Xin Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Xin Liu has a total of 287 co-authors. Here are the top co-authors by number of shared publications:
- Hairong Zheng
Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
25 shared publications - Dong Liang
Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences
20 shared publications Hairong Zheng
Paul C. Lauterbur Research Centre for Biomedical Imaging Shenzhen Institute of Advanced Technology, Chinese Academy of Science Shenzhen Guangdong China
11 shared publications- Ye Li
Chinese Academy of Sciences
9 shared publications - Chao Zou
Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences 1068 Xueyuan Avenue
8 shared publications Changjun Tie
Lauterbur Imaging Research Center, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences
7 shared publications
Latest Publications
- 26 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 74 (6) , 5551-5559
Enhanced Magnetic Resonance Imaging for Small Animals at 5 T Using a Near-Field Coupling Array
High-resolution magnetic resonance imaging (MRI) of rodent brains plays an important role in preclinical neuroscience. However, clinical MRI systems, such as 5 T scanners, offer lower magnetic field strength than ultra-high field small-bore systems (7T, 9.4T and 11.7 T), resulting in insufficient signal-to-noise ratio (SNR) for small-animal imaging. Moreover, conventional radio frequency (RF) coils are limited in their anatomical fit and suffer from signal loss due to bulky cables and connectors. In this study, we introduce a compact, passive triple-loop near-field coupling array (NFCA) that significantly enhances SNR for small-animal brain imaging. Unlike traditional single RF coils or coil arrays, the overlapping loop structure enabled uniform signal enhancement throughout the brain and conformed tightly to curved cranial surfaces without the need for cables, baluns, or preamplifiers. To evaluate compatibility, the NFCA was tested with various commercial RF receiver coils and showed a 4.80-fold SNR increase compared to the only flexible coil and a 2.53-fold SNR increase compared to the only 12-channel rat coil. In- vivo imaging of rodents at 5 T using the NFCA with a flexible commercial receiver coil demonstrated up to a 25% SNR increase in the rat brain compared to a 3-channel coil with the same size as the NFCA. The NFCA enabled high-quality anatomical structure imaging and diffusion-weighted imaging, demonstrating its effectiveness for both structural and functional neuroimaging. A high imaging resolution of 0.1 × 0.1 × 1.0 mm³ was successfully achieved using the NFCA combined with the flexible coil. This NFCA provides a practical and cost-effective solution approach to enhance preclinical MRI performance using standard clinical systems, offering new opportunities for high-resolution rodent neuroimaging in both research and translational applications.
- 26 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 74 (6) , 5551-5559
Enhanced Magnetic Resonance Imaging for Small Animals at 5 T Using a Near-Field Coupling Array
High-resolution magnetic resonance imaging (MRI) of rodent brains plays an important role in preclinical neuroscience. However, clinical MRI systems, such as 5 T scanners, offer lower magnetic field strength than ultra-high field small-bore systems (7T, 9.4T and 11.7 T), resulting in insufficient signal-to-noise ratio (SNR) for small-animal imaging. Moreover, conventional radio frequency (RF) coils are limited in their anatomical fit and suffer from signal loss due to bulky cables and connectors. In this study, we introduce a compact, passive triple-loop near-field coupling array (NFCA) that significantly enhances SNR for small-animal brain imaging. Unlike traditional single RF coils or coil arrays, the overlapping loop structure enabled uniform signal enhancement throughout the brain and conformed tightly to curved cranial surfaces without the need for cables, baluns, or preamplifiers. To evaluate compatibility, the NFCA was tested with various commercial RF receiver coils and showed a 4.80-fold SNR increase compared to the only flexible coil and a 2.53-fold SNR increase compared to the only 12-channel rat coil. In- vivo imaging of rodents at 5 T using the NFCA with a flexible commercial receiver coil demonstrated up to a 25% SNR increase in the rat brain compared to a 3-channel coil with the same size as the NFCA. The NFCA enabled high-quality anatomical structure imaging and diffusion-weighted imaging, demonstrating its effectiveness for both structural and functional neuroimaging. A high imaging resolution of 0.1 × 0.1 × 1.0 mm³ was successfully achieved using the NFCA combined with the flexible coil. This NFCA provides a practical and cost-effective solution approach to enhance preclinical MRI performance using standard clinical systems, offering new opportunities for high-resolution rodent neuroimaging in both research and translational applications.
