Cape Town - 2026 ISMRM-ISMRT Annual Meeting and Exhibition - Artificial Intelligence-Based Image Reconstruction in the Body

Artificial Intelligence-Based Image Reconstruction in the Body

Digital Poster

Body

Thursday, 14 May 2026

Digital Posters Row C

14:35 - 15:30

Session Number: 662-04

No CME/CE Credit

Similar Sessions

This session includes presentations that utilize Artificial Intelligence in body MRI.

		662-04-001. Delineating Dynamic Hyperpolarized Metabolic Signatures in Patient-derived Prostate Cancer Xenografts Using AI Erzsébet Merényi, Pratip Bhattacharya, Patrick Pilié, José Enriquez, Prasanta Dutta, Ziting Tang, Cesar Cardenas Rice University, Houston, United States of America Impact: Our AI approach can enable deep understanding and refinement of prostate cancer subtypes in response to therapy, from non-invasive HP-MR signatures. This has the potential to facilitate increased prediction accuracy and significant advancement of personalized medicine in a diverse population.
		662-04-002. First clinical translation of volumetric deep-learning super-resolution in 3D T2 breast MRI: faster acquisition,better images Narine Mesropyan, Oliver Weber, Christoph Katemann, Johannes Peeters, Can Yueksel, Alexander Isaak, Julian Luetkens University Hospital Bonn, Bonn, Germany Impact: A true 3D deep-learning super-resolution framework enables substantial scan-time reduction for 3D breast T2 MRI while improving image quality and maintaining BI-RADS consistency, facilitating clinical translation for either higher throughput or enhanced diagnostic image quality.
		662-04-003. Deep Learning-Reconstructed Thin-Slice VIBE Enhances Biliary Delineation and Lesion Detection in Hepatobiliary Phase MRI Haoran Dai, Jili Chen, Xuhao Song, Kai Liu, Caixia Fu, Mengsu Zeng Zhongshan Hospital, Shanghai, China Impact: This inline DL-reconstruction protocol resolves the spatial resolution-SNR trade-off, providing histology-grade lesion margin sharpness and surgical-grade biliary maps. It offers a immediate clinical upgrade for precision liver MRI without prolonging scan time.
		662-04-004. Self-Supervised Reconstruction and Denoising for High-Resolution Distortion-Free Prostate DWI Jingjia Chen, Haoyang Pei, Kun Zhou, Qiuting Wen, Mahesh Keerthivasan, Angela Tong, Hersh Chandarana, Li Feng New York University Grossman School of Medicine, New York, United States of America Impact: This work addresses the SNR limitations in distortion-free TGSE-PROPELLER-DWI by introducing a self-supervised learning-based reconstruction method. Improved image quality, enhanced resolution, and reduced scan times make it highly promising for advancing prostate imaging applications.
		662-04-005. Public-Trained, Clinic-Ready: Robust Diffusion-Prior Breast MRl Reconstruction Across Mulitple Sequences Zihao Wang, Shoujin Huang, Jiacai Cai, Lingyan Zhang, Hongyan Du, Yuqian Wang, Min Wang, Shaojun Liu, Mengye Lyu Shenzhen Technology University, Shenzhen, China Impact: Noise-adaptive diffusion reconstruction generalizes from public data to clinical breast MRI, enabling higher accelerations with better PSNR/SSIM, while preserving DCE TICs. These results may shorten exams, reduce motion, and motivate prospective studies.
		662-04-006. A unified fully automated 4D MRI for radiotherapy using pseudo-golden-angle radial k-space and deep-learning reconstruction Subin Erattakulangara, Victor Murray, Victoria Yu, Ricardo Otazo, Can Wu Memorial Sloan Kettering Cancer Center, New York, United States of America Impact: Accurate motion assessment is essential for precise radiotherapy of abdominal tumors. Our unified fully automated pipeline enables fast 4D MRI-based motion assessment of abdominal tumors and organs using standard clinical sequences, facilitating clinical translation of the technique for radiotherapy applications.
		662-04-007. Accelerated Abdominal MRI at 0.05 Tesla via Golden-angle Radial Sampling and Deep Learning Reconstruction Junhao Zhang, Yujiao Zhao, Vick Lau, Xiang Li, Jiahao Hu, Shi Su, Ye Ding, Alex T. L. Leong, Ed X Wu The University of Hong Kong, Hong Kong, China Impact: It is the first time that the proposed method could accelerate 3D T2W abdominal MRI at 0.05 Tesla for about 4 minutes per scan.
