Using Deep Learning to Accelerate Knee MRI at 3 T: Results of an Interchangeability Study

Recht, Michael P.; Žbontar, Jure; Sodickson, Daniel K.; Knoll, Florian; Yakubova, Nafissa; Sriram, Anuroop; Murrell, Tullie; Defazio, Aaron; Rabbat, Michael; Rybak, Leon; Kline, Mitchell; Ciavarra, Gina; Alaia, Erin F.; Samim, Mohammad; Walter, William R.; Lin, Dana J; Lui, Yvonne W.; Muckley, Matthew J.; Huang, Zhengnan; Johnson, Patricia M.; Stern, Ruben; Zitnick, C. Lawrence

doi:10.2214/ajr.20.23313

Cited by 128 publications

(96 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…DL-based reconstructions can be used for the acceleration of MRI sequences via the increased undersampling of the acquired data, e.g., via a variational neural network [16]. The clinical applicability of this technique was already demonstrated in knee MRI examinations by Recht et al [13]. However, in this study, the authors used retrospectively undersampled data.…”

Section: Discussionmentioning

confidence: 97%

See 1 more Smart Citation

Accelerated T2-Weighted TSE Imaging of the Prostate Using Deep Learning Image Reconstruction: A Prospective Comparison with Standard T2-Weighted TSE Imaging

Gassenmaier

Afat

Nickel

et al. 2021

Cancers

View full text Add to dashboard Cite

Multiparametric MRI (mpMRI) of the prostate has become the standard of care in prostate cancer evaluation. Recently, deep learning image reconstruction (DLR) methods have been introduced with promising results regarding scan acceleration. Therefore, the aim of this study was to investigate the impact of deep learning image reconstruction (DLR) in a shortened acquisition process of T2-weighted TSE imaging, regarding the image quality and diagnostic confidence, as well as PI-RADS and T2 scoring, as compared to standard T2 TSE imaging. Sixty patients undergoing 3T mpMRI for the evaluation of prostate cancer were prospectively enrolled in this institutional review board-approved study between October 2020 and March 2021. After the acquisition of standard T2 TSE imaging (T2S), the novel T2 TSE sequence with DLR (T2DLR) was applied in three planes. Overall, the acquisition time for T2S resulted in 10:21 min versus 3:50 min for T2DLR. The image evaluation was performed by two radiologists independently using a Likert scale ranging from 1–4 (4 best) applying the following criteria: noise levels, artifacts, overall image quality, diagnostic confidence, and lesion conspicuity. Additionally, T2 and PI-RADS scoring were performed. The mean patient age was 69 ± 9 years (range, 49–85 years). The noise levels and the extent of the artifacts were evaluated to be significantly improved in T2DLR versus T2S by both readers (p < 0.05). Overall image quality was also evaluated to be superior in T2DLR versus T2S in all three acquisition planes (p = 0.005–<0.001). Both readers evaluated the item lesion conspicuity to be superior in T2DLR with a median of 4 versus a median of 3 in T2S (p = 0.001 and <0.001, respectively). T2-weighted TSE imaging of the prostate in three planes with an acquisition time reduction of more than 60% including DLR is feasible with a significant improvement of image quality.

show abstract

Section: Discussionmentioning

confidence: 97%

“…In previous studies, it could be shown that DL, e.g., via variational networks, is able to significantly accelerate MRI protocols of the abdomen, knee or of the pituitary [12][13][14][15]. However, literature regarding the clinical impact of DL strategies in accelerated MRI protocols is still sparse.…”

Section: Introductionmentioning

confidence: 99%

Accelerated T2-Weighted TSE Imaging of the Prostate Using Deep Learning Image Reconstruction: A Prospective Comparison with Standard T2-Weighted TSE Imaging