- Open Access
- 21 October 2025
- Published by MDPI AG in Biomedicines
- Vol. 13 (10) , 2562
Preoperative Prediction of Axillary Lymph Node Metastasis in Breast Cancer Using Radiomics Features of Voxel-Wise DCE-MRI Time-Intensity-Curve Profile Maps
Objective: Axillary lymph node (ALN) status in breast cancer is pivotal for guiding treatment and determining prognosis. The study aimed to explore the feasibility and efficacy of a radiomics model using voxel-wise dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) time-intensity-curve (TIC) profile maps to predict ALN metastasis in breast cancer. Methods: A total of 615 breast cancer patients who underwent preoperative DCE-MRI from October 2018 to February 2024 were retrospectively enrolled and randomly allocated into training (n = 430) and testing (n = 185) sets (7:3 ratio). Based on wash-in rate, wash-out enhancement, and wash-out stability, each voxel within manually segmented 3D lesions that were categorized into 1 of 19 TIC subtypes from the DCE-MRI images. Three feature sets were derived: composition ratio (type-19), radiomics features of TIC subtypes (type-19-radiomics), and radiomics features of third-phase DCE-MRI (phase-3-radiomics). Student’s t-test and the least absolute shrinkage and selection operator (LASSO) was used to select features. Four models (type-19, type-19-radiomics, type-19-combined, and phase-3-radiomics) were constructed by a support vector machine (SVM) to predict ALN status. Model performance was assessed using sensitivity, specificity, accuracy, F1 score, and area under the curve (AUC). Results: The type-19-combined model significantly outperformed the phase-3-radiomics model (AUC = 0.779 vs. 0.698, p < 0.001; 0.674 vs. 0.559) and the type-19 model (AUC = 0.779 vs. 0.541, p < 0.001; 0.674 vs. 0.435, p < 0.001) in cross-validation and independent testing sets. The type-19-radiomics showed significantly better performance than the phase-3-radiomics model (AUC = 0.764 vs. 0.698, p = 0.002; 0.657 vs. 0.559, p = 0.037) and type-19 model (AUC = 0. 764 vs. 0.541, p < 0.001; 0.657 vs. 0.435, p < 0.001) in cross-validation and independent testing sets. Among four models, the type-19-combined model achieved the highest AUC (0.779, 0.674) in cross-validation and testing sets. Conclusions: Radiomics analysis of voxel-wise DCE-MRI TIC profile maps, simultaneously quantifying temporal and spatial hemodynamic heterogeneity, provides an effective, noninvasive method for predicting ALN metastasis in breast cancer.
- 21 October 2025
- Published by MDPI AG in Biomedicines
- Vol. 13 (10) , 2562
Open AccessPreoperative Prediction of Axillary Lymph Node Metastasis in Breast Cancer Using Radiomics Features of Voxel-Wise DCE-MRI Time-Intensity-Curve Profile Maps
Objective: Axillary lymph node (ALN) status in breast cancer is pivotal for guiding treatment and determining prognosis. The study aimed to explore the feasibility and efficacy of a radiomics model using voxel-wise dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) time-intensity-curve (TIC) profile maps to predict ALN metastasis in breast cancer. Methods: A total of 615 breast cancer patients who underwent preoperative DCE-MRI from October 2018 to February 2024 were retrospectively enrolled and randomly allocated into training (n = 430) and testing (n = 185) sets (7:3 ratio). Based on wash-in rate, wash-out enhancement, and wash-out stability, each voxel within manually segmented 3D lesions that were categorized into 1 of 19 TIC subtypes from the DCE-MRI images. Three feature sets were derived: composition ratio (type-19), radiomics features of TIC subtypes (type-19-radiomics), and radiomics features of third-phase DCE-MRI (phase-3-radiomics). Student’s t-test and the least absolute shrinkage and selection operator (LASSO) was used to select features. Four models (type-19, type-19-radiomics, type-19-combined, and phase-3-radiomics) were constructed by a support vector machine (SVM) to predict ALN status. Model performance was assessed using sensitivity, specificity, accuracy, F1 score, and area under the curve (AUC). Results: The type-19-combined model significantly outperformed the phase-3-radiomics model (AUC = 0.779 vs. 0.698, p < 0.001; 0.674 vs. 0.559) and the type-19 model (AUC = 0.779 vs. 0.541, p < 0.001; 0.674 vs. 0.435, p < 0.001) in cross-validation and independent testing sets. The type-19-radiomics showed significantly better performance than the phase-3-radiomics model (AUC = 0.764 vs. 0.698, p = 0.002; 0.657 vs. 0.559, p = 0.037) and type-19 model (AUC = 0. 764 vs. 0.541, p < 0.001; 0.657 vs. 0.435, p < 0.001) in cross-validation and independent testing sets. Among four models, the type-19-combined model achieved the highest AUC (0.779, 0.674) in cross-validation and testing sets. Conclusions: Radiomics analysis of voxel-wise DCE-MRI TIC profile maps, simultaneously quantifying temporal and spatial hemodynamic heterogeneity, provides an effective, noninvasive method for predicting ALN metastasis in breast cancer.
