Generalizable, sequence‐invariant deep learning image reconstruction for subspace‐constrained quantitative MRI.
Saved in:
| Title: | Generalizable, sequence‐invariant deep learning image reconstruction for subspace‐constrained quantitative MRI. |
|---|---|
| Authors: | Hu, Zheyuan1,2,3 (AUTHOR), Chen, Zihao1,2,3 (AUTHOR), Cao, Tianle1,2,3 (AUTHOR), Lee, Hsu‐Lei3 (AUTHOR), Xie, Yibin3 (AUTHOR), Li, Debiao2,3 (AUTHOR), Christodoulou, Anthony G.1,2,3 (AUTHOR) achristodoulou@mednet.ucla.edu |
| Source: | Magnetic Resonance in Medicine. Jul2025, Vol. 94 Issue 1, p89-104. 16p. |
| Subjects: | Standard deviations, Deep learning, Cardiac magnetic resonance imaging, Image reconstruction, Magnetic resonance imaging |
| Abstract: | Purpose: To develop a deep subspace learning network that can function across different pulse sequences. Methods: A contrast‐invariant component‐by‐component (CBC) network structure was developed and compared against previously reported spatiotemporal multicomponent (MC) structure for reconstructing MR Multitasking images. A total of 130, 167, and 16 subjects were imaged using T1, T1‐T2, and T1‐T2‐ T2*$$ {\mathrm{T}}_2^{\ast } $$‐fat fraction (FF) mapping sequences, respectively. We compared CBC and MC networks in matched‐sequence experiments (same sequence for training and testing), then examined their cross‐sequence performance and generalizability by unmatched‐sequence experiments (different sequences for training and testing). A "universal" CBC network was also evaluated using mixed‐sequence training (combining data from all three sequences). Evaluation metrics included image normalized root mean squared error and Bland–Altman analyses of end‐diastolic maps, both versus iteratively reconstructed references. Results: The proposed CBC showed significantly better normalized root mean squared error than MC in both matched‐sequence and unmatched‐sequence experiments (p < 0.001), fewer structural details in quantitative error maps, and tighter limits of agreement. CBC was more generalizable than MC (smaller performance loss; p = 0.006 in T1 and p < 0.001 in T1‐T2 from matched‐sequence testing to unmatched‐sequence testing) and additionally allowed training of a single universal network to reconstruct images from any of the three pulse sequences. The mixed‐sequence CBC network performed similarly to matched‐sequence CBC in T1 (p = 0.178) and T1‐T2 (p = 0121), where training data were plentiful, and performed better in T1‐T2‐T2*$$ {\mathrm{T}}_2^{\ast } $$‐FF (p < 0.001) where training data were scarce. Conclusion: Contrast‐invariant learning of spatial features rather than spatiotemporal features improves performance and generalizability, addresses data scarcity, and offers a pathway to universal supervised deep subspace learning. [ABSTRACT FROM AUTHOR] |
| Copyright of Magnetic Resonance in Medicine is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Purpose: To develop a deep subspace learning network that can function across different pulse sequences. Methods: A contrast‐invariant component‐by‐component (CBC) network structure was developed and compared against previously reported spatiotemporal multicomponent (MC) structure for reconstructing MR Multitasking images. A total of 130, 167, and 16 subjects were imaged using T1, T1‐T2, and T1‐T2‐ T2*$$ {\mathrm{T}}_2^{\ast } $$‐fat fraction (FF) mapping sequences, respectively. We compared CBC and MC networks in matched‐sequence experiments (same sequence for training and testing), then examined their cross‐sequence performance and generalizability by unmatched‐sequence experiments (different sequences for training and testing). A "universal" CBC network was also evaluated using mixed‐sequence training (combining data from all three sequences). Evaluation metrics included image normalized root mean squared error and Bland–Altman analyses of end‐diastolic maps, both versus iteratively reconstructed references. Results: The proposed CBC showed significantly better normalized root mean squared error than MC in both matched‐sequence and unmatched‐sequence experiments (p < 0.001), fewer structural details in quantitative error maps, and tighter limits of agreement. CBC was more generalizable than MC (smaller performance loss; p = 0.006 in T1 and p < 0.001 in T1‐T2 from matched‐sequence testing to unmatched‐sequence testing) and additionally allowed training of a single universal network to reconstruct images from any of the three pulse sequences. The mixed‐sequence CBC network performed similarly to matched‐sequence CBC in T1 (p = 0.178) and T1‐T2 (p = 0121), where training data were plentiful, and performed better in T1‐T2‐T2*$$ {\mathrm{T}}_2^{\ast } $$‐FF (p < 0.001) where training data were scarce. Conclusion: Contrast‐invariant learning of spatial features rather than spatiotemporal features improves performance and generalizability, addresses data scarcity, and offers a pathway to universal supervised deep subspace learning. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 07403194 |
| DOI: | 10.1002/mrm.30433 |
