skip to main content

Explore Research Products in the NSF-PAR

This content will become publicly available on October 17, 2024

Title: Coil sketching for computationally efficient MR iterative reconstruction
Abstract Purpose

Parallel imaging and compressed sensing reconstructions of large MRI datasets often have a prohibitive computational cost that bottlenecks clinical deployment, especially for three‐dimensional (3D) non‐Cartesian acquisitions. One common approach is to reduce the number of coil channels actively used during reconstruction as in coil compression. While effective for Cartesian imaging, coil compression inherently loses signal energy, producing shading artifacts that compromise image quality for 3D non‐Cartesian imaging. We propose coil sketching, a general and versatile method for computationally‐efficient iterative MR image reconstruction.

 Theory and Methods

We based our method on randomized sketching algorithms, a type of large‐scale optimization algorithms well established in the fields of machine learning and big data analysis. We adapt the sketching theory to the MRI reconstruction problem via a structured sketching matrix that, similar to coil compression, considers high‐energy virtual coils obtained from principal component analysis. But, unlike coil compression, it also considers random linear combinations of the remaining low‐energy coils, effectively leveraging information from all coils.

 Results

First, we performed ablation experiments to validate the sketching matrix design on both Cartesian and non‐Cartesian datasets. The resulting design yielded both improved computatioanal efficiency and preserved signal‐to‐noise ratio (SNR) as measured by the inverse g‐factor. Then, we verified the efficacy of our approach on high‐dimensional non‐Cartesian 3D cones datasets, where coil sketching yielded up to three‐fold faster reconstructions with equivalent image quality.

 Conclusion

Coil sketching is a general and versatile reconstruction framework for computationally fast and memory‐efficient reconstruction.

  more » « less
NSF-PAR ID:
10469557
Author(s) / Creator(s):
 ;  ;  ;  ;  ;  ;  ;  ;  
Publisher / Repository:
Wiley Blackwell (John Wiley & Sons)
Date Published:
Journal Name:
Magnetic Resonance in Medicine
Volume:
91
Issue:
2
ISSN:
0740-3194
Format(s):
Medium: X Size: p. 784-802
Size(s):
["p. 784-802"]
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ; ( , Magnetic Resonance in Medicine)
    Purpose

    To develop a scan‐specific model that estimates and corrects k‐space errors made when reconstructing accelerated MRI data.

     Methods

    Scan‐specific artifact reduction in k‐space (SPARK) trains a convolutional‐neural‐network to estimate and correct k‐space errors made by an input reconstruction technique by back‐propagating from the mean‐squared‐error loss between an auto‐calibration signal (ACS) and the input technique’s reconstructed ACS. First, SPARK is applied to generalized autocalibrating partially parallel acquisitions (GRAPPA) and demonstrates improved robustness over other scan‐specific models, such as robust artificial‐neural‐networks for k‐space interpolation (RAKI) and residual‐RAKI. Subsequent experiments demonstrate that SPARK synergizes with residual‐RAKI to improve reconstruction performance. SPARK also improves reconstruction quality when applied to advanced acquisition and reconstruction techniques like 2D virtual coil (VC‐) GRAPPA, 2D LORAKS, 3D GRAPPA without an integrated ACS region, and 2D/3D wave‐encoded imaging.

     Results

    SPARK yields SSIM improvement and 1.5 – 2× root mean squared error (RMSE) reduction when applied to GRAPPA and improves robustness to ACS size for various acceleration rates in comparison to other scan‐specific techniques. When applied to advanced reconstruction techniques such as residual‐RAKI, 2D VC‐GRAPPA and LORAKS, SPARK achieves up to 20% RMSE improvement. SPARK with 3D GRAPPA also improves RMSE performance by ~2×, SSIM performance, and perceived image quality without a fully sampled ACS region. Finally, SPARK synergizes with non‐Cartesian, 2D and 3D wave‐encoding imaging by reducing RMSE between 20% and 25% and providing qualitative improvements.

     Conclusion

    SPARK synergizes with physics‐based acquisition and reconstruction techniques to improve accelerated MRI by training scan‐specific models to estimate and correct reconstruction errors in k‐space.

     
    more » « less
  2. ; ; ; ( , Magnetic Resonance in Medicine)
    Purpose

    This work demonstrates a magnetization prepared diffusion‐weighted single‐shot fast spin echo (SS‐FSE) pulse sequence for the application of body imaging to improve robustness to geometric distortion. This work also proposes a scan averaging technique that is superior to magnitude averaging and is not subject to artifacts due to object phase.

     Theory and Methods

    This single‐shot sequence is robust against violation of the Carr‐Purcell‐Meiboom‐Gill (CPMG) condition. This is achieved by dephasing the signal after diffusion weighting and tipping the MG component of the signal onto the longitudinal axis while the non‐MG component is spoiled. The MG signal component is then excited and captured using a traditional SS‐FSE sequence, although the echo needs to be recalled prior to each echo. Extended Parallel Imaging (ExtPI) averaging is used where coil sensitivities from the multiple acquisitions are concatenated into one large parallel imaging (PI) problem. The size of the PI problem is reduced by SVD‐based coil compression which also provides background noise suppression. This sequence and reconstruction are evaluated in simulation, phantom scans, and in vivo abdominal clinical cases.

     Results

    Simulations show that the sequence generates a stable signal throughout the echo train which leads to good image quality. This sequence is inherently low‐SNR, but much of the SNR can be regained through scan averaging and the proposed ExtPI reconstruction. In vivo results show that the proposed method is able to provide diffusion encoded images while mitigating geometric distortion artifacts compared to EPI.

