ISMRM25 - Generating Synthetic Imaging Data: Part II

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

Generating Synthetic Imaging Data: Part II

Digital Poster

AI & Machine Learning

Wednesday, 14 May 2025

Exhibition Hall

08:15 - 09:15

Session Number: D-32

No CME/CE Credit

		Computer Number: 49 3232. Multi-Contrast Generation and On Demand Quantification of Saturation Transfer, Relaxivity, and Field Homogeneity using a Deep MRI on a Chip View Presentation Video D. Nagar, M. Zaiss, O. Perlman Tel Aviv University, Tel Aviv, Israel Impact: A deep learning framework was designed to provide rich biological information in less than 30 seconds. It can capture the magnetic signal dynamics in humans and decode the tissue response to RF excitation, constituting a deep MRI on a chip.
		Computer Number: 50 3233. Hyperpolarized 129Xe MR Image Denoising Model Based on Multimodal Edge Information View Presentation Video M. Zhang, S. Xiao, S. Shi, X. Zhou State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, National Center for Magnetic Resonance in Wuhan, Wuhan Institute of Physics and Mathematics, Innovation Academy for Precision Measurement Science and Technology, Chinese Academy of Sciences–Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology, 430071, wuhan, China Impact: Our proposed image denoising model combines multimodal edge information and uses the attention mechanism to increase the weight of important information, which effectively improves the image details and quality, and improves the help for the clinical assessment of related diseases.
		Computer Number: 51 3234. SRIP: A Self-Refined Iterative Pipeline to Integrate Whole Body MRI-CT Registration and Pseudo-CT Generation View Presentation Video J. Hu, H. Yousefi, T. Fraum, J. Crandall, R. Wahl, R. Laforest, Y. Chen, H. An Washington University in St. Louis, St. Louis, United States Impact: SRIP introduces a novel pCT generation pipeline that produces high-quality pCT images through a fully automated approach, demonstrating the potential for various applications, such as radiation therapy and PET attenuation correction.
		Computer Number: 52 3235. Diffusion-guided MR Brain Tumor Segmentation with Missing Modalities View Presentation Video Y. Zhang, Y. Huang, J. Qi GE Healthcare, Beijing, China Impact: This approach enhances brain tumor segmentation accuracy and consistency, even with missing MRI modalities, thereby improving diagnostic precision and supporting clinical decision-making.
		Computer Number: 53 3236. Attention-Based Inverted Minimum Intensity Projection-Guided GAN for 7T-Like SWI Generation from 3T SWI View Presentation Video W. Tang, R. Zheng, Y-H Chu, C. Wang, H. Wang Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University, Shanghai, China Impact: This study was an early exploration of artificial intelligence methods in the field of cross-field-strength SWI image generation. It provided insights for related research on enhancing the image quality and diagnostic capability of low-field MR images.
		Computer Number: 54 3237. Label-informed Data Augmentation for DBS Optimization: Synthesizing fMRI Maps with SPADE-VAE Network to Improve DBS Parameter Classification View Presentation Video J. Qiu, A. Ajala, J. Germann, B. Santyr, D. Yeo, L. Marinelli, A. Boutet, A. Lozano GE HealthCare, Niskayuna, United States Impact: By producing realistic synthetic DBS-fMRI maps, the trained SPADE-VAE model addresses data gap issues in DL-based fMRI-DBS optimization, corrects class imbalances, enhances classification accuracy and ultimately reduces the time to optimization per patient in Parkinson’s disease treatment.
		Computer Number: 55 3238. DTI with Minimal Data: Image Translation Based Distortion Correction and FA Map Generation for Clinical Efficiency Y. Cui, H. Qi, Z. Zhang, H. Zhang, S. Yuan, B. Cai, J. Li, M. Zhang, Z. Wang, L. Tong, J. Luo School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai, China Impact: Success of the proposed pipeline will enable much shorter DTI acquisition time for patients who cannot stay still throughout a multidirectional DTI scans, which holds great potential to becoming a promising tool for clinical applications.
		Computer Number: 56 3239. Generation of 3D Super-resolution FLAIR Images from Heterogeneous 2D FLAIR Acquisitions in Epilepsy Patients View Presentation Video S. Morris, T-Y Su, A. Alexopoulos, I. Najm, Z. Wang Cleveland Clinic, Cleveland, United States Impact: Our method can assist clinicians to evaluate the existence and the extent of epileptic lesions in data-limited situations in various health care settings. The generated images may also serve as inputs to inform AI models for automated lesion detection.
