Accelerated Patient-specific Non-Cartesian Magnetic Resonance Imaging Reconstruction Using Implicit Neural Representations.

Purpose: Accelerating magnetic resonance acquisition is essential for image guided therapeutic applications. Compressed sensing (CS) has been developed to minimize image artifacts in accelerated scans, but the required iterative reconstruction is computationally complex and difficult to generalize. Convolutional neural networks (CNNs)/Transformers-based deep learning methods emerged as a faster alternative but face challenges in modeling continuous k-space, a problem amplified with non-Cartesian sampling commonly used in accelerated acquisition. In comparison, implicit neural representations (INRs) can model continuous signals in the frequency domain and thus are compatible with arbitrary k-space sampling patterns. The current study developed novel k-space generative-adversarially trained INRs (k-GINR) for de novo undersampled non-Cartesian k-space reconstruction.

Methods and materials: k-GINR consists of 2 stages: 1) supervised training on an existing patient cohort; 2) self-supervised patient-specific optimization. The StarVIBE T1-weighted liver data set, consisting of 118 prospectively acquired scans and corresponding coil data, was employed for testing. k-GINR is compared with 2 INR-based methods, Neural Representation learning methodology with Prior embedding (NeRP) and k-space NeRP, an unrolled deep learning method, Deep Cascade CNN, and CS.

Results: k-GINR consistently outperformed the baselines with a larger performance advantage observed at very high accelerations (peak-signal-to-noise ratio: 6.8%-15.2% higher at 3 times, 15.1%-48.8% at 10 times, and 29.3%-60.5% higher at 20 times). The reconstruction times for k-GINR, NeRP, k-NeRP, CS, and Deep Cascade CNN were approximately 3 minutes, 4-10 minutes, 3 minutes, 4 minutes and 3 second, respectively.

Conclusions: k-GINR, an innovative 2-stage INR network incorporating adversarial training, was designed for direct non-Cartesian k-space reconstruction for new incoming patients. It demonstrated superior image quality compared to CS and Deep Cascade CNN across a wide range of acceleration ratios.