- 6 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Signal Processing
- Vol. 19 (8) , 1694-1709
DUN-SRE: Deep Unrolling Network With Spatiotemporal Rotation Equivariance for Dynamic MRI Reconstruction
Dynamic Magnetic Resonance Imaging (MRI) exhibits transformation symmetries, including spatial rotation symmetry within individual frames and temporal symmetry along the time dimension. Explicit incorporation of these symmetry priors in the reconstruction model can significantly improve image quality, especially under aggressive undersampling scenarios. Recently, Equivariant convolutional neural network (ECNN) has shown great promise in exploiting spatial symmetry priors. However, existing ECNNs critically fail to model temporal symmetry, arguably the most universal and informative structural prior in dynamic MRI reconstruction. To tackle this issue, we propose a novel Deep Unrolling Network with Spatiotemporal Rotation Equivariance (DUN-SRE) for Dynamic MRI Reconstruction. The DUN-SRE establishes spatiotemporal equivariance through a (2+1)D equivariant convolutional architecture. In particular, it integrates both the data consistency and proximal mapping module into a unified deep unrolling framework. This architecture ensures rigorous propagation of spatiotemporal rotation symmetry constraints throughout the reconstruction process, enabling more physically accurate modeling of cardiac motion dynamics in cine MRI. In addition, a high-fidelity group filter parameterization mechanism is developed to maintain representation precision while enforcing symmetry constraints. Comprehensive experiments on Cardiac CINE MRI datasets demonstrate that DUN-SRE achieves state-of-the-art performance, particularly in preserving rotation-symmetric structures, offering strong generalization capability to a broad range of dynamic MRI reconstruction tasks.
- 6 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Signal Processing
- Vol. 19 (8) , 1694-1709
DUN-SRE: Deep Unrolling Network With Spatiotemporal Rotation Equivariance for Dynamic MRI Reconstruction
Dynamic Magnetic Resonance Imaging (MRI) exhibits transformation symmetries, including spatial rotation symmetry within individual frames and temporal symmetry along the time dimension. Explicit incorporation of these symmetry priors in the reconstruction model can significantly improve image quality, especially under aggressive undersampling scenarios. Recently, Equivariant convolutional neural network (ECNN) has shown great promise in exploiting spatial symmetry priors. However, existing ECNNs critically fail to model temporal symmetry, arguably the most universal and informative structural prior in dynamic MRI reconstruction. To tackle this issue, we propose a novel Deep Unrolling Network with Spatiotemporal Rotation Equivariance (DUN-SRE) for Dynamic MRI Reconstruction. The DUN-SRE establishes spatiotemporal equivariance through a (2+1)D equivariant convolutional architecture. In particular, it integrates both the data consistency and proximal mapping module into a unified deep unrolling framework. This architecture ensures rigorous propagation of spatiotemporal rotation symmetry constraints throughout the reconstruction process, enabling more physically accurate modeling of cardiac motion dynamics in cine MRI. In addition, a high-fidelity group filter parameterization mechanism is developed to maintain representation precision while enforcing symmetry constraints. Comprehensive experiments on Cardiac CINE MRI datasets demonstrate that DUN-SRE achieves state-of-the-art performance, particularly in preserving rotation-symmetric structures, offering strong generalization capability to a broad range of dynamic MRI reconstruction tasks.
