Accelerated chemical shift encoded cardiovascular magnetic resonance imaging with use of a resolution enhancement network
ABSTRACT: Background: Cardiovascular magnetic resonance (CMR) chemical shift encoding (CSE) enables myocardial fat imaging. We sought to develop a deep learning network (fast chemical shift encoding [FastCSE]) to accelerate CSE. Methods: FastCSE was built on a super-resolution generative adversaria...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-01-01
|
| Series: | Journal of Cardiovascular Magnetic Resonance |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1097664724011177 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850255733223849984 |
|---|---|
| author | Manuel A. Morales Scott Johnson Patrick Pierce Reza Nezafat |
| author_facet | Manuel A. Morales Scott Johnson Patrick Pierce Reza Nezafat |
| author_sort | Manuel A. Morales |
| collection | DOAJ |
| description | ABSTRACT: Background: Cardiovascular magnetic resonance (CMR) chemical shift encoding (CSE) enables myocardial fat imaging. We sought to develop a deep learning network (fast chemical shift encoding [FastCSE]) to accelerate CSE. Methods: FastCSE was built on a super-resolution generative adversarial network extended to enhance complex-valued image sharpness. FastCSE enhances each echo image independently before water-fat separation. FastCSE was trained with retrospectively identified cines from 1519 patients (56 ± 16 years; 866 men) referred for clinical 3T CMR. In a prospective study of 16 participants (58 ± 19 years; 7 females) and 5 healthy individuals (32 ± 17 years; 5 females), dual-echo CSE images were collected with 1.5 × 1.5 mm2, 2.5 × 1.5 mm2, and 3.8 × 1.9 mm2 resolution using generalized autocalibrating partially parallel acquisition (GRAPPA). FastCSE was applied to images collected with resolution of 2.5 × 1.5 mm2 and 3.8 × 1.9 mm2 to restore sharpness. Fat images obtained from two-point Dixon reconstruction were evaluated using a quantitative blur metric and analyzed with a five-way analysis of variance. Results: FastCSE successfully reconstructed CSE images inline. FastCSE acquisition, with a resolution of 2.5 × 1.5 mm2 and 3.8 × 1.9 mm2, reduced the number of breath-holds without impacting visualization of fat by approximately 1.5-fold and 3-fold compared to GRAPPA acquisition with a resolution of 1.5 × 1.5 mm2, from 3.0 ± 0.8 breath-holds to 2.0 ± 0.2 and 1.1 ± 0.4 breath-holds, respectively. FastCSE improved image sharpness and removed ringing artifacts in GRAPPA fat images acquired with a resolution of 2.5 × 1.5 mm2 (0.32 ± 0.03 vs 0.35 ± 0.04, P < 0.001) and 3.8 × 1.9 mm2 (0.32 ± 0.03 vs 0.43 ± 0.06, P < 0.001). Blurring in FastCSE images was similar to blurring in images with 1.5 × 1.5 mm2 resolution (0.32 ± 0.03 vs 0.31 ± 0.03, P = 0.57; 0.32 ± 0.03 vs 0.31 ± 0.03, P = 0.66). Conclusion: We showed that a deep learning-accelerated CSE technique based on complex-valued resolution enhancement can reduce the number of breath-holds in CSE imaging without impacting the visualization of fat. FastCSE showed similar image sharpness compared to a standardized parallel imaging method. |
| format | Article |
| id | doaj-art-e1c6406243f542a7a70e7c4791669f70 |
| institution | OA Journals |
| issn | 1097-6647 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Journal of Cardiovascular Magnetic Resonance |
| spelling | doaj-art-e1c6406243f542a7a70e7c4791669f702025-08-20T01:56:48ZengElsevierJournal of Cardiovascular Magnetic Resonance1097-66472024-01-0126210109010.1016/j.jocmr.2024.101090Accelerated chemical shift encoded cardiovascular magnetic resonance imaging with use of a resolution enhancement networkManuel A. Morales0Scott Johnson1Patrick Pierce2Reza Nezafat3Department of Medicine (Cardiovascular Division), Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, USADepartment of Medicine (Cardiovascular Division), Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, USADepartment of Medicine (Cardiovascular Division), Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, USACorresponding author. Department of Medicine (Cardiovascular Division), Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, Massachusetts, 02215, USA.; Department of Medicine (Cardiovascular