Gassenmaier

Afat

Nickel

et al. 2021

Cancers

View full text Add to dashboard Cite

show abstract

“…For the standard multicoil tracks in the 2019 challenge, we observed that although there were many high-quality submissions at 4X, all of the submissions began missing pathology at 8X acceleration [ 27 ]. Since this time, 4X machine learning methods have been validated for clinical interchangeability [ 48 ]. This suggests that the current upper limit of 2D machine learning image reconstruction performance remains between 4-fold and 8-fold acceleration rates.…”

Section: Methodsmentioning

confidence: 99%

Results of the 2020 fastMRI Challenge for Machine Learning MR Image Reconstruction

Muckley

Riemenschneider

Radmanesh

et al. 2021

IEEE Trans. Med. Imaging

Self Cite

194

108

View full text Add to dashboard Cite

Accelerating MRI scans is one of the principal outstanding problems in the MRI research community. Towards this goal, we hosted the second fastMRI competition targeted towards reconstructing MR images with subsampled k-space data. We provided participants with data from 7,299 clinical brain scans (de-identified via a HIPAA-compliant procedure by NYU Langone Health), holding back the fully-sampled data from 894 of these scans for challenge evaluation purposes. In contrast to the 2019 challenge, we focused our radiologist evaluations on pathological assessment in brain images. We also debuted a new Transfer track that required participants to submit models evaluated on MRI scanners from outside the training set. We received 19 submissions from eight different groups. Results showed one team scoring best in both SSIM scores and qualitative radiologist evaluations. We also performed analysis on alternative metrics to mitigate the effects of background noise and collected feedback from the participants to inform future challenges. Lastly, we identify common failure modes across the submissions, highlighting areas of need for future research in the MRI reconstruction community.

show abstract

“…When compared to radiologists, their model achieved comparable results to radiologists in diagnosing abnormalities but underperformed in detecting ACL and meniscus tears, specifically in the sensitivity metric. Another study provided further evidence for the ability of deep learning models to perform diagnosis tasks at a level of accuracy similar to radiologists [4]. They developed a two-stream convolutional neural network model capable of diagnosing cartilage lesions, including cartilage softening, fibrillation, and fissuring.…”

Section: Introductionmentioning

confidence: 99%

Incorporating the Laplacian Filter with a Three-Stream Multi-Channel Convolutional Neural Network for Improved Abnormality Detection in Knee MRIs

Kumar¹,

Rm²

2021

Preprint

View full text Add to dashboard Cite

Despite ACL and meniscus tears being among the most common movement induced injuries, they are often the most difficult to diagnose due to the variable severity with which these tears occur. Typically, magnetic resonance imaging (MRI) scans are used for diagnosing ligament tears, but performing and analyzing these scans is time consuming and expensive due to the necessitation of a radiologist or professional orthopedic specialist. Consequently, we developed a custom three-stream convolutional neural network (CNN) architecture that contains multiple channels to automate the diagnosis of ACL and meniscus tears from MRI scans. Our algorithm utilizes the sagittal, coronal, and axial slices to maximize feature extraction. Furthermore, we apply the Laplace Operator on the MRI scan images to evaluate and compare its propensity in different medical imaging modalities. The algorithm attained an accuracy of 92.80\%, significantly higher than that of orthopedic diagnosis accuracy. Our results point towards the feasibility of shallow, multi-channel CNNs and the ability of the Laplace Operator to improve performance metrics for MRI scan diagnosis.

show abstract

Using Deep Learning to Accelerate Knee MRI at 3 T: Results of an Interchangeability Study

Cited by 128 publications

References 29 publications

Accelerated T2-Weighted TSE Imaging of the Prostate Using Deep Learning Image Reconstruction: A Prospective Comparison with Standard T2-Weighted TSE Imaging

Accelerated T2-Weighted TSE Imaging of the Prostate Using Deep Learning Image Reconstruction: A Prospective Comparison with Standard T2-Weighted TSE Imaging

Results of the 2020 fastMRI Challenge for Machine Learning MR Image Reconstruction

Incorporating the Laplacian Filter with a Three-Stream Multi-Channel Convolutional Neural Network for Improved Abnormality Detection in Knee MRIs

Contact Info

Product

Resources

About