- 10 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-11
High-Power Radio Frequency Amplifier for 5-T MRI Whole-Body Scanner
Magnetic resonance imaging (MRI) at ultra-high field (UHF) requires radio-frequency power amplifiers (RFPA) with high power capability for whole-body imaging. Insufficient RF power restricts the peak transmit magnetic field (B1+) strength, thereby impeding the ability to achieve sufficient high flip angles for body imaging. This study presents an RFPA that can achieve a peak power capability of 64 kW, the highest reported value for UHF MRI systems. A novel impedance matching circuit design method was proposed to optimize the power capability of each RF transistor. The designed RFPA features an innovative transmit architecture that accommodates various applications by configuring independent transmit channels in 8-channel, 4-channel, or 2-channel modes. High gain linearity of 0.2 dB / 3.6°, 0.3 dB / 5.2°, and 0.2 dB / 4° were achieved across different operating modes. The B1+ maps and body images acquired using a 5 Tesla whole-body MRI scanner demonstrated that sufficient flip-angles for whole body imaging can be achieved with the proposed RFPA. The body images also exhibited good uniformity in contrast.
- 10 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-11
High-Power Radio Frequency Amplifier for 5-T MRI Whole-Body Scanner
Magnetic resonance imaging (MRI) at ultra-high field (UHF) requires radio-frequency power amplifiers (RFPA) with high power capability for whole-body imaging. Insufficient RF power restricts the peak transmit magnetic field (B1+) strength, thereby impeding the ability to achieve sufficient high flip angles for body imaging. This study presents an RFPA that can achieve a peak power capability of 64 kW, the highest reported value for UHF MRI systems. A novel impedance matching circuit design method was proposed to optimize the power capability of each RF transistor. The designed RFPA features an innovative transmit architecture that accommodates various applications by configuring independent transmit channels in 8-channel, 4-channel, or 2-channel modes. High gain linearity of 0.2 dB / 3.6°, 0.3 dB / 5.2°, and 0.2 dB / 4° were achieved across different operating modes. The B1+ maps and body images acquired using a 5 Tesla whole-body MRI scanner demonstrated that sufficient flip-angles for whole body imaging can be achieved with the proposed RFPA. The body images also exhibited good uniformity in contrast.
- 22 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Microwave Theory and Techniques
- Vol. PP (00189480) , 1-12
A 2-kW High-Efficiency and Broadband GaN Linear RF Power Amplifier for Multinuclear Magnetic Resonance Imaging
Multinuclear magnetic resonance imaging (MRI) requires high-power, broadband radio frequency (RF) amplification to transmit signals efficiently and linearly. In this study, we propose an innovative 2-kW highly efficient broadband gallium nitride (GaN) linear power amplifier designed for multinuclear MRI at 5.0 T, showing its first application in multinuclear MRI. The amplifier design is optimized through a multiobjective optimization load–pull analysis, enabling optimal impedance matching across a wideband range (30–300 MHz) to achieve high power, efficiency, and gain flatness. In addition, a broadband high-power combiner with compensation inductors was developed to enhance phase and amplitude consistency while optimizing return loss, insertion loss, and isolation. Furthermore, an adaptive frequency-based pre-compensation linearization technique was implemented to improve amplitude linearity. Experimental validations were performed for 1H and 2H spectroscopy and imaging at 5.0T MRI. The amplifier achieves 2-kW output power across 30–300 MHz with efficiencies of 68.1% (1H), 70.3% (2H), 63.7% (23Na), and 64.8% (31P). Furthermore, gain variation and phase changes remain within ±0.9 dB and 10° over a 40-dB dynamic range. The proposed GaN amplifier offers broadband amplification and high efficiency, making it a suitable candidate for multinuclear MRI.