Division), Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts, USAABSTRACT: Background: Cardiovascular magnetic resonance (CMR) chemical shift encoding (CSE) enables myocardial fat imaging. We sought to develop a deep learning network (fast chemical shift encoding [FastCSE]) to accelerate CSE. Methods: FastCSE was built on a super-resolution generative adversarial network extended to enhance complex-valued image sharpness. FastCSE enhances each echo image independently before water-fat separation. FastCSE was trained with retrospectively identified cines from 1519 patients (56 ± 16 years; 866 men) referred for clinical 3T CMR. In a prospective study of 16 participants (58 ± 19 years; 7 females) and 5 healthy individuals (32 ± 17 years; 5 females), dual-echo CSE images were collected with 1.5 × 1.5 mm2, 2.5 × 1.5 mm2, and 3.8 × 1.9 mm2 resolution using generalized autocalibrating partially parallel acquisition (GRAPPA). FastCSE was applied to images collected with resolution of 2.5 × 1.5 mm2 and 3.8 × 1.9 mm2 to restore sharpness. Fat images obtained from two-point Dixon reconstruction were evaluated using a quantitative blur metric and analyzed with a five-way analysis of variance. Results: FastCSE successfully reconstructed CSE images inline. FastCSE acquisition, with a resolution of 2.5 × 1.5 mm2 and 3.8 × 1.9 mm2, reduced the number of breath-holds without impacting visualization of fat by approximately 1.5-fold and 3-fold compared to GRAPPA acquisition with a resolution of 1.5 × 1.5 mm2, from 3.0 ± 0.8 breath-holds to 2.0 ± 0.2 and 1.1 ± 0.4 breath-holds, respectively. FastCSE improved image sharpness and removed ringing artifacts in GRAPPA fat images acquired with a resolution of 2.5 × 1.5 mm2 (0.32 ± 0.03 vs 0.35 ± 0.04, P < 0.001) and 3.8 × 1.9 mm2 (0.32 ± 0.03 vs 0.43 ± 0.06, P < 0.001). Blurring in FastCSE images was similar to blurring in images with 1.5 × 1.5 mm2 resolution (0.32 ± 0.03 vs 0.31 ± 0.03, P = 0.57; 0.32 ± 0.03 vs 0.31 ± 0.03, P = 0.66). Conclusion: We showed that a deep learning-accelerated CSE technique based on complex-valued resolution enhancement can reduce the number of breath-holds in CSE imaging without impacting the visualization of fat. FastCSE showed similar image sharpness compared to a standardized parallel imaging method.http://www.sciencedirect.com/science/article/pii/S1097664724011177Deep learningChemical shift encodingSuper-resolutionDixonMyocardial fat |
| spellingShingle | Manuel A. Morales Scott Johnson Patrick Pierce Reza Nezafat Accelerated chemical shift encoded cardiovascular magnetic resonance imaging with use of a resolution enhancement network Journal of Cardiovascular Magnetic Resonance Deep learning Chemical shift encoding Super-resolution Dixon Myocardial fat |
| title | Accelerated chemical shift encoded cardiovascular magnetic resonance imaging with use of a resolution enhancement network |
| title_full | Accelerated chemical shift encoded cardiovascular magnetic resonance imaging with use of a resolution enhancement network |
| title_fullStr | Accelerated chemical shift encoded cardiovascular magnetic resonance imaging with use of a resolution enhancement network |
| title_full_unstemmed | Accelerated chemical shift encoded cardiovascular magnetic resonance imaging with use of a resolution enhancement network |
| title_short | Accelerated chemical shift encoded cardiovascular magnetic resonance imaging with use of a resolution enhancement network |
| title_sort | accelerated chemical shift encoded cardiovascular magnetic resonance imaging with use of a resolution enhancement network |
| topic | Deep learning Chemical shift encoding Super-resolution Dixon Myocardial fat |
| url | http://www.sciencedirect.com/science/article/pii/S1097664724011177 |
| work_keys_str_mv | AT manuelamorales acceleratedchemicalshiftencodedcardiovascularmagneticresonanceimagingwithuseofaresolutionenhancementnetwork AT scottjohnson acceleratedchemicalshiftencodedcardiovascularmagneticresonanceimagingwithuseofaresolutionenhancementnetwork AT patrickpierce acceleratedchemicalshiftencodedcardiovascularmagneticresonanceimagingwithuseofaresolutionenhancementnetwork AT rezanezafat acceleratedchemicalshiftencodedcardiovascularmagneticresonanceimagingwithuseofaresolutionenhancementnetwork |