Rapid 3D T₁ mapping using deep learning-assisted Look-Locker inversion recovery MRI

Purpose: Conventional 3D Look-Locker inversion recovery (LLIR) T₁ mapping requires multi-repetition data acquisition to reconstruct images at different inversion times for T₁ fitting. To ensure B₁ robustness, sufficient time of delay (TD) is needed between repetitions, which prolongs scan time. This work proposes a novel deep learning-assisted LLIR MRI approach for rapid 3D T₁ mapping without TD. Theory and Methods: The proposed approach is based on the fact that (Formula presented.), the effective T₁ in LLIR imaging, is independent of TD and can be estimated from both LLIR imaging with and without TD, while accurate conversion of (Formula presented.) to T₁ requires TD. Therefore, deep learning can be used to learn the conversion of (Formula presented.) to T₁, which eliminates the need for TD. This idea was implemented for inversion-recovery-prepared Golden-angel RAdial Sparse Parallel T₁ mapping (GraspT₁). 39 GraspT₁ datasets with a TD of 6 s (GraspT₁-TD6) were used for training, which also incorporates additional anatomical images. The trained network was applied for T₁ estimation in 14 GraspT₁ datasets without TD (GraspT₁-TD0). The robustness of the trained network was also tested. Results: Deep learning-based T₁ estimation from GraspT₁-TD0 is accurate compared to the reference. Incorporation of additional anatomical images improves the accuracy of T₁ estimation. The technique is also robust against slight variation in spatial resolution, imaging orientation and scanner platform. Conclusion: Our approach eliminates the need for TD in 3D LLIR imaging without affecting the T₁ estimation accuracy. It represents a novel use of deep learning towards more efficient and robust 3D LLIR T₁ mapping.