2024 ISMRM Honors

Summa Cum Laude


Implicit Neural Representations of GRAPPA Kernels for Rapid Non-Cartesian and Time-Segmented Reconstructions

Daniel Abraham1, Mark Nishimura1, Xiaozhi Cao2, Congyu Liao2, and Kawin Setsompop1,2
1Electrical Engineering, Stanford University, Stanford, CA, United States, 2Radiology, Stanford University, Stanford, CA, United States

Keywords: Image Reconstruction, Image Reconstruction

Motivation: Using non-Cartesian trajectories allows for motion robustness, and a more efficient encoding. However, these non-Cartesian acquisitions necessitate the use of NUFFTs and field correction techniques, leading to costly reconstruction times.

Goal(s): We aim to remove the need for NUFFTs in non-Cartesian MRI, and drastically reduce the computational footprint of field correction.

Approach: Our approach is to correct the raw k-space data of phase due to field imperfections and off-grid sampling using an implicit representation of GRAPPA kernels.

Results: We show an order of magnitude increase in comparison to current standard techniques with near identical reconstructions quality.

Impact: This work aims to significantly reduce the computational requirement for reconstructing non-Cartesian data. This will help with the adoption of long readout non-Cartesian acquisitions, which naturally accelerate MRI exams.


Transfer learning for non-parametric prediction of joint distributions of g-ratios and axon diameters from MRI

Gustavo Chau Loo Kung1,2, Emmanuelle M.M. Weber2, Ankita Batra3, Lijun Ni3, Michael Zeineh2, Juliet Knowles3, and Jennifer A. McNab2
1Bioengineering Department, Stanford University, Stanford, CA, United States, 2Radiology Department, Stanford University, Stanford, CA, United States, 3Neurology Department, Stanford University, Stanford, CA, United States

Keywords: Analysis/Processing, Microstructure, Histology, Diffusion Imaging, g-ratio, axon diameter

Motivation: Machine learning approaches are an alternative to conventional biophysical model fitting used to generate MRI microstructural maps, but the lack of paired MRI-histology data complicates end-to-end training of these models.

Goal(s): Develop a nonparametric deep learning based prediction of joint distributions of g-ratios and axon diameters from multimodal MRI data.

Approach: Histology-based synthetic MRI data was used to pretrain a conditioned normalizing flow model. Transfer learning was then performed on limited paired MRI-histology data.

Results: The joint distribution shows good visual agreement with actual samples and the distances between the marginal probabilities and their respective samples exhibit a Jensen-Shannon distance smaller than 0.22.

Impact: We present an optimized model to obtain non-parametric joint distributions of g-ratios and axon diameters from multimodal MRI from limited experimental data. The approach can easily be adapted to other microstructural modeling tasks.


Robust Multi-Shot Diffusion Weighted Imaging of the Abdomen with Region-Based Shot Rejection

Philip Kenneth Lee1, Xuetong Zhou1,2, and Brian Andrew Hargreaves1,2,3
1Radiology, Stanford University, Stanford, CA, United States, 2Bioengineering, Stanford University, Stanford, CA, United States, 3Electrical Engineering, Stanford University, Stanford, CA, United States

Keywords: Diffusion Reconstruction, Diffusion/other diffusion imaging techniques

Motivation: To improve the motion robustness of multi-shot DWI in the abdomen and reduce signal dropouts and ADC overestimation caused by unresolved shot-to-shot phase.

Goal(s): Demonstrate that region-based weighting of different shots improves diffusion contrast in rapidly moving abdominal organs.

Approach: Shot rejection was evaluated in the pancreas. Multiple shot rejection formulations were tested, and compared using conventional monopolar, and motion-compensated diffusion encodings.

Results: Shot rejection allows conventional monopolar encoding to achieve diffusion weighting and ADCs similar to the motion-compensated encoding in the pancreas. The reconstruction is linear, requires no modifications to the sequence, and is applicable to many encoding trajectories.

Impact: Shot rejection may improve the consistency and robustness of multi-shot abdominal DWI in the clinic, as well as its ability to differentiate pathologies. This will improve repeatability of DWI studies of rapidly moving organs, such as the pancreas and heart.

Magna Cum Laude


Improving Accuracy and Repeatability of Cartilage T2 Mapping in the OAI Dataset through Extended Phase Graph Modeling

Marco Barbieri1, Anthony A Gatti1, and Feliks Kogan1
1Department of Radiology, Stanford University, Stanford, CA, United States

Keywords: Osteoarthritis, Osteoarthritis, Cartilage, MSK, Quantitative Imaging, Data Processing

Motivation: Current methods for T2 fitting in the OAI dataset are based on exponential models, which are inherently sub-optimal as they do not account for stimulated echoes and B1 inhomogeneities. 

Goal(s): To study whether EPG-Model fitting methods improve accuracy and repeatability of T2 mapping in the OAI dataset compared to conventional methods.