- 22 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Microwave Theory and Techniques
- Vol. PP (00189480) , 1-12
A 2-kW High-Efficiency and Broadband GaN Linear RF Power Amplifier for Multinuclear Magnetic Resonance Imaging
Multinuclear magnetic resonance imaging (MRI) requires high-power, broadband radio frequency (RF) amplification to transmit signals efficiently and linearly. In this study, we propose an innovative 2-kW highly efficient broadband gallium nitride (GaN) linear power amplifier designed for multinuclear MRI at 5.0 T, showing its first application in multinuclear MRI. The amplifier design is optimized through a multiobjective optimization load–pull analysis, enabling optimal impedance matching across a wideband range (30–300 MHz) to achieve high power, efficiency, and gain flatness. In addition, a broadband high-power combiner with compensation inductors was developed to enhance phase and amplitude consistency while optimizing return loss, insertion loss, and isolation. Furthermore, an adaptive frequency-based pre-compensation linearization technique was implemented to improve amplitude linearity. Experimental validations were performed for 1H and 2H spectroscopy and imaging at 5.0T MRI. The amplifier achieves 2-kW output power across 30–300 MHz with efficiencies of 68.1% (1H), 70.3% (2H), 63.7% (23Na), and 64.8% (31P). Furthermore, gain variation and phase changes remain within ±0.9 dB and 10° over a 40-dB dynamic range. The proposed GaN amplifier offers broadband amplification and high efficiency, making it a suitable candidate for multinuclear MRI.
- 4 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-12
A Quadruple-Tuned RF Coil System for ¹H/¹⁹F/²³Na/³¹P Synchronous MR Imaging and Spectroscopy at 3 T
The simultaneous acquisition of biological information at different levels of tumor molecular targets, energy metabolism, ion perturbations and structural function is significant for tumor research, as MRI and MRS are crucial tools for comprehensive characterization of tumor pathology and monitoring tumor treatment response. As critical component of RF signal acquisition, the RF transceivers provide the detection capability of signals. Most of the acquisitions for detecting weak nuclei mainly use double tuned probes or RF coils and multi-tuned coils with limitation. There is currently no reported quadruple-tuned multi-channel RF coil design to support simultaneous imaging of four nuclei for animal studies at 3 T. In this work, we developed an innovative quadruple-tuned multi-channel transceiver coil array for synchronous 1H/19F/23Na/31P MR imaging in animals at 3.0 T, which provided a potential tool for tumor research at different levels of tumor molecular targets, energy metabolism, ion perturbations and structural function in MRI application. In contrast to receive-only RF coils, the proposed RF coil system supports multichannel parallel excitation and parallel reception for all four nuclei (1H, 19F, 23Na, 31P) and enables B1+ shimming to control the transmit field distribution. In order to maximize the SNR for all nuclei, detuning control strategy was innovatively utilized to counteract the electromagnetic interference in the RF system. Bench tests and numerical simulations were conducted to assess the performance of the developed quadruple-tuned RF coil system. The phantom and in vivo imaging experiments were performed on a 3T scanner with multi-nuclear imaging capability to evaluate the practicability of our developed design, where SNR of the quadruple-tuned coil was compared with 1H/19F doubletuned coil array, 23Na /31P single-tuned coil arrays. The results demonstrated that the proposed quadruple-tuned RF coil system possessed high sensitivity and relatively homogeneous RF transmission and reception profiles with the ability to image 1H, 19F, 23Na, and 31P synchronously, which has the potential to simultaneously provide physical, biochemical, functional information for tumor research in addition to structural information.
- 4 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-12
A Quadruple-Tuned RF Coil System for ¹H/¹⁹F/²³Na/³¹P Synchronous MR Imaging and Spectroscopy at 3 T
The simultaneous acquisition of biological information at different levels of tumor molecular targets, energy metabolism, ion perturbations and structural function is significant for tumor research, as MRI and MRS are crucial tools for comprehensive characterization of tumor pathology and monitoring tumor treatment response. As critical component of RF signal acquisition, the RF transceivers provide the detection capability of signals. Most of the acquisitions for detecting weak nuclei mainly use double tuned probes or RF coils and multi-tuned coils with limitation. There is currently no reported quadruple-tuned multi-channel RF coil design to support simultaneous imaging of four nuclei for animal studies at 3 T. In this work, we developed an innovative quadruple-tuned multi-channel transceiver coil array for synchronous 1H/19F/23Na/31P MR imaging in animals at 3.0 T, which provided a potential tool for tumor research at different levels of tumor molecular targets, energy metabolism, ion perturbations and structural function in MRI application. In contrast to receive-only RF coils, the proposed RF coil system supports multichannel parallel excitation and parallel reception for all four nuclei (1H, 19F, 23Na, 31P) and enables B1+ shimming to control the transmit field distribution. In order to maximize the SNR for all nuclei, detuning control strategy was innovatively utilized to counteract the electromagnetic interference in the RF system. Bench tests and numerical simulations were conducted to assess the performance of the developed quadruple-tuned RF coil system. The phantom and in vivo imaging experiments were performed on a 3T scanner with multi-nuclear imaging capability to evaluate the practicability of our developed design, where SNR of the quadruple-tuned coil was compared with 1H/19F doubletuned coil array, 23Na /31P single-tuned coil arrays. The results demonstrated that the proposed quadruple-tuned RF coil system possessed high sensitivity and relatively homogeneous RF transmission and reception profiles with the ability to image 1H, 19F, 23Na, and 31P synchronously, which has the potential to simultaneously provide physical, biochemical, functional information for tumor research in addition to structural information.
