Computer Science Faculty Publications
Quantitative Phosphocreatine Mapping with a Transformer-Based Regression Network
Document Type
Conference Proceeding
Publication Date
3-10-2026
Abstract
Phosphocreatine (PCr) is the key element of buffering adenosine triphosphate (ATP) concentration for sustaining the high energy demand of cardiomyocytes, thus maintaining normal mechanical function. CEST MRI has been employed to capture potential changes in myocardial PCr. A quantitative PCr network (QPCrNet) has been developed to provide accurate and fast PCr maps for the assessment of cardiac energy metabolism. In the simulation study, QPCrNet showed superior accuracy (lowest mean absolute error) in predicting PCr concentration and exchange rate from PCr pool to water pool. The transfer learning of QPCrNet on a canine stenosis model further proved the network capability of predicting regional difference between the diseased and remote areas and accurately predicted PCr reductions from the resting state to Dobutamine-induced stress.
Recommended Citation
Huang, Qi, Haoteng Tang, Han Tang, Caleb Berberet, Liya Dai, Thomas Schindler, Linda Peterson et al. "Quantitative Phosphocreatine Mapping with a Transformer-Based Regression Network." In 2025 IEEE International Conference on Data Mining Workshops (ICDMW), pp. 2015-2019. IEEE, 2025. https://doi.org/10.1109/ICDMW69685.2025.00244
Publication Title
2025 IEEE International Conference on Data Mining Workshops (ICDMW)
DOI
10.1109/ICDMW69685.2025.00244

Comments
©2025 IEEE