Approach: We set up three EPG modelling approaches: nonlinear-least-square, dictionary matching, and deep learning. We used simulations and data from the OAI dataset to evaluate accuracy, repeatability. 

Results: We found that EPG-based methods had higher accuracy and repeatability than exponential-based methods commonly used to compute T2 maps in the OAI dataset. 

Impact: We have demonstrated that EPG-based methods improved accuracy and repeatability of T2 mapping in the OAI dataset over the commonly used mono-exponential fitting methods. This permits more robust analysis of T2 information in the OAI dataset, especially in longitudinal analyses.


Neural Shape Models Meaningfully Localize Features Relevant to Osteoarthritis Disease: Data from the Osteoarthritis Initiative

Anthony A Gatti1, Louis Blankemeier1, Dave Van Veen1, Brian A Hargreaves1, Scott L Delp1, Feliks Kogan1, Garry E Gold1, and Akshay S Chaudhari1
1Stanford University, Stanford, CA, United States

Keywords: Osteoarthritis, MSK, shape model, MOAKS, osteophytes

Motivation: Osteoarthritis is a whole joint disease that requires quantification, localization, and visualization of disease related features of bones and cartilage.

Goal(s): To develop a novel neural shape model (NSM) that can encode and reconstruct bone and cartilage shape, while quantifying localized features of OA.

Approach: We trained a NSM on 6,325 knees and compared its reconstructions to a conventional statistical shape model and its ability to predict localized disease to a convolutional neural network.

Results: The NSM reconstructed tissues with cartilage thickness correlations >0.993. NSM representations accurately diagnosed OA and predicted localized severity of osteophytes and cartilage defects better than a CNN.

Impact: Our NSM can reconstruct whole bone and cartilage morphology, while encoding localized pathology specific information. Research use of the NSM can unlock novel insights into OA pathophysiology. Clinical deployment would enable automated insights into whole joint health.


Deep Learning-based Disambiguation for Multiple AD Radiotracers using PET/MRI

Ashwin Kumar1, Donghoon Kim1, Elizabeth Mormino2, Akshay Chaudhari1, Christina Young2, Kevin Chen3, Mehdi Khalighi1, and Greg Zaharchuk1
1Radiology, Stanford University, Stanford, CA, United States, 2Neurology, Stanford University, Stanford, CA, United States, 3Biomedical Engineering, National Taiwan University, Taipei, Taiwan

Keywords: PET/MR, PET/MR

Motivation: AD patients must undergo repeated visits for amyloid and tau radiotracer imaging, leading to high costs and dose concerns due to PET's inability to simultaneously acquire multiple radiotracers during a single session.

Goal(s): Using PET/MRI scans, we used deep learning to create separate amyloid and tau PET images from a simulated combined dual-tracer image.

Approach: We simulated a combined amyloid-tau image by blending co-registered list-mode data and employed a 2.5D U-Net architecture for effective separation.

Results: Mixed-dose models, incorporating physics-inspired data augmentation and MR information, exhibited enhanced anatomical preservation and reduced variability in quantitative metrics.

Impact: The demonstrated separation of a simulated combined amyloid and tau PET/MRI study into its individual components using DL may allow for simultaneous injection of multiple radiotracers in a single acquisition, streamlining the imaging process for AD patients.


Association of Patella Bone Shape and MR-Diagnosed Patellar Tendinopathy with Patellar Cartilage T2/T1ρ in Elite Basketball Players

Andrew M. Schmidt1, Elka B Rubin1, Mackenzie Little1,2, Madison George3, Hayden Zheng4, Katherine Young1, Arjun D. Desai1,5,6, Feliks Kogan1, Sharmila Majumdar7, Hollis G Potter8, Garry E. Gold1,3, and Anthony A. Gatti1
1Radiology, Stanford University, Stanford, CA, United States, 2University of Sydney, Sydney, Australia, 3Bioengineering, Stanford University, Stanford, CA, United States, 4Human Biology, Stanford University, Stanford, CA, United States, 5Electrical Engineering, Stanford University, Stanford, CA, United States, 6Computer Science, Stanford University, Stanford, CA, United States, 7Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, United States, 8Hospital for Special Surgery, New York, NY, United States

Keywords: Tendon/Ligament, Tendon/Ligament, Bone shape, T2 mapping, T1ρ mapping

Motivation: Patellar tendinopathy (PT) is a common injury in basketball that can lead to early retirement. Basketball influences bone shape and cartilage composition, yet the association between these factors and PT is unknown.

Goal(s): We examined the association between PT, bone shape, and patellar cartilage composition in collegiate basketball players.

Approach: We developed a measure of bone shape indicative of PT and investigated whether PT-associated bone shape is associated with patellar cartilage T2/T.

Results: We effectively separated grades of PT using bone shape and identified bone shape features associated with PT. We found patellar cartilage composition is independent of PT and bone shape.

Impact: We developed a measure to identify varying grades of PT based on bone shape in collegiate basketball players. Future work will determine the association of our PT-bone shape score with MR-identifiable measures to identify athlete specific PT risk factors.