- 3 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44 (2) , 1019-1031
SPIRiT-Diffusion: Self-Consistency Driven Diffusion Model for Accelerated MRI
Diffusion models have emerged as a leading methodology for image generation and have proven successful in the realm of magnetic resonance imaging (MRI) reconstruction. However, existing reconstruction methods based on diffusion models are primarily formulated in the image domain, making the reconstruction quality susceptible to inaccuracies in coil sensitivity maps (CSMs). k -space interpolation methods can effectively address this issue but conventional diffusion models are not readily applicable in k -space interpolation. To overcome this challenge, we introduce a novel approach called SPIRiT-Diffusion, which is a diffusion model for k -space interpolation inspired by the iterative self-consistent SPIRiT method. Specifically, we utilize the iterative solver of the self-consistent term (i.e., k -space physical prior) in SPIRiT to formulate a novel stochastic differential equation (SDE) governing the diffusion process. Subsequently, k -space data can be interpolated by executing the diffusion process. This innovative approach highlights the optimization model’s role in designing the SDE in diffusion models, enabling the diffusion process to align closely with the physics inherent in the optimization model-a concept referred to as model-driven diffusion. We evaluated the proposed SPIRiT-Diffusion method using a 3D joint intracranial and carotid vessel wall imaging dataset. The results convincingly demonstrate its superiority over image-domain reconstruction methods, achieving high reconstruction quality even at a substantial acceleration rate of 10. Our code are available at https://github.com/zhyjSIAT/SPIRiT-Diffusion.
- 3 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44 (2) , 1019-1031
SPIRiT-Diffusion: Self-Consistency Driven Diffusion Model for Accelerated MRI
Diffusion models have emerged as a leading methodology for image generation and have proven successful in the realm of magnetic resonance imaging (MRI) reconstruction. However, existing reconstruction methods based on diffusion models are primarily formulated in the image domain, making the reconstruction quality susceptible to inaccuracies in coil sensitivity maps (CSMs). k -space interpolation methods can effectively address this issue but conventional diffusion models are not readily applicable in k -space interpolation. To overcome this challenge, we introduce a novel approach called SPIRiT-Diffusion, which is a diffusion model for k -space interpolation inspired by the iterative self-consistent SPIRiT method. Specifically, we utilize the iterative solver of the self-consistent term (i.e., k -space physical prior) in SPIRiT to formulate a novel stochastic differential equation (SDE) governing the diffusion process. Subsequently, k -space data can be interpolated by executing the diffusion process. This innovative approach highlights the optimization model’s role in designing the SDE in diffusion models, enabling the diffusion process to align closely with the physics inherent in the optimization model-a concept referred to as model-driven diffusion. We evaluated the proposed SPIRiT-Diffusion method using a 3D joint intracranial and carotid vessel wall imaging dataset. The results convincingly demonstrate its superiority over image-domain reconstruction methods, achieving high reconstruction quality even at a substantial acceleration rate of 10. Our code are available at https://github.com/zhyjSIAT/SPIRiT-Diffusion.