		Computer Number: 57 3240. Generation of 123I-IMP brain SPECT from 3D T1-weighted imaging using a machine-learning-based model View Presentation Video S. Okuchi, Y. Fushimi, K. Fujimoto, T. Seguchi, A. Iohara, S. Nakajima, A. Sakata, T. Yamamoto, S. Otani, A. Sakurama, S. Ikeda, S. Ito, M. Umehana, Y. Ma, S. Morooka, J. Fujimoto, S. Shinjo, Y. Nakamoto Graduate School of Medicine, Kyoto University, Kyoto, Japan Impact: This study shows that a machine learning model trained on a combined dementia and Parkinsonism dataset can generate accurate SPECT-like images from MPRAGE, potentially reducing the need for traditional SPECT imaging in neurological assessments.
		Computer Number: 58 3241. Turing Testing the Realism of U-Net, GAN, and Real Images for Virtual Contrast-Enhanced Breast MRI View Presentation Video A. George, H. Schreiter, J. Hossbach, T-T Nguyen, I. Horishnyi, C. Ehring, S. Heidarikahkesh, L. Kapsner, F. Laun, S. Ohlmeyer, M. Uder, S. Bickelhaupt, A. Liebert Institute of Radiology, Uniklinikum Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany, Erlangen, Germany Impact: DL-based vCE provides an effective alternative to significantly reduce reliance on contrast agents, addressing concerns of cost, time, and contraindications. Incorporating adversarial training enables the model to learn intricate details of contrast-enhanced images, resulting in more realistic outputs.
		Computer Number: 59 3242. Self-Supervised Isotropic MRI Volume Restoration from Complementary Contrast-Plane Acquisition using Two-Phase Fast Score-based Models View Presentation Video H. Kim, Y. Song, T. Çukur, D. Hwang Yonsei University, Seoul, Korea, Republic of Impact: This method enhances clinical MRI protocols by enabling isotropic multi-contrast volume reconstruction from anisotropic data, improving diagnostic consistency across contrasts. It reduces the need for extended scan times, maximizing data utility and facilitating broader clinical insights in routine practice.
		Computer Number: 60 3243. AI-based Synthetic Contrast-enhanced MR Images from Noncontrast MR Images for Various Brain Tumors beyond Gliomas View Presentation Video H. Takita, H. Tatekawa, K. Nakajo, T. Uda, Y. Mitsuyama, S. Walston, Y. Miki, D. Ueda Graduate School of Medicine, Osaka Metropolitan University, Osaka, Japan Impact: Our study suggests potential clinical applicability of AI-based synthetic contrast-enhanced MR images generated from noncontrast MR images across various brain tumor types, offering a promising alternative for patients unable to receive gadolinium-based contrast agents.
		Computer Number: 61 3244. Generating Synthetic Late-Gadolinium Enhancement Images for Training Scar Segmentation Networks View Presentation Video I. Margolis, L. Dal Toso, S. Buoso, S. Kozerke ETH Zurich, Zurich, Switzerland Impact: This study presents a GAN-based framework to generate high-variability synthetic LGE images for scar segmentation, achieving realistic scar patterns and outperforming real-image networks and standard augmentation.
		Computer Number: 62 3245. Synthesizing PET images with ASL CBF for AD diagnosis View Presentation Video Q. Shou, D. Wang University of Southern California, Los Angeles, United States Impact: This study provides the proof-of-concept that noninvasive ASL can be applied to synthesize PET images for the diagnosis and management of AD without the cost and radiation exposure of an actual PET scan.
		Computer Number: 63 3246. Prediction of EM Field in a Simple Homogenous Phantom at UHF MRI Using Physics-Informed Neural Networks (PINNs): Methodology in Data Generation View Presentation Video F. Jabbarigargari, A. Dulny, M. Terekhov, A. Krause, A. Hotho, L. Schreiber Comprehensive Heart Failure Center, University Hospital Wuerzburg, Wuerzburg, Germany Impact: This study proposes a deep learning-based method for EM field prediction, which, by significantly reducing the computational time, can enable safer and more accessible 7T MRI.
		Computer Number: 64 3247. Counterfactual MRI Data Augmentation using Conditional Denoising Diffusion Models:training AI for different clinical scenarios View Presentation Video J. Santinha, P. Morão, Y. Forghani, N. Loução, T. Correia, P. Gouveia, M. Figueiredo Breast Unit/Digital Surgery LAB, Champalimaud Foundation, Lisboa, Portugal Impact: This method can improve the performance of DL models in clinical settings by enabling them to generalize across different acquisition settings. This could lead to more reliable and robust diagnoses, particularly in scenarios with limited access to diverse training data.

Similar Session(s)