Physics-informed Model Selection for Robust Deep Learning Segmentation of Multi-center Perfusion CMR: Initial Findings from the SCMR Registry
Motivation: Accurate segmentation of first-pass perfusion CMR is critical for reliable myocardial blood flow analysis. Goal(s): To enable generalizaton to multi-center datasets. Approach: Incorporating cardiac magnetic resonance physics-informed features in the deep learning model selection process. Results: Our results, which leverage multi-center perfusion CMR datasets from the SCMR registry, suggest that combining physics-informed guidance with uncertainty-based deep learning model selection improves the segmentation performance in deep learning-based analysis of multi-center stress FPP datasets. Impact: The proposed hybrid approach has the potential to improve the reliability of fully automated stress/rest FPP analysis in clinical settings and in multi-center clinical trials.
Authors
- Luis Zamudio (ORCID: https://orcid.org/0000-0003-0206-9493)
- Orlando P. Simonetti (ORCID: https://orcid.org/0000-0002-8994-0095)
- Michael D. Elliott (ORCID: https://orcid.org/0000-0002-9057-3915)
- Robert M. Judd
- Dipan J. Shah (ORCID: https://orcid.org/0000-0002-6179-2393)
- Dilek M. Yalcinkaya
- Matthew Tong (ORCID: https://orcid.org/0000-0003-2482-1324)
- Venkateshwar Polsani (ORCID: https://orcid.org/0009-0004-6576-1297)
- Behzad Sharif (ORCID: https://orcid.org/0000-0002-8945-8178)
- Rohan Dharmakumar (ORCID: https://orcid.org/0000-0003-4120-4635)
- Khalid Youssef (ORCID: https://orcid.org/0000-0003-4429-367X)
- Arian M. Sohi
Publication Details
- Journal
- Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition
- Published
- 2025-09-16
- DOI
- https://doi.org/10.58530/2025/4280
- Primary Topic
- Advanced MRI Techniques and Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00