- 5 December 2023
- Published by Wiley in Journal of Magnetic Resonance Imaging
- Vol. 60 (3) , 1027-1034
Fatty Acids Composition of the Sacroiliac Joint in Axial Spondyloarthritis: Analysis Using 3.0 T Chemical Shift‐Encoded MRI
Background: Axial spondyloarthritis (axSpA) is a group of inflammatory diseases that may lead to ankylosis of the sacroiliac joint and spine. Fat lesion in the sacroiliac joint is an important feature in diagnosis and disease progression of axSpA. However, whether there is alteration of fatty acids (FAs) composition has not been investigated using MRI.Purpose: To investigate bone marrow FA composition of the sacroiliac joint in patients with axSpA compared to controls.Study Type: Prospective.Subjects: Eighty five participants (mean age, 32.3 ± 6.1 years): 48 axSpA (25 male, 23 female) and 37 non‐SpA controls (18 male, 19 female).Field Strength/Sequence: 3.0 T/Two multiple gradient‐echo chemical shift‐encoded (CSE) MRI which differed only in echo times (TEs) were scanned consecutively.Assessment: Axial multi‐echo CSE MRI was performed in the sacroiliac joints in vivo. Regions of interest (ROIs) were manually placed on subchondral bone with and without fat lesion in axSpA patients, and on subchondral bone without fat lesion in controls. FA composition was computed within the ROIs using a nonlinear least square method from literature.Statistical Tests: Intergroup comparisons were performed using t tests.Results: In axSpA, male patients had significantly higher monounsaturated FA compared to controls in areas with fat lesion in the sacrum (+12%) and in the ilium (+9%), and in areas without fat lesion in the sacrum (+10%). Significantly lower polyunsaturated FAs were found in areas with fat lesion in the sacrum (−10%) and ilium (−11%), and lower saturated FAs were found in areas without fat lesion in the sacrum (−6%). In female, patients with axSpA had significantly higher saturated FAs in areas with fat lesion in the ilium (+7%) in comparison to controls.Data Conclusion: FA composition of the sacroiliac joint alters in patients with axSpA, and it can be detected using CSE MRI based analysis.
- 5 December 2023
- Published by Wiley in Journal of Magnetic Resonance Imaging
- Vol. 60 (3) , 1027-1034
Fatty Acids Composition of the Sacroiliac Joint in Axial Spondyloarthritis: Analysis Using 3.0 T Chemical Shift‐Encoded MRI
Background: Axial spondyloarthritis (axSpA) is a group of inflammatory diseases that may lead to ankylosis of the sacroiliac joint and spine. Fat lesion in the sacroiliac joint is an important feature in diagnosis and disease progression of axSpA. However, whether there is alteration of fatty acids (FAs) composition has not been investigated using MRI.Purpose: To investigate bone marrow FA composition of the sacroiliac joint in patients with axSpA compared to controls.Study Type: Prospective.Subjects: Eighty five participants (mean age, 32.3 ± 6.1 years): 48 axSpA (25 male, 23 female) and 37 non‐SpA controls (18 male, 19 female).Field Strength/Sequence: 3.0 T/Two multiple gradient‐echo chemical shift‐encoded (CSE) MRI which differed only in echo times (TEs) were scanned consecutively.Assessment: Axial multi‐echo CSE MRI was performed in the sacroiliac joints in vivo. Regions of interest (ROIs) were manually placed on subchondral bone with and without fat lesion in axSpA patients, and on subchondral bone without fat lesion in controls. FA composition was computed within the ROIs using a nonlinear least square method from literature.Statistical Tests: Intergroup comparisons were performed using t tests.Results: In axSpA, male patients had significantly higher monounsaturated FA compared to controls in areas with fat lesion in the sacrum (+12%) and in the ilium (+9%), and in areas without fat lesion in the sacrum (+10%). Significantly lower polyunsaturated FAs were found in areas with fat lesion in the sacrum (−10%) and ilium (−11%), and lower saturated FAs were found in areas without fat lesion in the sacrum (−6%). In female, patients with axSpA had significantly higher saturated FAs in areas with fat lesion in the ilium (+7%) in comparison to controls.Data Conclusion: FA composition of the sacroiliac joint alters in patients with axSpA, and it can be detected using CSE MRI based analysis.
