View Presentation Video

Xitong Wang¹, Junyu Wang¹, Shen Zhao¹, Ruixi Zhou², Yang Yang³, and Michael Salerno^1
1Stanford Univeristy, Stanford, CA, United States, ²Beijing University of Posts and Telecommunications, Beijing, China, ³Department of Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States

Keywords: Machine Learning/Artificial Intelligence, Cardiovascular

We developed a rapid 3D self-gated cardiac cine technique, using a variable-density undersampled randomized stack of spiral gradient echo sequence to perform cine evaluation of the whole left ventricle. Our proposed slice-by-slice deep learning-based imaging reconstruction technique for self-gated free-breathing 3D stack of spiral cardiac cine imaging can produce cine images with high temporal (40 ms) and spatial resolution (1.25x1.25x8 mm) within a 20s acquisition time with <1s deep learning inference time.

View Presentation Video

Mahmoud Mostapha¹, Gregor Koerzdoerfer², Boris Mailhe¹, Esther Raithel², Inge M. Brinkmann², Nirmal Janardhanan¹, Maria Ringholz², Mariappan S. Nadar¹, and Jan Fritz^3
1Siemens Healthineers, Princeton, NJ, United States, ²Siemens Healthcare GmbH, Erlangen, Germany, ³Department of Radiology, NYU Grossman School of Medicine, New York, NY, United States

Keywords: Machine Learning/Artificial Intelligence, MSK

Combining parallel imaging (PI) and simultaneous multislice (SMS) acceleration realized a clinical 4-fold accelerated 2D TSE MRI of the knee. However, 8-fold acceleration with conventional reconstruction methods suffers from significant image quality degradation. We propose a complete DL approach for combined slice separation and k-space-to-image reconstruction of SMS-PI-accelerated knee MRI with tunable denoising strength and super-resolution image enhancement. The proposed methods enable artifact-free image reconstruction of 8-fold accelerated 2D TSE MR images in multiple planes and with multiple image contrasts. Clinical evaluations suggest equivalence of image quality and detection rates of 8-fold S2P4 DL reconstructions compared to the reference standard.

View Presentation Video

Linfang Xiao¹, Yilong Liu², Zhizun Zhang¹, Ruixing Zhu¹, Weijun Chen¹, Zeyao Ma¹, Congying Mao¹, and Ke Wang^1
1Hangzhou Weiying Medical Technology Co., Ltd, Hangzhou, China, ²Guangdong-Hongkong-Macau Institute of CNS Regeneration, Key Laboratory of CNS Regeneration (Ministry of Education), Jinan University, Guangzhou, China

Keywords: Machine Learning/Artificial Intelligence, Image Reconstruction

MR Image reconstruction of uniformly undersampled data often relies on prior information estimated from additional calibration data, leading to compromised acquisition efficiency and flexibility. Here, we propose a joint multi-slice deep learning strategy for MR image reconstruction from uniformly undersampled data with complementary undersampling across adjacent slices. Specifically, we design a slice fusion block to fully exploit the structural and phase similarity in adjacent slices and a slice shift block to further suppress the aliasing artifacts introduced by uniform undersampling. Consequently, the proposed strategy enables accurate MR image reconstruction for both image magnitude and phase without additional calibration information.

View Presentation Video

Tobias Wech¹, Oliver Schad^1,2, and Jonas Kleineisel^1
1Department of Diagnostic and Interventional Radiology, University Hospital Würzburg, Würzburg, Germany, ²Experimental Physics 5, University of Würzburg, Würzburg, Germany

Keywords: Machine Learning/Artificial Intelligence, Image Reconstruction

We trained a score-based generative diffusion model with cardiac MR images, which allows generating new, randomized instances of the given data distribution. By conditioning each step of the underlying reverse time stochastic differential equation with a physics-informed data consistency step, undersampled MR data can be reconstructed.  An initial estimation of the complex phase, which slowly transfers into the actual phase of the image, allows to train the diffusion model with magnitude data only. The approach was evaluated in fast spiral dynamic cardiac MRI at 1.5T, where it provided superior SNR-levels compared to alternative acceleration techniques.
 

View Presentation Video

Keywords: Machine Learning/Artificial Intelligence, Machine Learning/Artificial Intelligence, SPIRiT, Diffusion Models, Vessel Wall Imaging

Diffusion model is the most advanced method in image generation and has been successfully applied to MRI reconstruction. However, the existing methods do not consider the characteristics of multi-coil acquisition of MRI data. Therefore, we give a new diffusion model, called SPIRiT-Diffusion, based on the SPIRiT iterative reconstruction algorithm. Specifically, SPIRiT-Diffusion characterizes the prior distribution of coil-by-coil images by score matching and characterizes the k-space redundant prior between coils based on self-consistency. With sufficient prior constraint utilized, we achieve superior reconstruction results on the joint Intracranial and Carotid Vessel Wall imaging dataset.

