ISMRM25 - Image Reconstruction Using AI

ISMRM & ISMRT Annual Meeting & Exhibition • 10-15 May 2025 • Honolulu, Hawai'i

Image Reconstruction Using AI

Digital Poster

Acquisition & Reconstruction

Thursday, 15 May 2025

Exhibition Hall

14:15 - 15:15

Session Number: D-15

No CME/CE Credit

		Computer Number: 17 4602. Physics-Coupled Synthetic Data Generation Method Using Natural Images Enabling In-vivo Data-Free Complex MRI Denoising View Presentation Video S. Jung, D-H Kim Yonsei University, Seoul, Korea, Republic of Impact: Our method generates physics-coupled synthetic data from natural images, enabling effective complex MRI denoising without in-vivo data, achieving performance on par with in-vivo-trained models. This approach reduces dependence on large in-vivo datasets and addresses practical challenges in in-vivo data collection.
		Computer Number: 18 4603. Single-breathhold cardiac T1, T2, T2, and fat fraction mapping at 0.55T using dual rosette trajectory MRF and a deep image prior reconstruction View Presentation Video E. Cummings, G. Lima da Cruz, J. Hamilton, N. Seiberlich University of Michigan, Ann Arbor, United States Impact:* To reduce the time and number of breathholds required for a cardiac scan, we introduce a method for acquiring cardiac T₁, T₂, T₂*, and proton density fat fraction maps at 0.55T from a single breathhold, 16-heartbeat sequence.
		Computer Number: 19 4604. Developing a multi-channel deep-learning model for automatically quantifying malignant bone disease from Multiparametric Whole-Body MRI View Presentation Video A. Candito, R. Holbrey, L. D’Erme, S. Bottazzi, L. Russo, F. Castagnoli, A. Dragan, C. Messiou, N. Tunariu, D-M Koh, M. Blackledge The Institute of Cancer Research, London, United Kingdom Impact: Our multi-channel model can automatically quantify TDV and ADC from suspected malignant bone lesions across treatment with accuracy close to 70%, which can assist with clinical decision-making in patients with systemic cancer spread.
		Computer Number: 20 4605. Noise-Adaptive MRI Denoising Using Self-Supervised Learning with Average-to-Average (Avg2Avg) Loss View Presentation Video N. Janjusevic, M. Bruno, Y. Huang, J. Chen, Y. Wang, H. Chandarana, L. Feng Bernard and Irene Schwartz Center for Biomedical Imaging, New York University Grossman School of Medicine, New York, United States Impact: The proposed denoising technique could greatly encourage the use of 0.55T MRI and other low-SNR MRI scanners, making imaging more affordable and accessible.
		Computer Number: 21 4606. Deep-ERx2: Deep Learning Reconstruction for fast high-resolution non-Cartesian Compressed-Sensing MR Spectroscopic Imaging at 3T and 7T View Presentation Video P. Weiser, G. Langs, S. Motyka, B. Strasser, W. Bogner, P. Golland, N. Singh, J. Dietrich, E. Uhlmann, T. Batchelor, D. Cahill, G. Ungan, M. Hoffmann, A. Klauser, O. Andronesi Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Boston, MA, USA, Cambridge, United States Impact: Deep-ER enables high-resolution (3.4 mm isotropic) metabolic imaging with clinically feasible acquisition (4-9 min) and reconstruction times (1 min) at 3T and 7T. These times are compatible with the clinical workflow.
		Computer Number: 22 4607. Deep learning for magnetic resonance vessel wall image: image reconstruction, stenosis diagnosis and plaque calculation View Presentation Video f. fu, z. lin, x. yang, b. li Ruijin Hospital affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China Impact: A deep learning algorithm for magnetic resonance vessel wall interpretation accurately determined image reconstruction, vessel stenosis and plaque calculation, which achieved automatic postprocessing and had equivalent diagnostic performance when compared with experienced radiologists.
		Computer Number: 23 4608. Deep Learning-reconstruction of rapid 3DEPI acquisitions for enhanced QSM in the clinical assessment of Multiple Sclerosis View Presentation Video D. Gkotsoulias, M. Weigel, A. Cagol, N. de Oliveira Soares Siebenborn, J. Pfeuffer, C. Granziera University of Basel, Basel, Switzerland Impact: Deep learning-based reconstruction, denoising and super-resolution pipeline substantially enhances the quality of QSM maps obtained from fast 3DEPI. This holds promise for advancing the broader implementation of QSM in the clinical management of Multiple Sclerosis.
