fastMRI+, Clinical pathology annotations for knee and brain fully sampled magnetic resonance imaging data

Zhao, Ruiyang; Zhang, Yuxin; Stewart, Russell J.; Dixon, Angela; Knoll, Florian; Huang, Zhengnan; Lui, Yvonne W.; Hansen, Michael S.; Lungren, Matthew P.

doi:10.1038/s41597-022-01255-z

Cited by 29 publications

(11 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We investigated adversarial noise and rotation attacks as described in Section 2 on the fastMRI knee dataset [19]. We explored attacking the entire centercropped image, as well as targeted attacks on diagnostically relevant regions, which we obtained from the pathology annotations in the fastMRI+ dataset [20]. We evaluated the adversarial robustness for the E2E-VN [16] as well as the UNet [7] and for acceleration factors of 4× and 8×.…”

Section: Experiments and Resultsmentioning

confidence: 99%

“…Here, k is the vector of all k-space data k i from all coils i, η defines the L 2adversarial noise level relative to every coil measurement k i , and S is a binary region selection mask to limit the attack to diagnostically relevant regions. Specifically, we use the bounding boxes indicating pathologies provided by the fastMRI+ dataset [20] to define S. In contrast to [6,18], we restrict the noise-vector z i for every coil individually instead of the complete measurement, thereby allowing for differences in the coil-sensitivities.…”

Section: Investigated Reconstruction Methodsmentioning

confidence: 99%

“…In this paper, we aim to shed further light on this question by analysing adversarial robustness on the commonly investigated fastMRI knee dataset [19,20]. In contrast to prior work we show that when attacking state-of-the-art MRI re-construction algorithms in a way targeting regions with diagnosed pathologies directly, very small levels of adversarial noise on the k-space (of comparable magnitude to the thermal noise always present in MR acquisition) can substantially degrade those regions and in some instances alter diagnostically relevant features.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Adversarial Robustness of MR Image Reconstruction under Realistic Perturbations

Morshuis¹,

Gatidis²,

Hein³

et al. 2022

Preprint

View full text Add to dashboard Cite

Deep Learning (DL) methods have shown promising results for solving ill-posed inverse problems such as MR image reconstruction from undersampled k-space data. However, these approaches currently have no guarantees for reconstruction quality and the reliability of such algorithms is only poorly understood. Adversarial attacks offer a valuable tool to understand possible failure modes and worst case performance of DL-based reconstruction algorithms. In this paper we describe adversarial attacks on multi-coil k-space measurements and evaluate them on the recently proposed E2E-VarNet and a simpler UNet-based model. In contrast to prior work, the attacks are targeted to specifically alter diagnostically relevant regions. Using two realistic attack models (adversarial k-space noise and adversarial rotations) we are able to show that current state-of-the-art DL-based reconstruction algorithms are indeed sensitive to such perturbations to a degree where relevant diagnostic information may be lost. Surprisingly, in our experiments the UNet and the more sophisticated E2E-VarNet were similarly sensitive to such attacks. Our findings add further to the evidence that caution must be exercised as DL-based methods move closer to clinical practice.

show abstract

Section: Experiments and Resultsmentioning

confidence: 99%

Section: Investigated Reconstruction Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Adversarial Robustness of MR Image Reconstruction under Realistic Perturbations

Morshuis¹,

Gatidis²,

Hein³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Most prior DL segmentation studies report model accuracies based on the annotations of a singular reader 37,38 . However, in most practical applications, different readers perform segmentation across different studies 39,40 .…”

Section: Discussionmentioning

confidence: 99%

“…Most prior DL segmentation studies report model accuracies based on the annotations of a singular reader. 37,38 However, in most practical applications, different readers perform segmentation across different studies. 39,40 In our generalizability analysis, we depicted the performance of segmentation algorithms across studies that used different readers as annotators of the ground truth cartilage surfaces.…”

Section: Discussionmentioning

confidence: 99%

Generalizability of Deep Learning Segmentation Algorithms for Automated Assessment of Cartilage Morphology and MRI Relaxometry