View Presentation Video

Keywords: Machine Learning/Artificial Intelligence, Image Reconstruction, MRI reconstruction, self-supervised, deep learning

 Although Deep learning (DL) techniques are powerful for MRI reconstruction, their development is hindered by the need for large training datasets. We propose a self-supervised method for training DL reconstruction networks using only limited-resolution data, acquired in k-space “bands”, which are generally easier to acquire than variable-density full-resolution data. Although the network is trained using low-resolution data, during inference it can reconstruct high-resolution images. Comprehensive experiments demonstrate that k-band is robust to various acceleration factors, outperforms two other methods trained on low-resolution data, and obtains comparable performance with SSDU and MoDL, while also reducing the need for full-resolution data.

View Presentation Video

Omer Burak Demirel^1,2, Chi Zhang^1,2, Burhaneddin Yaman^1,2, Toygan Kilic^1,2, Steen Moeller², Chetan Shenoy³, Sebastian Weingärtner⁴, Tim Leiner⁵, and Mehmet Akçakaya^1,2
1Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN, United States, ²Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States, ³Department of Medicine (Cardiology), University of Minnesota, Minneapolis, MN, United States, ⁴Delft University of Technology, Delft, Netherlands, ⁵Department of Radiology, Mayo Clinical, Rochester, MN, United States

Keywords: Machine Learning/Artificial Intelligence, Machine Learning/Artificial Intelligence

Myocardial perfusion CMR is used to functionally assess coronary artery disease. However, its resolution and coverage remain limited and require rapid imaging .At high accelerations for whole-heart coverage and high spatio-temporal resolution, conventional reconstructions suffer from noise and aliasing artifacts. Physics-guided deep learning (PG-DL) reconstruction has shown improved image quality in fast MRI, but its application to perfusion CMR has been limited due to substantial differences in breathing and contrast uptakes among subjects. In this work, we tackle these challenges by adopting subject-specific self-supervised PG-DL that does not require a training database for simultaneous multi-slice accelerated myocardial perfusion CMR.

View Presentation Video

Jianping Xu¹, Tao Zu¹, Yi-Cheng Hsu², Yi Sun², and Yi Zhang^1
1Key Laboratory for Biomedical Engineering of Ministry of Education, Department of Biomedical Engineering, College of Biomedical Engineering & Instrument Science, Zhejiang University, Hangzhou, Zhejiang, China, Hangzhou, China, ²MR Collaboration, Siemens Healthcare Ltd., Shanghai, China, Shanghai, China

Keywords: Machine Learning/Artificial Intelligence, Image Reconstruction

The widespread clinical adoption of chemical exchange saturation transfer (CEST) imaging has been hampered by its prolonged scan time, due to multiple data acquisitions over the varying saturation offset frequencies. In this work, we utilize the artifact suppression algorithm and propose an effective deep reconstruction framework with self-calibration mechanisms (DEISM). The DEISM method was validated on brain tumor patients at 3T. In conjunction with deep-learning multi-coil image reconstruction and data-driven artifact suppression mechanisms, DEISM can provide reliable reconstructions of highly accelerated CEST data, yielding superior performance compared to state-of-the-art methods.

View Presentation Video

Wei Jiang¹, Yang Gao¹, and Hongfu Sun^1
1University of Queensland, Brisbane, Australia

Keywords: Machine Learning/Artificial Intelligence, Image Reconstruction, diffusion models, compressed sensing, MRI

Recovering MR images from partially acquired k-space data can reduce scan time, leading to significant applications. Deep learning approaches have been proposed to train neural networks using under-sampled and full-sampled image pairs in a supervised manner. However, these supervised models usually have poor generalizability when deployed to acquisitions different from the training datasets. Here, we propose an unsupervised Denoising Diffusion Probabilistic Model (DDPM) capable of not only generating random high-fidelity MR images but also reconstructing images corresponding to k-space data from arbitrary acquisition patterns.

View Presentation Video

Chang Gao^1,2, Zhengyang Ming^1,2, Kim-Lien Nguyen^1,2, Xiaodong Zhong³, and John Paul Finn^1,2
1Department of Radiological Sciences, University of California Los Angeles, Los Angeles, CA, United States, ²Department of Physics and Biology in Medicine, University of California Los Angeles, Los Angeles, CA, United States, ³MR R&D Collaborations, Siemens Medical Solutions USA, Inc., Los Angeles, CA, United States

Keywords: Machine Learning/Artificial Intelligence, Image Reconstruction

Ferumoxytol can support high quality SGE cardiac cine with blood-myocardial CNR similar to SSFP cine, but free of off-resonance artifact. Much work is ongoing to accelerate the acquisition of multi-slice, breath held cardiac cine. Deep learning-based reconstruction methods can accelerate the image acquisition and reconstruction but needs a large amount of data to train. When compared with compressed sensing and low rank reconstructions, our network showed sharper images and higher consistency with the reference and significantly better quantitative evaluation metrics. We showed that a network trained with non-contrast images could generalize to accelerated ferumoxytol-enhanced cardiac cine MRI with 10x acceleration.