		Computer Number: 24 4609. High efficient abdominal T1 and T2 mapping via LMC-MOLED acquisition and GAL Multi-UNet reconstruction Y. Zheng, W. Chen, Q. Lin, L. Zhu, L. Lin, J. Wang, Z. Chen, S. Cai, C. Cai Xiamen University, Xiamen, China Impact: We proposed a new method for rapid simultaneous T₁ and T₂ mapping of abdomen, addressing the issue of long acquisition time in abdominal multi-parametric quantitative imaging, with significant potential value for clinical diagnosis.
		Computer Number: 25 4610. Joint attention deep learning reconstruction of highly-accelerated pre- and post-contrast T1-weighted 3D images of brain tumors View Presentation Video A. Mekhanik, R. Otazo Memorial Sloan Kettering Cancer Center, New York, United States Impact: A joint-attention deep learning reconstruction method can exploit correlations across sequences and enable significant reductions in MRI protocols.
		Computer Number: 26 4611. Fast and accurate reconstruction of accelerated 7T susceptibility-weighted imaging using multi-scale hybrid CNN-Transformer network C. Duan, D. Zhang, X. Bian, J. Qu, X. Lou The First Medical Center, Chinese PLA General Hospital, Beijing, China Impact: The potentially reduced scan time with MSCT-Net offers new possibilities for widely adoption of 7T SWI in clinical brain imaging. Furthermore, the approach has the potential to guide design of reconstruction models for other high-resolution 7T MRI data.
		Computer Number: 27 4612. Accelerated T2* Mapping of the Human Brain at 7T Using Deep Learning: Achieving 0.6 mm Isotropic Resolution in Under 6 Minutes View Presentation Video A. Klauser, E. Sleight, T. Yu, N. Pato Montemayor, J. Phillippe, D. Nickel, L. Bacha, T. Di Noto, B. Maréchal, T. Kober, T. Hilbert, G. F. Piredda Siemens Healthineers, Lausanne, Switzerland Impact: We demonstrate the efficacy of deep learning-based reconstruction for highly accelerated acquisitions, enabling 0.6mm isotropic R2* mapping of the brain in 6 minutes at 7T. This method highlights submillimeter T2* contrast, potentially enhancing its application in detecting microstructural alterations.
		Computer Number: 28 4613. CRNN with Bidirectional Frame-By-Frame Diffusion-Model-Based Refinement for Cardiac cine MRI Reconstruction H. Li, H. Sun, Z. Li, R. Yang, H. Chen Tsinghua University, Beijing, China Impact: A novel diffusion-model-based method of accelerated cardiac cine MRI reconstruction has been developed and improved the quality of reconstructed images. This may facilitate the application of accelerated cardiac cine MRI in clinical medicine, reducing patient discomfort and motion related artifacts.
		Computer Number: 29 4614. AI-assisted Collaborative Reconstruction for Highly-accelerated DW-PROPELLER-EPI View Presentation Video H. Xiong, L. Liang, S. Chen, X. Xu, C. Yuan, Y. Li, T. Liu, H-C Chang The Chinese University of Hong Kong, Hong Kong, China Impact: High geometric fidelity and high-resolution brain DWI with reasonable scan time may benefit clinical applications and neuroscience research.
		Computer Number: 30 4615. Robust Image Synthesis Method of Multi-Modal MRI Utilizing a Transformer Architecture View Presentation Video H. Zhang, H. Guo, M. Li, X. Liu Shenyang University of Technology, Shenyang, China Impact: This study introduces a novel technique for generating missing modality high quality images in MRI, which is robust to motion artifacts. The provision of structurally intact images enables clinicians to identify lesions more efficiently, augment diagnostic precision.
		Computer Number: 31 4616. Integrating Inline Quality Control at the MRI Scanner: Global and Local Assessment of Motion Artifacts Using Deep Learning View Presentation Video V. Ecker, M. Ganz, H. Eichhorn, E. Marchetto, T. Huelnhagen, B. Yang, S. Gatidis, T. Küstner University Hospital of Tübingen, Tübingen, Germany Impact: Our inline integration assessment for global and local image quality in MRI scans enables reliable detection of motion artifacts. This advancement allows for immediate corrective actions, improving diagnostic accuracy and optimizing imaging workflows.
		Computer Number: 32 4617. Unsupervised reconstruction of highly undersampled 3D cones cardiac image navigators using a dual-branch joint training framework View Presentation Video X. Guo, C. Sheagren, J. Patel, L. Li, G. Wright, F. Guo Huazhong University of Science and Technology, Wuhan, HuBei, China Impact: Our approach provides a way to reconstruct highly undersampled 3D cardiac images with sufficient quality for retrospective motion correction, suggesting the utility of our approach in various scenarios where fully sampled data is unavailable.

Similar Session(s)