Purpose To accelerate coronary MRI acquisitions with arbitrary undersampling patterns by using a novel reconstruction algorithm that applies coil self-consistency using subject-specific neural networks. Methods Self-consistent robust artificial-neural-networks for k-space interpolation (sRAKI) performs iterative parallel imaging reconstruction by enforcing self-consistency among coils. The approach bears similarity to SPIRiT, but extends the linear convolutions in SPIRiT to nonlinear interpolation using convolutional neural networks (CNNs). These CNNs are trained individually for each scan using the scan-specific autocalibrating signal (ACS) data. Reconstruction is performed by imposing the learned self-consistency and data-consistency, which enables sRAKI to support random undersampling patterns. Fully-sampled targeted right coronary artery MRI was acquired in six healthy subjects. The data were retrospectively undersampled, and reconstructed using SPIRiT, l₁-SPIRiT and sRAKI for acceleration rates of 2 to 5. Additionally, prospectively undersampled whole-heart coronary MRI was acquired to further evaluate reconstruction performance. Results sRAKI reduces noise amplification and blurring artifacts compared with SPIRiT and l₁-SPIRiT, especially at high acceleration rates in targeted coronary MRI. Quantitative analysis shows that sRAKI outperforms these techniques in terms of normalized mean-squared-error (~44% and ~21% over SPIRiT and ‘₁-SPIRiT at rate 5) and vessel sharpness (~10% and ~20% over SPIRiT and l₁-SPIRiT at rate 5). Whole-heart data shows the sharpest coronary arteries when resolved using sRAKI, with 11% and 15% improvement in vessel sharpness over SPIRiT and l₁-SPIRiT, respectively. Conclusion sRAKI is a database-free neural network-based reconstruction technique that may further accelerate coronary MRI with arbitrary undersampling patterns, while improving noise resilience over linear parallel imaging and image sharpness over l₁ regularization techniques.

Parallel imaging is a widely-used acceleration technique for magnetic resonance imaging (MRI). Conventional linear reconstruction approaches in parallel imaging suffer from noise amplification. Recently, a non-linear method that utilizes subject- specific convolutional neural networks for k-space reconstruction, Robust Artificial-neural-networks for k-space Interpolation (RAKI) was proposed and shown to improve noise resilience over linear methods. However, the linear convolutions still provide a sufficient baseline image quality and interpretability. In this paper, we sought to utilize a residual network architecture to combine the benefits of both the linear and non-linear RAKI reconstructions. This hybrid method, called residual RAKI (rRAKI) offers significant improvement in image quality compared to linear method, and improves upon RAKI in highly- accelerated simultaneous multi-slice imaging. Furthermore, it establishes an interpretable view for the use of CNNs in parallel imaging, as the CNN component in the residual network removes the noise amplification arising from the linear part.

Background Robust Artificial-neural-networks for k-space Interpolation (RAKI) is a recently proposed deep-learning-based reconstruction algorithm for parallel imaging. Its main premise is to perform k-space interpolation using convolutional neural networks (CNNs) trained on subject-specific autocalibration signal (ACS) data. Since training is performed individually for each subject, the reconstruction time is longer than approaches that pre-train on databases. In this study, we sought to reduce the computational time of RAKI. Methods RAKI was implemented using CPU multi-processing and process pooling to maximize the utility of GPU resources. We also proposed an alternative CNN architecture that interpolates all output channels jointly for specific skipped k-space lines. This new architecture was compared to the original CNN architecture in RAKI, as well as to GRAPPA in phantom, brain and knee MRI datasets, both qualitatively and quantitatively. Results The optimized GPU implementations were approximately 2-to-5-fold faster than a simple GPU implementation. The new CNN architecture further improved the computational time by 4-to-5-fold compared to the optimized GPU implementation using the original RAKI CNN architecture. It also provided significant improvement over GRAPPA both visually and quantitatively, although it performed slightly worse than the original RAKI CNN architecture. Conclusions The proposed implementations of RAKI bring the computational time towards clinically acceptable ranges. The new CNN architecture yields faster training, albeit at a slight performance loss, which may be acceptable for faster visualization in some settings.