Deep Learning-Based Parameter Mapping With Uncertainty Estimation For Fat Quantification Using Accelerated Free-Breathing Radial MRI

Shih, Shu‐Fu; Kafalı, Sevgi Gökçe; Armstrong, Tess; Zhong, Xiaodong; Calkins, Kara L.; Wu, Holden H.

doi:10.1109/isbi48211.2021.9433938

Cited by 8 publications

(11 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The "black box" nature of DL-based methods for MRI is an important concern and potential barrier to clinical translation. Uncertainty estimation in DL networks [35][36][37][38] presents a promising approach to provide context and assess confidence in DL outputs for clinical applications that demand a high level of numerical accuracy, including the use of quantitative maps for diagnostic decisions. In this study, we showed that with calibration, UP-Net uncertainty scores predicted quantification errors in a separate testing dataset.…”

Section: Discussionmentioning

confidence: 99%

“…Second, our proposed network has built‐in uncertainty estimation that generates pixel‐wise uncertainty maps for different quantitative parameters. Uncertainty estimation to assess the confidence levels in DL‐based MRI reconstruction and quantitative parameter mapping results is a nascent direction 35–38 . We specifically investigated the application of uncertainty estimation in DL‐based PDFF and R 2 * quantification and demonstrated that a calibration method for the UP‐Net uncertainty scores can be used to predict absolute liver PDFF and R 2 * quantification errors in UP‐Net parameter maps to within 1% and 3 s −1 , respectively, compared to actual errors with respect to reference methods.…”

Section: Discussionmentioning

confidence: 99%

“…A recent work showed promising results of using uncertainty estimation for quantitative MRI PDFF maps obtained from DL. 38 The study indicated that the uncertainty scores were related to the noise levels in the input data. However, the relationship between the uncertainty scores and quantification accuracy was not established.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Uncertainty‐aware physics‐driven deep learning network for free‐breathing liver fat and R₂* quantification using self‐gated stack‐of‐radial MRI

Shih

Kafalı

Calkins

et al. 2022

Magnetic Resonance in Med

Self Cite

View full text Add to dashboard Cite

Purpose To develop a deep learning‐based method for rapid liver proton‐density fat fraction (PDFF) and R2* quantification with built‐in uncertainty estimation using self‐gated free‐breathing stack‐of‐radial MRI. Methods This work developed an uncertainty‐aware physics‐driven deep learning network (UP‐Net) to (1) suppress radial streaking artifacts because of undersampling after self‐gating, (2) calculate accurate quantitative maps, and (3) provide pixel‐wise uncertainty maps. UP‐Net incorporated a phase augmentation strategy, generative adversarial network architecture, and an MRI physics loss term based on a fat–water and R2* signal model. UP‐Net was trained and tested using free‐breathing multi‐echo stack‐of‐radial MRI data from 105 subjects. UP‐Net uncertainty scores were calibrated in a validation dataset and used to predict quantification errors for liver PDFF and R2* in a testing dataset. Results Compared with images reconstructed using compressed sensing (CS), UP‐Net achieved structural similarity index >0.87 and normalized root mean squared error <0.18. Compared with reference quantitative maps generated using CS and graph‐cut (GC) algorithms, UP‐Net achieved low mean differences (MD) for liver PDFF (−0.36%) and R2* (−0.37 s−1). Compared with breath‐holding Cartesian MRI results, UP‐Net achieved low MD for liver PDFF (0.53%) and R2* (6.75 s−1). UP‐Net uncertainty scores predicted absolute liver PDFF and R2* errors with low MD of 0.27% and 0.12 s−1 compared to CS + GC results. The computational time for UP‐Net was 79 ms/slice, whereas CS + GC required 3.2 min/slice. Conclusion UP‐Net rapidly calculates accurate liver PDFF and R2* maps from self‐gated free‐breathing stack‐of‐radial MRI. The pixel‐wise uncertainty maps from UP‐Net predict quantification errors in the liver.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Uncertainty‐aware physics‐driven deep learning network for free‐breathing liver fat and R₂* quantification using self‐gated stack‐of‐radial MRI

Shih

Kafalı

Calkins

et al. 2022

Magnetic Resonance in Med

Self Cite

View full text Add to dashboard Cite

show abstract

“…181,215,222,223 However, in applications where the phase information is important for the subsequent parameter estimation step, complex neural networks have been developed that exploit the correlation between the real and imaginary parts of a complex image by processing the real and imaginary images with complex convolutions inspired by the multiplication of complex numbers. 219 Various architectures have been investigated for the neural network, including U-NET, 215,224 fully connected convolutional neural network, 223,225 multiscale residual network, 222 deep cascade of residual dense network, 221 and deep complex residual network. 219…”