     Conclusion

    This work presents a diffusion‐prepared SS‐FSE sequence that is robust against the violation of the CPMG condition while providing diffusion contrast in clinical cases. Magn Reson Med 79:3032–3044, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

     
    more » « less
  3. ; ; ; ; ; ; ; ( , Medical Physics)
    Abstract Background

    Magnetic resonance imaging (MRI) scans are known to suffer from a variety of acquisition artifacts as well as equipment‐based variations that impact image appearance and segmentation performance. It is still unclear whether a direct relationship exists between magnetic resonance (MR) image quality metrics (IQMs) (e.g., signal‐to‐noise, contrast‐to‐noise) and segmentation accuracy.

     Purpose

    Deep learning (DL) approaches have shown significant promise for automated segmentation of brain tumors on MRI but depend on the quality of input training images. We sought to evaluate the relationship between IQMs of input training images and DL‐based brain tumor segmentation accuracy toward developing more generalizable models for multi‐institutional data.

     Methods

    We trained a 3D DenseNet model on the BraTS 2020 cohorts for segmentation of tumor subregions enhancing tumor (ET), peritumoral edematous, and necrotic and non‐ET on MRI; with performance quantified via a 5‐fold cross‐validated Dice coefficient. MRI scans were evaluated through the open‐source quality control tool MRQy, to yield 13 IQMs per scan. The Pearson correlation coefficient was computed between whole tumor (WT) dice values and IQM measures in the training cohorts to identify quality measures most correlated with segmentation performance. Each selected IQM was used to group MRI scans as “better” quality (BQ) or “worse” quality (WQ), via relative thresholding. Segmentation performance was re‐evaluated for the DenseNet model when (i) training on BQ MRI images with validation on WQ images, as well as (ii) training on WQ images, and validation on BQ images. Trends were further validated on independent test sets derived from the BraTS 2021 training cohorts.

     Results

    For this study, multimodal MRI scans from the BraTS 2020 training cohorts were used to train the segmentation model and validated on independent test sets derived from the BraTS 2021 cohort. Among the selected IQMs, models trained on BQ images based on inhomogeneity measurements (coefficient of variance, coefficient of joint variation, coefficient of variation of the foreground patch) and the models trained on WQ images based on noise measurement peak signal‐to‐noise ratio (SNR) yielded significantly improved tumor segmentation accuracy compared to their inverse models.

     Conclusions

    Our results suggest that a significant correlation may exist between specific MR IQMs and DenseNet‐based brain tumor segmentation performance. The selection of MRI scans for model training based on IQMs may yield more accurate and generalizable models in unseen validation.

     
    more » « less
  4. ; ; ; ; ; ( , Magnetic Resonance in Medicine)
    Purpose

    To develop and evaluate a technique for 3D dynamic MRI of the full vocal tract at high temporal resolution during natural speech.

     Methods

    We demonstrate 2.4 × 2.4 × 5.8 mm3spatial resolution, 61‐ms temporal resolution, and a 200 × 200 × 70 mm3FOV. The proposed method uses 3D gradient‐echo imaging with a custom upper‐airway coil, a minimum‐phase slab excitation, stack‐of‐spirals readout, pseudo golden‐angle view order inkx‐ky, linear Cartesian order alongkz, and spatiotemporal finite difference constrained reconstruction, with 13‐fold acceleration. This technique is evaluated using in vivo vocal tract airway data from 2 healthy subjects acquired at 1.5T scanner, 1 with synchronized audio, with 2 tasks during production of natural speech, and via comparison with interleaved multislice 2D dynamic MRI.

     Results

    This technique captured known dynamics of vocal tract articulators during natural speech tasks including tongue gestures during the production of consonants “s” and “l” and of consonant–vowel syllables, and was additionally consistent with 2D dynamic MRI. Coordination of lingual (tongue) movements for consonants is demonstrated via volume‐of‐interest analysis. Vocal tract area function dynamics revealed critical lingual constriction events along the length of the vocal tract for consonants and vowels.

     Conclusion

    We demonstrate feasibility of 3D dynamic MRI of the full vocal tract, with spatiotemporal resolution adequate to visualize lingual movements for consonants and vocal tact shaping during natural productions of consonant–vowel syllables, without requiring multiple repetitions.

     
    more » « less
  5. ; ; ( , IEEE Transactions on Computational Imaging)
    This work develops a novel set of algorithms, al- ternating Gradient Descent (GD) and minimization for MRI (altGDmin-MRI1 and altGDmin-MRI2), for accelerated dynamic MRI by assuming an approximate low-rank (LR) model on the matrix formed by the vectorized images of the sequence. The LR model itself is well-known in the MRI literature; our contribution is the novel GD-based algorithms which are much faster, memory-efficient, and ‘general’ compared with existing work; and careful use of a 3-level hierarchical LR model. By ‘general’, we mean that, with a single choice of parameters, our method provides accurate reconstructions for multiple acceler- ated dynamic MRI applications, multiple sampling rates and sampling schemes. We show that our methods outperform many of the popular existing approaches while also being faster than all of them, on average. This claim is based on comparisons on 8 different retrospectively undersampled multi-coil dynamic MRI applications, sampled using either 1D Cartesian or 2D pseudo- radial undersampling, at multiple sampling rates. Evaluations on some prospectively undersampled datasets are also provided. Our second contribution is a mini-batch subspace tracking extension that can process new measurements and return reconstructions within a short delay after they arrive. The recovery algorithm itself is also faster than its batch counterpart. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email