Cape Town - 2026 ISMRM-ISMRT Annual Meeting and Exhibition

Breast MRI and the Role of AI

Digital Poster

Body

Wednesday, 13 May 2026

Digital Posters Row C

09:15 - 10:10

Session Number: 562-02

No CME/CE Credit

Similar Sessions

This session presents the cutting edge AI applications for breast MRI and their potential value for enhancing patient care.

		562-02-001. Deep Learning System for Predicting ER Negative, -low, and -high Positive Breast Cancer Using MRI Yi Dai, Zhenwei Shi Peking University Shenzhen Hospital, Shenzhen, China Impact: This non-invasive deep learning system enables preoperative, precise ER status classification, directly informing personalized therapy decisions. It empowers clinicians to identify elusive ER-low positive patients and opens new avenues for radiogenomics research, advancing image-guided precision oncology.
		562-02-002. Voxel-wise Variability in Breast DCE-MRI: Monte Carlo Comparison of Patlak, ETM-FXL, and ETM-NXL Pharmacokinetic Models Rachaita Podder, Sai Man Cheung, Kangwa Nkonde, Andrew Blamire, Jiabao He Newcastle University, Newcastle upon Tyne, United Kingdom Impact: Voxel-wise variability mapping informs the reliability of pharmacokinetic parameters in quantitative DCE-MRI. The lower variability observed in pharmacokinetic parameters from Patlak and ETM-FXL models supports the potential clinical utility as robust imaging biomarkers for precision treatment monitoring in breast cancer.
		562-02-003. Intratumoral Heterogeneity of DCE-MRI for Predicting Pathological Response to Neoadjuvant Chemotherapy in Breast Cancer Wenjie Jiang, Chen Bai, Xueying Zhao, Xinwei Tang, Xingke Huang, Xiaoying Li, Xiang Zhang, jun shen Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University,, Guangzhou, China Impact: Integrating habitat radiomics and intratumoral heterogeneity from DCE-MRI pharmacokinetic maps enhances the accuracy of pathological complete response prediction after neoadjuvant chemotherapy in breast cancer, supporting personalized treatment planning and advancing noninvasive precision oncology.
		562-02-004. Integrating Global and Habitat-Based Radiomics from DCE-MRI for Improved Characterization of Breast Lesions Kexin Chen, Ya Ren, Xuetong Tao, Meng Wang, Lin Li, Shuluan Chen, Jie Wen, wei cui, Zhanli Hu, Xin Liu, Dong Liang, Hairong Zheng, Qian Wan, Zhou Liu, Na Zhang State Key Laboratory of Biomedical Imaging Science and System, Key Laboratory of Biomedical Imaging Science and System, Chinese Academy of Sciences, Shenzhen, China Impact: This novel approach provides an intuitive and interpretable framework that integrates global and habitat-based radiomics to visualize and quantify spatial and temporal hemodynamic heterogeneity, enhancing diagnostic performance and offering potential for broader clinical translation.
		562-02-005. Integrating BI-RADS, Radiomics, and Deep Learning for MRI-Based HER2 Phenotype Differentiation Thais Bezerra, Soraia Damião, Claudio Patrocínio, Adriana Bueno, Marina De Brot, Almir Bitencourt University of Campinas, Campinas, Brazil Impact: This study compares conventional, radiomic, and deep learning models using breast MRI to differentiate HER2 expression levels in invasive breast cancer. It highlights MRI’s potential as a non-invasive surrogate for molecular profiling and personalized treatment planning.
		562-02-006. Integrating DCE-MRI Multiparametric Maps and Digital Pathological Images to Predict Pathological Response in Breast Cancer Huidong Wang, Siyao Du, Ruimeng Zhao, LINA ZHANG the Fourth Affiliated Hospital of China Medical University, Shenyang, China Impact: Combining hemodynamic information from DCE-MRI with digital pathological data results in the optimal model. This model can effectively predict the efficacy of neoadjuvant chemotherapy (NAC) for breast cancer at an early stage, which has clinical guidance.
		562-02-007. MRI-based Habitat radiomics for risk stratification of axillary lymph node metastasis in breast cancer: A multicenter study Yuqin Peng, Zhiyao Liu, Xueying Zhao, miaomiao ding, Shuxin He, Xiang Zhang Shenshan Medical Center, Memorial Hospital of Sun Yat-Sen University, Shanwei, China Impact: The model stratifies patients into low-, intermediate-, and high-risk groups: low-risk patients may be exempted from SLNB; intermediate-risk patients could omit SLND; and high-risk patients should undergo ALND directly. It can be used preoperatively to assess ALNM in breast cancer.
		562-02-008. Prediction of Residual Axillary Metastasis Following NAT for Breast Cancer: Habitat Radiomics Analysis Based on breast MRI Junjie Zhang Shanxi Province Cancer Hospital/ Shanxi Hospital Affiliated to Cancer Hospital, Chinese Academy of Medical Sciences/Cancer Hospital Affiliated to Shanxi Medical University, Taiyuan, China Impact: The DCE-MRI-derived habitat model demonstrates great potential in distinguishing between axillary pCR and residual axillary lymph node metastasis, which is conducive to assisting clinicians in formulating individualized axillary management strategies.
		562-02-009. Distinguishing benign and malignant masses of breast tumors in DCE-MRI using a spatial-temporal encoding model Tianyi Zhang, Lin Li, Yuan Guo, Jianming Rong, Yuanhao Li, Ya Ren, Meng Wang, Shuluan Chen, Jie Wen, wei cui, Dehong Luo, Zhou Liu, Zhenxing Huang Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China Impact: We developed and clinically validated a novel Spatial-Temporal Encoding Methodology (STEM) for automated processing of DCE-MRI time-intensity curve data to achieve simultaneous differentiation of benign and malignant breast tumors while providing accurate tumor staging capabilities.
		562-02-010. Evaluating Synthetic T2-Weighted Breast MRI: A Multi-Reader, Multicenter Study Antonio Portaluri, Luyi Han, Jarek van Dijk, Xinglong Liang, Marialena Tsarouchi, Carla Carla Sitges, Miguel Braga, Noemi Schmidt, Maria Adele Marino, Tao Tan, Tianyu Zhang, Ritse Mann the Netherlands Cancer Institute, Amsterdam, Netherlands Impact: By restoring T2 information, IMPORTANT-NET could expand access to multiparametric breast MRI without prolonged scan-time. This multicenter reader study demonstrates visual equivalence and no artifact penalty for generated T2 sequences, with real-versus-synthetic identification at chance level and consistent diagnostic acceptability.
		562-02-011. Artificial Intelligence Detection of Breast Cancer on Abbreviated MRI Zahra Aghdam, Xin Wang, Luuk Balkenende, Jonas Teuwen, Koen Eppenhof, Kevin Groot Lipman, Ritse Mann Radboud University Medical Center, Nijmegen, Netherlands, Netherlands Impact: The AI system showed acceptable discriminatory performance in detecting cancer on abbreviated breast MR examination, the findings emphasize the importance of further clinical validation to account for the risk of false positives and ensure safe integration into practice.
		562-02-012. Prediction of lymphovascular invasion in invasive breast cancer using synthetic MRI-derived histogram parameters Zhou Lu, Xiao Yang, Siwei Zhang, Zekai Wang, Jie Shi, Zongqiong sun Department of radiology, Affiliated hospital of Jiangnan university, Wuxi, China Impact: The integration of SyMRI-derived histogram parameters into a clinical model offers a promising, non-invasive approach to improve the preoperative detection of LVI, potentially leading to more personalized treatment strategies for breast cancer patients.