Introduction

Various sequences have been proposed for whole heart coronary magnetic resonance angiography (WHC-MRA)^1-4. In conventional WHC-MRA, it is typically performed during free-breathing, and respiratory motion is prospectively compensated by using a 1D-right-diaphragmatic-navigator (NAV)⁵. However, an important limitation of the NAV method is its relatively long acquisition time⁶. Single-breath-hold technique using 3D Turbo-field-echo-planar-imaging (TFEPI) could significantly reduce the acquisition time of WHC-MRA while providing image quality similar to that of conventional free-breathing gradient echo sequence⁷. TFEPI sequence is a hybrid technique that combines TFE and EPI readouts⁷. It enables highly accelerated imaging thanks to EPI readout, resulting in shorter scan time by allowing more echoes to be acquired during one heartbeat. Furthermore, EPI with Compressed SENSE (C-SENSE) has been introduced^8,9. This approach can be extended to 3D TFEPI without modifying its sampling pattern and ffurther accelerated scan time without degrading the image quality by denoising as previously introduced¹⁰. However, TFEPI with C-SENSE is still challenging to obtain high spatial resolution like conventional NAV-WHC-MRA with clinically acceptable signal-to-ratio (SNR). Recently, integrating artificial intelligence (AI) into the Compressed SENSE reconstruction (CS-AI), based on Adaptive-CS-Net^11,12 (fig.1), has been introduced and CS-AI dramatically reduced noise artifacts and significantly improved image quality^13-15. In addition, it has been shown recently that deep learning-based denoising is an effective tool for image quality improvement in WHC-MRA¹⁶. We hypothesize that the combination of TFEPI and CS-AI can provide further high spatial resolution of single-breath-hold WHC-MRA while maintaining clinically acceptable SNR and scan time. In this study, we investigated the feasibility of single-breath-hold TFEPI with CS-AI for WHC-MRA by comparing with single-breath-hold TFEPI with C-SENSE and SENSE.

Methods

A total of six volunteers (5 males and 1 female; age range: 28~45) were examined on a 3.0T-MRI (Ingenia Elition X, Philips Healthcare). The study was approved by the local IRB, and written informed consent was obtained from all subjects. We compared image quality among WHC-MRA using TFEPI with CS-AI, C-SENSE or SENSE. Three sequences were acquired with cardiac triggering and single-breath-hold for respiratory compensation. Imaging-parameters; TFEPI: FOV=300×280mm2, voxel-size=1.7×1.7×1.7mm3, TR/TE/FA=7.7/3.5/20, TFE-factor=15-22, EPI-factor=7, shot-duration=100~130ms, reduction factor=4.1×2.0, scan time:17~23sec. Curved Planer Reconstructions (CPRs) were performed using Ziostation2 (Ziosoft Co, Tokyo) and image quality was evaluated by visual score at the 10-points (RCA: #1/2/3/4 LAD: #5/6/7/8 CX: #11/13) based on American-Heart-Association classification. For overall image quality, sharpness and noise and artifacts, we evaluated them as 4-point grades (grade “4” was excellent, “1” was severe) by two blinded readers. Visual evaluation was assessed by Steel-Dwass test. For quantitative comparison, contrast-to-noise ratio (CNR) were measured. The CNR was estimated for blood and epicardial fat (CNRblood-epicardial fat) and blood and myocardium (CNRblood-myocardium). To allow quantitative CNR measurements, we used a noise-measurement-method proposed by Zwanenburg et al¹⁷. Each sequence was repeated with exactly the same receiver gain, but without any RF and gradient-pulses. The reconstructed images showed only noise, including the noise added due to the C-SENSE/SENSE-reconstruction. The standard-deviation of a region of interest of the corresponding area in the noise image was used as metric for the noise. The CNRblood-epicardial fat and CNRblood-myocardium were calculated by the following equations: CNRA-B = [SI(A) - SI(B)] / 0.5 [SDnoise(A) + SDnoise(B)] Where SI are the mean average signal intensity of the blood, epicardial fat and myocardium respectively, and the corresponding SDnoise is the standard-deviation at the same location on the noise images. The CNR were assessed by one-way repeated measures analysis of variance (ANOVA) and the post-hoc Tukey-Kramer test.