Section: View-sharing Reconstructionmentioning

confidence: 99%

Systematic review of reconstruction techniques for accelerated quantitative MRI

Shafieizargar

Byanju

Sijbers

et al. 2023

Magnetic Resonance in Med

View full text Add to dashboard Cite

To systematically review the techniques that address undersampling artifacts in accelerated quantitative magnetic resonance imaging (qMRI). Methods: A literature search was conducted using the Embase, Medline, Web of Science Core Collection, Coherence Central Register of Controlled Trials, and Google Scholar databases for studies, published before July 2022 proposing reconstruction techniques for accelerated qMRI. Studies are reviewed according to inclusion criteria, and included studies are categorized based on the methodology used. Results: A total of 292 studies included in the review are categorized. A technical overview of each category is provided, and the categories are described in a unified mathematical framework. The distribution of the reviewed studies over time, application domain, and parameters of interest is illustrated. Conclusion: An increasing trend in the number of articles that propose new techniques for accelerated qMRI reconstruction indicates the importance of acceleration in qMRI. The techniques are mostly validated for relaxometry parameters and brain scans. The categories of techniques are compared based on theoretical grounds, highlighting existing trends and potential gaps in the field.

show abstract

“…In 29 , authors showed that using complex data outperformed real data on its own for phase-based applications and reconstructions of 2D data for three different datasets. The authors in 30 propose using CNNs for parameter mapping and uncertainty estimation for fat quantification. Another recent effort for fat-water separation in 2D data that simultaneously estimates R2* and field decay has been proposed by 31 .…”

Section: /12mentioning

confidence: 99%

Swap-Free Fat-Water Separation in Dixon MRI using Conditional Generative Adversarial Networks

Basty,

Thanaj,

Cule

et al. 2021

Preprint

View full text Add to dashboard Cite

Dixon MRI is widely used for body composition studies. Current processing methods associated with large whole-body volumes are time intensive and prone to artifacts during fat-water separation performed on the scanner, making the data difficult to analyse. The most common artifact are fat-water swaps, where the labels are inverted at the voxel level. It is common for researchers to discard swapped data (generally around 10%), which can be wasteful and lead to unintended biases. The UK Biobank is acquiring Dixon MRI for over 100,000 participants, and thousands of swaps will occur. If those go undetected, errors will propagate into processes such as abdominal organ segmentation and dilute the results in population-based analyses. There is a clear need for a fast and robust method to accurately separate fat and water channels. In this work we propose such a method based on style transfer using a conditional generative adversarial network. We also introduce a new Dixon loss function for the generator model. Using data from the UK Biobank Dixon MRI, our model is able to predict highly accurate fat and water channels that are free from artifacts. We show that the model separates fat and water channels using either single input (in-phase) or dual input (in-phase and opposed-phase), with the latter producing improved results. Our proposed method enables faster and more accurate downstream analysis of body composition from Dixon MRI in population studies by eliminating the need for visual inspection or discarding data due to fat-water swaps.

show abstract

Deep Learning-Based Parameter Mapping With Uncertainty Estimation For Fat Quantification Using Accelerated Free-Breathing Radial MRI

Cited by 8 publications

References 14 publications

Uncertainty‐aware physics‐driven deep learning network for free‐breathing liver fat and R₂* quantification using self‐gated stack‐of‐radial MRI

Uncertainty‐aware physics‐driven deep learning network for free‐breathing liver fat and R₂* quantification using self‐gated stack‐of‐radial MRI

Systematic review of reconstruction techniques for accelerated quantitative MRI

Swap-Free Fat-Water Separation in Dixon MRI using Conditional Generative Adversarial Networks

Contact Info

Product

Resources

About

Deep Learning-Based Parameter Mapping With Uncertainty Estimation For Fat Quantification Using Accelerated Free-Breathing Radial MRI

Cited by 8 publications

References 14 publications

Uncertainty‐aware physics‐driven deep learning network for free‐breathing liver fat and R2* quantification using self‐gated stack‐of‐radial MRI

Uncertainty‐aware physics‐driven deep learning network for free‐breathing liver fat and R2* quantification using self‐gated stack‐of‐radial MRI

Systematic review of reconstruction techniques for accelerated quantitative MRI

Swap-Free Fat-Water Separation in Dixon MRI using Conditional Generative Adversarial Networks

Contact Info

Product

Resources

About

Uncertainty‐aware physics‐driven deep learning network for free‐breathing liver fat and R₂* quantification using self‐gated stack‐of‐radial MRI

Uncertainty‐aware physics‐driven deep learning network for free‐breathing liver fat and R₂* quantification using self‐gated stack‐of‐radial MRI