View Presentation Video

Lijun Zhang¹ and Sha Wang^1
1Research and Development Center, Canon Medical Systems (China), Beijing, China

Keywords: Machine Learning/Artificial Intelligence, Image Reconstruction

Model driven deep learning reconstruction methods usually utilize residual learning within single cascade which is composed by a neural network and a data consistency module, here we propose a model based end-to-end residual learning variational network(E2E-ResVarNet), a k-space residual is passed through cascades, then added to acquired under-sampled k-space after the last cascade output. It was demonstrated that the image quality is significantly improved at 4x/6x/8x acceleration factors trained with brain data set.

View Presentation Video

Victor Murray¹, Syed Siddiq¹, Ramin Jafari¹, Can Wu¹, and Ricardo Otazo^1,2
1Department of Medical Physics, Memorial Sloan Kettering Cancer Center, New York, NY, United States, ²Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY, United States

Keywords: Machine Learning/Artificial Intelligence, Radiotherapy

Motion-resolved 4D MRI enables free-breathing imaging and access to important physiological information. However, long reconstruction times for 4D MRI techniques like XD-GRASP have restricted routine clinical use. Even with unrolled convolutional networks, reconstruction enforcing data consistency in a high-dimensional space is still long. This work presents a deep learning approach named MRI-movienet that exploits spatial-time-coil correlations without enforcing data consistency to enable 2-fold scan acceleration compared to XD-GRASP and 4D reconstruction in less than 2 seconds. MRI-movienet uses the intrinsic separation into static and dynamic components to avoid hallucinations. MRI-movienet high performance will promote 4D MRI for routine clinical use.

View Presentation Video

Eze Ahanonu¹, Ute Goerke², Kevin Johnson³, Brian Toner⁴, Diego Martin⁵, Vibhas Deshpande⁶, Ali Bilgin^1,3,7, and Maria Altbach^3,7
1Department of Electrical and Computer Engineering, The University of Arizona, Tucson, AZ, United States, ²Siemens Healthineers, Tucson, AZ, United States, ³Department of Medical Imaging, The University of Arizona, Tucson, AZ, United States, ⁴Applied Math Program, The University of Arizona, Tucson, AZ, United States, ⁵Department of Radiology, Houston Methodist Hospital, Houston, TX, United States, ⁶Siemens Healthineers, Austin, TX, United States, ⁷Department of Biomedical Engineering, The University of Arizona, Tucson, AZ, United States

Keywords: Machine Learning/Artificial Intelligence, Quantitative Imaging

Comprehensive liver evaluation with T1 mapping requires full abdominal coverage with sufficiently high spatial resolution for detection of pathology. Existing methods for abdominal T1 mapping are only able to achieve partial coverage, primarily limited by the breath hold and the time required to sample the T1 recovery curve (T1RC) for accurate T1 estimation. We present a radial Look-Locker T1 mapping framework which utilizes short T1RC sampling combined with deep learning based T1 estimation to achieve full abdominal coverage within a single 20s breath hold period.

View Presentation Video

Yanhong Lin¹, Qinqin Yang¹, Wenhua Geng¹, Haitao Huang¹, Jianfeng Bao², Shuhui Cai¹, Zhong Chen¹, and Congbo Cai^1
1Department of Electronic Science, Fujian Provincial Key Laboratory of Plasma and Magnetic Resonance, Xiamen University, Xiamen, China, ²Department of Magnetic Resonance Imaging, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China

Keywords: Quantitative Imaging, Quantitative Imaging

T₂-weighted imaging via conventional FLAIR sequence can suppress the signal of cerebrospinal fluid (CSF), making it easier for identifying long T₂ lesions in the vicinity of the CSF. But it was usually used for qualitative analysis because of its inevitable time-consuming acquisition. In this study, we applied inversion recovery overlapping-echo acquisition together with deep learning-based reconstruction to achieve ultra-fast T₂-FLAIR mapping. In vivo results from a healthy volunteer and two glioma patients demonstrate the good accuracy and robustness of our proposed method.

View Presentation Video

Keywords: Machine Learning/Artificial Intelligence, Machine Learning/Artificial Intelligence

High-resolution, multi-contrast magnetic resonance imaging (MRI) protocols are required for accurate clinical diagnoses, but are limited by long scan times. Recovering high-quality, multi-contrast images from low-quality accelerated acquisitions is a promising approach to mitigate this limitation. Prior studies have demonstrated deep-learning for tasks such as contrast synthesis, image super-resolution, and image reconstruction. However, each of these tasks requires different architectures and training paradigms. Motivated by these challenges, we introduce a unified conditional denoising diffusion probabilistic model (DDPM) for inverse MR image recovery. Experiments performed on three image recovery tasks demonstrate that DDPMs achieve superior performance compared to prior state-of-the-art approaches.