Results and Discussion

Figure 2 shows the results of visual evaluation. Regarding overall image quality, sharpness and noise and artifacts, TFEPI with CS-AI was significantly higher than TFEPI with C-SNESE and SENSE (p<0.0001). Figure 3 shows CNR comparison among three sequences. For the CNRblood-epicardial fat and CNRblood-myocardium, TFEPI with CS-AI was significantly higher than TFEPI with C-SNESE and SENSE. Figure 4 shows the representative images and actual scan time using TFEPI with CS-AI/C-SENSE/SENSE. Figure 5 shows the representative images and actual scan time of TFEPI with CS-AI and conventional NAV WHC-MRA. TFEPI with CS-AI had high visual score and better image contrast as a result of high SNR and CNR due to the effect of AI-based denoising. Furthermore, the image quality of single-breath-hold TFEPI with CS-AI was comparable to that of conventional TFE with NAV.

Conclusion

3D TFEPI with CS-AI has a great potential to obtain 3D high resolution isotropic single-breath-hold WHC-MRA with clinical acceptable image quality with short breath-holding time, around 20 seconds, and it could reduce the burden on the patients.

Acknowledgements

No acknowledgements found.

References

(1) Kazuo Kodaira, et al. Whole heart coronary MRA using image-based 2D navigator (iNav) and conventional Nav system: comparison of image quality and scan time. Proceedings of the Annual Meeting & Exhibition of ISMRM (2021). (2) Oliver M. Weber, PhD et al. Free-Breathing, Three-Dimensional Coronary Artery Magnetic Resonance Angiography: Comparison of Sequences. Journal of Magenetic Resonance Imaging. 2004; 20: 395-402. (3) Teresa Correia et al. Technical note: Accelerated nonrigid motion-compensated isotropic 3D coronary MR angiography. Med Phys. 2018 Jan; 45(1): 214-222. (4) Yuji Iyama et al. Single-breath-hold whole-heart coronary MRA in healthy volunteers at 3.0-T MRI. Springerplus. 2014; 3: 667. (5) Markus Henningsson et al. Prospaective Respiratory Motion Correction for Coronary MR Angiography Using a 2D Image Navigator. Magn Reson Med. 2013; 69: 486-494. (6) Mehdi H. Monghari et al. Three-dimensional Heart Locator and Compressed Sensing for Whole-heart Magnetic Resonance Angiography. Magn Reson Med. 2016 May; 75(5): 2086–2093. (7) Yuji Iyama et al. Single-Breath-Hold Whole-heart Unenhanced Coronary MRA Using Multi-shot Gradient Echo EPI at 3T: Comparison with Free-breathing Turbo-field-echo Coronary MRA on Healthy Volunteers. Magn Reson Med Sci. 2018; 17: 161-167. (8) Kaga T, et al. Diffusion-weighted imaging of the abdomen using echo planar imaging with compressed SENSE: Feasibility, image quality, and ADC value evaluation. Eur J Radiol. 2021 Sep;142:109889. doi: 10.1016/j.ejrad.2021.109889. (9) Yoneyama M, et al. Noise Reduction in Prostate Single-Shot DW-EPI utilizing Compressed SENSE Framework. Proc Intl Soc Mag Reson Med. 2019:1634. (10) Kazuo Kodaira, et al. Single-breath-hold whole heart coronary MRA using 3D turbo-field-echo-planar-imaging (TFEPI) with Compressed SENSE framework. Proceedings of the Joint Annual Meeting ISMRM-ESMRMB & ISMRT 31st Annual Meeting (2022). (11) Pezzotti N, et al. An Adaptive Intelligence Algorithm for Undersampled Knee MRI Reconstruction. IEEE Access. 2020;8:204825-204838. (12) Pezzotti N, et al. Adaptive-CS-Net: FastMRI with Adaptive Intelligence. arxiv. 2019;(NeurIPS). (13) Yoneyama M, et al. Rapid submillimeter high-resolution prostate T2 mapping with a deep learning constrained Compressed SENSE reconstruction. Proc Intl Soc Mag Reson Med. 2021:4117. (14) Yoneyama M, et al. SNR boost in whole-body DWIBS utilizing deep learning constrained Compressed SENSE reconstruction. Proceedings of the Joint Annual Meeting ISMRM-ESMRMB & ISMRT 31st Annual Meeting (2022). (15) Yoneyama M, et al. SNR enhancement in rapid high b-value prostate single-shot DW-EPI utilizing deep learning constrained Compressed SENSE reconstruction. Proceedings of the Joint Annual Meeting ISMRM-ESMRMB & ISMRT 31st Annual Meeting (2022). (16) Xi Wu, et al. Feasibility of accelerated non-contrast-enhanced whole-heart bSSFP coronary MR angiography by deep learning-constrained compressed sensing. Euro Radiol 2023, 33:8180-8190. (17) Zwanenburg et al. MR Angiography of the cerebral perforating arteries with magnetization prepared anatomical reference at 7T: comparison with time-of-flight. J Magn Reson Imaging. 2008; 28: 1519–1526.