Schmidt

Desai

Watkins

et al. 2022

Magnetic Resonance Imaging

View full text Add to dashboard Cite

Background: Deep learning (DL)-based automatic segmentation models can expedite manual segmentation yet require resource-intensive fine-tuning before deployment on new datasets. The generalizability of DL methods to new datasets without fine-tuning is not well characterized. Purpose: Evaluate the generalizability of DL-based models by deploying pretrained models on independent datasets varying by MR scanner, acquisition parameters, and subject population. Study Type: Retrospective based on prospectively acquired data. Population: Overall test dataset: 59 subjects (26 females); Study 1: 5 healthy subjects (zero females), Study 2: 8 healthy subjects (eight females), Study 3: 10 subjects with osteoarthritis (eight females), Study 4: 36 subjects with various knee pathology (10 females). Field Strength/Sequence: A 3-T, quantitative double-echo steady state (qDESS). Assessment: Four annotators manually segmented knee cartilage. Each reader segmented one of four qDESS datasets in the test dataset. Two DL models, one trained on qDESS data and another on Osteoarthritis Initiative (OAI)-DESS data, were assessed. Manual and automatic segmentations were compared by quantifying variations in segmentation accuracy, volume, and T2 relaxation times for superficial and deep cartilage. Statistical Tests: Dice similarity coefficient (DSC) for segmentation accuracy. Lin's concordance correlation coefficient (CCC), Wilcoxon rank-sum tests, root-mean-squared error-coefficient-of-variation to quantify manual vs. automatic T2 and volume variations. Bland-Altman plots for manual vs. automatic T2 agreement. A P value < 0.05 was considered statistically significant. Results: DSCs for the qDESS-trained model, 0.79-0.93, were higher than those for the OAI-DESS-trained model, 0.59-0.79. T2 and volume CCCs for the qDESS-trained model, 0.75-0.98 and 0.47-0.95, were higher than respective CCCs for the OAI-DESS-trained model, 0.35-0.90 and 0.13-0.84. Bland-Altman 95% limits of agreement for superficial and deep cartilage T2 were lower for the qDESS-trained model, AE2.4 msec and AE4.0 msec, than the OAI-DESS-trained model, AE4.4 msec and AE5.2 msec. Data Conclusion:The qDESS-trained model may generalize well to independent qDESS datasets regardless of MR scanner, acquisition parameters, and subject population.

show abstract

Accelerated Musculoskeletal Magnetic Resonance Imaging

Yoon,

Gold,

Chaudhari

2023

Magnetic Resonance Imaging

View full text Add to dashboard Cite

With a substantial growth in the use of musculoskeletal MRI, there has been a growing need to improve MRI workflow, and faster imaging has been suggested as one of the solutions for a more efficient examination process. Consequently, there have been considerable advances in accelerated MRI scanning methods. This article aims to review the basic principles and applications of accelerated musculoskeletal MRI techniques including widely used conventional acceleration methods, more advanced deep learning‐based techniques, and new approaches to reduce scan time. Specifically, conventional accelerated MRI techniques, including parallel imaging, compressed sensing, and simultaneous multislice imaging, and deep learning‐based accelerated MRI techniques, including undersampled MR image reconstruction, super‐resolution imaging, artifact correction, and generation of unacquired contrast images, are discussed. Finally, new approaches to reduce scan time, including synthetic MRI, novel sequences, and new coil setups and designs, are also reviewed. We believe that a deep understanding of these fast MRI techniques and proper use of combined acceleration methods will synergistically improve scan time and MRI workflow in daily practice.Evidence Level3Technical EfficacyStage 1

show abstract

fastMRI+, Clinical pathology annotations for knee and brain fully sampled magnetic resonance imaging data

Cited by 29 publications

References 17 publications

Adversarial Robustness of MR Image Reconstruction under Realistic Perturbations

Adversarial Robustness of MR Image Reconstruction under Realistic Perturbations

Generalizability of Deep Learning Segmentation Algorithms for Automated Assessment of Cartilage Morphology and MRI Relaxometry

Accelerated Musculoskeletal Magnetic Resonance Imaging

Contact Info

Product

Resources

About