Longitudinal guided sparsity improves accelerated brain MRI reconstruction
LGGS: Longitudinal MRI Reconstruction via Guided Group Sparsity
Computer Vision and Pattern Recognition
Summary
MRI scans can take a long time, which makes it hard for patients to get quick exams. The authors introduced a new method called LGGS that uses previous scans of the same patient to help reconstruct new MRI images faster and more accurately. Their technique handles changes between visits and differences in image alignment and phase. This approach performed better than other methods on brain MRI data while using fewer model parameters.
What this means in practice
- •For radiology teams: Speed up follow-up brain MRI exams using prior scans to improve image quality with fewer data and faster reconstruction.
- •For medical device manufacturers: Incorporate LGGS into MRI scanner software to reduce scan times for serial imaging in therapies like MR-guided radiotherapy.$Commercial implications: Enables faster MRI products that leverage prior patient images to improve reconstruction, creating a more competitive clinical imaging device.
Authors
Nikola Janjušević, Hersh Chandarana, Yao Wang, Li Feng
Abstract
Serial MRI is routinely performed in current clinical practice for disease surveillance and applications such as MR-guided radiotherapy. However, long MRI scan times increase costs and limit patient access, making accelerated image reconstruction of broad clinical importance. Previous scans of the same patient contain image information highly correlated with the current exam, yet existing reconstruction methods rarely exploit this information. Major challenges in the effective use of prior scans include inter-visit misregistration, pathological changes, and image phase variation. In this work, we propose Longitudinal Guided Group Sparsity (LGGS), an interpretable deep-learning framework for longitudinally informed MRI reconstruction derived from classical group sparsity. The prior scan is incorporated into the reconstruction regularizer through learned phase-invariant circulant attention that identifies corresponding features across exams. This formulation provides robustness to local in-plane and through-plane misregistration as well as inter-visit phase variation. Across acceleration rates, LGGS achieves state-of-the-art performance on longitudinal T1-weighted multi-coil brain MRI data, outperforming the evaluated unguided and longitudinally informed reconstruction methods, with substantially fewer learned parameters. We further evaluated the contributions of multiple guide slices and phase-invariant similarity and demonstrated that accelerated prior images can also serve as effective guides with minimal degradation in reconstruction quality.