Search

Nuclear magnetic resonance signal dynamics as described by the Bloch equations are highly complex and often are without closed form solutions. This is especially the case for quantitative magnetic resonance fingerprinting (MRF) scans in which acquisition parameters are varied to efficiently probe the parameter space. As previously demonstrated, optimization of the pattern in which the MRI acquisition parameters are varied relative to the variance in the target quantitative parameters can improve experiment design. This process, however, relies on large scale non-linear optimizations with hundreds to thousands of unknowns. As such, the numerical optimization is extremely time consuming, highly sensitive to guesses, and prone to getting caught in local minima. Here, we describe a method to reduce the complexity of the optimal MRF experiment design by constraining the solutions to span a predetermined low-dimensional subspace. Compared with standard flip angle patterns in calibration phantom experiments, precision in T2 can be increased by optimizing as few as 8 coefficients in less than one minute of computation time.

Purpose: The goal of this work is to extend the capabilities of RAKI, a k-space interpolating neural network, to reconstruct high-quality images from in-plane accelerated simultaneous multislice imaging acquisitions. This method is referred to as slice-RAKI. Methods: A three-layer convolutional neural network was designed to output k-space signals for separate slices given the input of a multicoil slice-aliased k-space. The output of the slice-interpolation network is passed into a separate in-plane interpolating network for each slice. The proposed framework was tested in retrospective acceleration experiments in vivo, and in prospectively accelerated phantom and in vivo experiments. Results: The neural network interpolation based reconstruction quantitatively outperforms conventional parallel imaging reconstruction algorithms for all tested in-plane and simultaneous multislice acceleration factors. Visually, the neural network reconstructions are of superior quality compared to parallel imaging reconstructions in prospectively accelerated acquisitions. Conclusion: Slice-RAKI provides a patient specific neural network based non-linear reconstruction which improves image quality compared with conventional linear parallel imaging algorithms. It could find use in MR-guided interventions such as MR-guided radiation therapy for use in rapid real-time cine imaging for motion monitoring., Comment: 22 pages, 7 figures

Purpose: Magnetic resonance imaging (MRI) exams include multiple series with varying contrast and redundant information. For instance, T2-FLAIR contrast is based upon tissue T2 decay and the presence of water, also present in T2- and diffusion-weighted contrasts. T2-FLAIR contrast can be hypothetically modeled through deep learning models trained with diffusion- and T2-weighted acquisitions. Methods: Diffusion-, T2-, T2-FLAIR-, and T1-weighted brain images were acquired in 15 individuals. A convolutional neural network was developed to generate a T2-FLAIR image from other contrasts. Two datasets were withheld from training for validation. Results: Inputs with physical relationships to T2-FLAIR contrast most significantly impacted performance. The best model yielded results similar to acquired T2-FLAIR images, with a structural similarity index of 0.909, and reproduced pathology excluded from training. Synthetic images qualitatively exhibited lower noise and increased smoothness compared to acquired images. Conclusion: This suggests that with optimal inputs, deep learning based contrast generation performs well with creating synthetic T2-FLAIR images. Feature engineering on neural network inputs, based upon the physical basis of contrast, impacts the generation of synthetic contrast images. A larger, prospective clinical study is needed., Comment: Submitted to Magnetic Resonance in Medicine for peer review on 15 January 2019

Intrafraction motion can play a pivotal role in the success of abdominal and thoracic radiation therapy. Hybrid magnetic resonance-guided radiotherapy systems have the potential to control for intrafraction motion. Recently, we introduced an MRI sequence capable of acquiring real-time cine imaging in two orthogonal planes (SOPI). We extend SOPI here to permit dynamic updating of slice positions in one plane while keeping the other plane position fixed. In this implementation, cine images from the static plane are used for motion monitoring and as image navigators to sort stepped images in the other plane, producing dynamic 4D image volumes for use in dose reconstruction. A custom 3D-printed target filled with radiochromic FXG gel was interfaced to a dynamic motion phantom. 4D-SOPI was acquired in a dynamic motion phantom driven by an actual patient respiratory waveform displaying amplitude/frequency variations and drifting and in a healthy volunteer. Unique 4D-MRI epochs were reconstructed from a time series of phantom motion. Dose from a static 4cmx15cm field was calculated on each 4D respiratory phase bin and epoch image, scaled by the time spent in each bin, and then rigidly accumulated. The phantom was then positioned on an Elekta MR Linac and irradiated while moving. Following irradiation, actual dose deposited to the FXG gel was determined by applying a R1 versus dose calibration curve to R1 maps of the phantom. The 4D-SOPI cine images produced a respiratory motion navigator that was highly correlated with the actual phantom motion (CC=0.9981). The mean difference between the accumulated and measured dose inside the target was 4.4% of the maximum prescribed dose. These results provide early validation that 4D-SOPI simultaneously enables real-time motion monitoring and truth-in-delivery analysis for integrated MR-guided radiation therapy (MR-gRT) systems., Comment: 26 pages, 12 figures, submitted to Physics in Medicine and Biology