		662-04-008. Prospective Comparison of Conventional and Deep Learning-Reconstructed Thin-Slice 3D T1-weighted imaging of the Breast Nan Zhang, Xuhao Song, Caixia Fu, Marcel Dominik Nickel, Mengsu Zeng Zhongshan Hospital, Shanghai, China Impact: thin-slice VIBE_DL sequence achieved better image quality, noise reduction, image sharpness, artifacts mitigation, and diagnostic confidence, as well as lesion conspicuity and detection compared to the conventional VIBE sequence.
		662-04-009. Super-Resolution Deep Learning Reconstruction Improves Assessment of Myometrial Invasion in Endometrial Cancer Takumi Tanigaki, Atsushi Nakamoto, Hideyuki Fukui, Takashi Ota, Toru Honda, Kengo Kiso, Eriko Yoshidome, Masatoshi Hori, noriyuki tomiyama The University of Osaka Graduate School of Medicine, Suita, Japan Impact: SR-DLR enables simultaneous improvement in signal-to-noise ratio and spatial resolution for endometrial cancer imaging, potentially improving staging accuracy and treatment planning while maintaining clinically feasible scan times, representing a significant advance in gynecologic oncology imaging.
		662-04-010. Synergizing PROPELLER and deep learning reconstruction to optimize 1.5T lumbar spine T2w-STIR imaging Qingbin Meng, Shan Luo, Qian Zhang, Guangxu Han, Min Li, Fang Shen Shaanxi Provincial Rehabilitation Hospital for Disabled Veterans, Weinan, China Impact: PDLR-T2w-STIR improves image quality and reduces scan time in 1.5T lumbar MRI, offering a promising solution to motion artifacts and workflow inefficiencies without compromising diagnostic performance.
		662-04-011. Deep Learning–Based Accelerated Diffusion-Weighted Imaging for Liver Lesion Evaluation: Improved Image Quality and Diagnostic Wenjie Xu, Fuquan Wei, Guoqun Mao, Yijiang Huang, Nan Chen, Yunzhu Wu, Thomas Benkert Tongde Hospital of Zhejiang Province, Hangzhou, China Impact: DL-based DWI enables faster, clearer liver imaging, improving lesion detection and supporting more confident clinical diagnosis.
		662-04-012. Complexed Signal Average for Prostatic DWI: Improving Diagnostic Performance of Malignant from Benign Prostatic Areas on DWIs Takahiro Ueda, Natsuka Yazawa, Kaori Yamamoto, Yuichiro Sano, Masato Ikedo, Masanori Ozaki, Masahiko Nomura, Takeshi Yoshikawa, Daisuke Takenaka, Masahiro Endo, Yoshiyuki Ozawa, Yoshiharu Ohno Fujita Health University School of Medicine, Toyoake, Japan Impact: Complexed signal average (CSA)mainly improves image quality and differentiation capability of malignant from benign prostatic areas on DWI with standard and ultra-high b values.
		662-04-013. Evaluation of a Deep Learning Based Accelerated 3D Acquisition Strategy for T2-Weighted MRI of the Prostate Eugene Milshteyn, Trevor Kolupar, Arnaud Guidon, Ty Cashen, Ajeetkumar Gaddipati, Nabih Nakrour, Mukesh Harisinghani, Rory Cochran GE HealthCare, San Ramon, United States of America Impact: Deep learning accelerated 3D T2-weighted prostate imaging can provide a new avenue for fast, high resolution prostate mpMRI and improve lesion localization and diagnostic confidence.
		662-04-014. Robust High-Resolution Multi-Organ Diffusion MRI Using Synthetic-Data-Tuned Prompt Learning Chen Qian, Haoyu Zhang, Junnan Ma, Ke Jiang, Mingyang Han, xianwang jiang, Lv Li, Yingsa Li, Qin Xu, Hai Zhong, Xiaobo Qu Fujian Provincial Key Laboratory of Plasma and Magnetic Resonance, Xiamen University, Xiamen, China Impact: The approach eliminates navigator signals and realistic data supervision, providing an interpretable, robust solution for high-resolution multi-organ multi-shot DWI. Its scanner-agnostic performance signifies transformative potential for precision oncology
		662-04-015. Self-Supervised Deep Learning Model for Estimating Microstructural Parameters in Breast Cancer Juanhua Zhang, Lei Wu, Yuan Lian, Fan Liu, Peijiang Ma, Kaihan Yang, Haihua Bao, Diwei Shi, Hua Guo Tsinghua University, Beijing, China Impact: This self-supervised deep learning model surpasses NLLS fitting in estimating IMPULSED microstructural parameters, enabling higher accuracy and shorter computation time.
		662-04-016. Breast DWI using complex signal averaging and super resolution DLR - impact on quality with breast phantom MASAKO KATAOKA, Maya Honda, Koji Fujimoto, Kanae Miyake, Sachi Okuchi, Mami Iima, Hajime Morizumi, Rimika Imai, Yuichiro Sano, Yuji Nakamoto Kyoto University Hospital, Kyoto, Japan Impact: The phantom study simulating breast DWI indicates that "Complex signal averaging (CSA)" with super-resolution DLR can reduce noise, improve image quality while allow reliable ADC quantification. Sufficient evidence to apply CSA and DLR to clinical breast DWI is demonstrated.