- Open Access
- 8 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 890-894
X-Ray Spectral Estimation Using Dictionary Learning
As computational tools for X-ray computed tomography (CT) become more quantitatively accurate, knowledge of the source-detector spectral response is critical for quantitative system-independent reconstruction and material characterization capabilities. Directly measuring the spectral response of a CT system is hard, which motivates spectral estimation using transmission data obtained from a collection of known homogeneous objects. However, the associated inverse problem is ill-conditioned, making accurate estimation of the spectrum challenging, particularly in the absence of a close initial guess.In this paper, we describe a dictionary-based spectral estimation method that yields accurate results without the need for any initial estimate of the spectral response. Our method utilizes a MAP estimation framework that combines a physics-based forward model along with an L 0 sparsity constraint and a simplex constraint on the dictionary coefficients. Our method uses a greedy support selection method and a new pair-wise iterated coordinate descent method to compute the above estimate. We demonstrate that our dictionary-based method outperforms a state-of-the-art method as shown in a cross-validation experiment on four real datasets collected at beamline 8.3.2 of the Advanced Light Source (ALS).
- 8 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 890-894
Open AccessX-Ray Spectral Estimation Using Dictionary Learning
As computational tools for X-ray computed tomography (CT) become more quantitatively accurate, knowledge of the source-detector spectral response is critical for quantitative system-independent reconstruction and material characterization capabilities. Directly measuring the spectral response of a CT system is hard, which motivates spectral estimation using transmission data obtained from a collection of known homogeneous objects. However, the associated inverse problem is ill-conditioned, making accurate estimation of the spectrum challenging, particularly in the absence of a close initial guess.In this paper, we describe a dictionary-based spectral estimation method that yields accurate results without the need for any initial estimate of the spectral response. Our method utilizes a MAP estimation framework that combines a physics-based forward model along with an L 0 sparsity constraint and a simplex constraint on the dictionary coefficients. Our method uses a greedy support selection method and a new pair-wise iterated coordinate descent method to compute the above estimate. We demonstrate that our dictionary-based method outperforms a state-of-the-art method as shown in a cross-validation experiment on four real datasets collected at beamline 8.3.2 of the Advanced Light Source (ALS).
- Open Access
- 21 September 2023
- Published by MDPI AG in Bioengineering
- Vol. 10 (9) , 1107
Joint Image Reconstruction and Super-Resolution for Accelerated Magnetic Resonance Imaging
Magnetic resonance (MR) image reconstruction and super-resolution are two prominent techniques to restore high-quality images from undersampled or low-resolution k-space data to accelerate MR imaging. Combining undersampled and low-resolution acquisition can further improve the acceleration factor. Existing methods often treat the techniques of image reconstruction and super-resolution separately or combine them sequentially for image recovery, which can result in error propagation and suboptimal results. In this work, we propose a novel framework for joint image reconstruction and super-resolution, aiming to efficiently image recovery and enable fast imaging. Specifically, we designed a framework with a reconstruction module and a super-resolution module to formulate multi-task learning. The reconstruction module utilizes a model-based optimization approach, ensuring data fidelity with the acquired k-space data. Moreover, a deep spatial feature transform is employed to enhance the information transition between the two modules, facilitating better integration of image reconstruction and super-resolution. Experimental evaluations on two datasets demonstrate that our proposed method can provide superior performance both quantitatively and qualitatively.
- 21 September 2023
- Published by MDPI AG in Bioengineering
- Vol. 10 (9) , 1107
Open AccessJoint Image Reconstruction and Super-Resolution for Accelerated Magnetic Resonance Imaging
Magnetic resonance (MR) image reconstruction and super-resolution are two prominent techniques to restore high-quality images from undersampled or low-resolution k-space data to accelerate MR imaging. Combining undersampled and low-resolution acquisition can further improve the acceleration factor. Existing methods often treat the techniques of image reconstruction and super-resolution separately or combine them sequentially for image recovery, which can result in error propagation and suboptimal results. In this work, we propose a novel framework for joint image reconstruction and super-resolution, aiming to efficiently image recovery and enable fast imaging. Specifically, we designed a framework with a reconstruction module and a super-resolution module to formulate multi-task learning. The reconstruction module utilizes a model-based optimization approach, ensuring data fidelity with the acquired k-space data. Moreover, a deep spatial feature transform is employed to enhance the information transition between the two modules, facilitating better integration of image reconstruction and super-resolution. Experimental evaluations on two datasets demonstrate that our proposed method can provide superior performance both quantitatively and qualitatively.