 1.5-T whole-body MR scanner (Magnetom Avanto, Siemens AG) with a non-contrast enhanced 2D TrueFISP pulse sequence in coronal view and free-breathing (acquisition time 45 s, 250 images). Three image-registration algorithms were used to compensate the spatial variation of the lungs (fMRLung 3.0, ANTs, and Elastix). Quality control for image registration was performed by edge detection (ED), quotient image criterion (QI), and dice similarity coefficient (DSC). Ventilation, perfusion, and a ventilation/perfusion quotient (V/Q) were calculated using the three registered datasets. Results Average computing times for the three image-registration algorithms were 1.0 ± 1.6 min, 38.0 ± 13.5 min, and 354 ± 78 min for fMRLung, ANTs, and Elastix, respectively. No significant difference in the quality of motion correction provided by different image-registration algorithms occurred. Significant differences were observed between fMRLung- and Elastix-based perfusion values ​​of the left lung as well as fMRLung- and ANTs-based V/Q quotient of the right and the entire lung ( P < 0.05). Other ventilation and perfusion values were not significantly different. Conclusion The mandatory motion correction for functional FD-MRI of the lung can be achieved through different image-registration algorithms with consistent quality. However, a significantly difference in computing time between the image-registration algorithms still requires an optimization.

Phase‐cycled balanced SSFP imaging for non‐contrast‐enhanced functional lung imaging

“…For this reason, the phase‐cycled FD method suffers from a lower scan efficiency compared with the conventional FD method. Please note that for successful applications of FD MRI, multiple respiratory and cardiac cycles are needed to be captured with sufficient temporal resolution, 52 and commonly, images in the order of hundreds are acquired for FD‐MRI applications 1,39 . Here, the total number of acquisitions prescribed was within the range of values previously used for conventional acquisition studies, 53 and no additional time was required to accommodate additional phase cycle increments.…”

“…In conventional FD MRI, 2D bSSFP sequence with single RF phase increment (hereafter referred to as single‐phase acquisition) constitutes the basis of the time‐resolved acquisitions to assess the signal variations arising from respiratory and cardiac cycles 5 . Conventionally, a total of $$$ K $$$ images are acquired with a frame rate of 3–4 images per second during free breathing, 14 and registered using a nonrigid registration software 39 to align lung structures, thus enabling the observation of regional parenchymal density changes. Thereafter, temporal Fourier transform is used to spectrally analyze and obtain the signal changes corresponding to respiratory and cardiac frequencies, resulting in ventilation‐weighted and perfusion‐weighted maps.…”

“…Please note that for successful applications of FD MRI, multiple respiratory and cardiac cycles are needed to be captured with sufficient temporal resolution, 52 and commonly, images in the order of hundreds are acquired for FD-MRI applications. 1,39 Here, the total number of acquisitions prescribed was within the range of values previously used for conventional acquisition studies, 53 and no additional time was required to accommodate additional phase cycle increments. Another limitation of the current implementation originates from the phase cycle increments used in this study.…”

Artificial intelligence in functional imaging of the lung

“…Likewise, different approaches have been proposed based on traditional functional optimization that shows stable quality results. 35 Deep learning registration offers an alternative with low computational cost during the inference stage once the registration model is trained. 36,37 Deep learning in MRI also has been attempted to estimate quantitative tissue parameters using quantitative susceptibility mapping (QMS) and MRI fingerprinting to achieve more standardized biomarkers.…”

Evaluation of image registration algorithms for 3D phase‐resolved functional lung ventilation magnetic resonance imaging in healthy volunteers and chronic obstructive pulmonary disease patients

“…Image registration demands up to several hours to enable pixel-wise analysis of acquired images from a single subject, even for 2D techniques. 11 The current state-of-the-art registration algorithms are implemented in Advanced Normalization Tools (ANTs) open-source software library. 12 To date, ANTs development does not permit taking advantage of a graphics processing unit (GPU), which hinders the use of new free-breathing proton MR techniques in the clinical setting.…”