Peng Hu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Peng Hu has a total of 110 co-authors. Here are the top co-authors by number of shared publications:
Fei Han
Department of Radiological Sciences, University of California, Los Angeles, CA, USA
19 shared publications- Kim-Lien Nguyen
Division of Cardiology, David Geffen School of Medicine at UCLA and VA Greater Los Angeles Healthcare System, Los Angeles, USA
13 shared publications Ziwu Zhou
Department of Radiological Sciences, University of California, Los Angeles, CA, USA
12 shared publicationsJ. Paul Finn
Department of Radiological Sciences David Geffen School of Medicine University of California Los Angeles California USA
11 shared publicationsYu Gao
Department of Radiological Sciences, University of California, Los Angeles, CA, USA
11 shared publicationsYingli Yang
Physics and Biology in Medicine IDP, University of California, Los Angeles, CA, USA
10 shared publications
Latest Publications
- 5 October 2021
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 87 (2) , 984-998
Automatic segmentation of peripheral arteries and veins in ferumoxytol‐enhanced MR angiography
Purpose: To automate the segmentation of the peripheral arteries and veins in the lower extremities based on ferumoxytol‐enhanced MR angiography (FE‐MRA).Methods: Our automated pipeline has 2 sequential stages. In the first stage, we used a 3D U‐Net with local attention gates, which was trained based on a combination of the Focal Tversky loss with region mutual loss under a deep supervision mechanism to segment the vasculature from the high‐resolution FE‐MRA datasets. In the second stage, we used time‐resolved images to separate the arteries from the veins. Because the ultimate segmentation quality of the arteries and veins relies on the performance of the first stage, we thoroughly evaluated the different aspects of the segmentation network and compared its performance in blood vessel segmentation with currently accepted state‐of‐the‐art networks, including Volumetric‐Net, DeepVesselNet‐FCN, and Uception.Results: We achieved a competitive F1 = 0.8087 and recall = 0.8410 for blood vessel segmentation compared with F1 = (0.7604, 0.7573, 0.7651) and recall = (0.7791, 0.7570, 0.7774) obtained with Volumetric‐Net, DeepVesselNet‐FCN, and Uception. For the artery and vein separation stage, we achieved F1 = (0.8274/0.7863) in the calf region, which is the most challenging region in peripheral arteries and veins segmentation.Conclusion: Our pipeline is capable of fully automatic vessel segmentation based on FE‐MRA without need for human interaction in <4 min. This method improves upon manual segmentation by radiologists, which routinely takes several hours.
- 5 October 2021
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 87 (2) , 984-998
Automatic segmentation of peripheral arteries and veins in ferumoxytol‐enhanced MR angiography
Purpose: To automate the segmentation of the peripheral arteries and veins in the lower extremities based on ferumoxytol‐enhanced MR angiography (FE‐MRA).Methods: Our automated pipeline has 2 sequential stages. In the first stage, we used a 3D U‐Net with local attention gates, which was trained based on a combination of the Focal Tversky loss with region mutual loss under a deep supervision mechanism to segment the vasculature from the high‐resolution FE‐MRA datasets. In the second stage, we used time‐resolved images to separate the arteries from the veins. Because the ultimate segmentation quality of the arteries and veins relies on the performance of the first stage, we thoroughly evaluated the different aspects of the segmentation network and compared its performance in blood vessel segmentation with currently accepted state‐of‐the‐art networks, including Volumetric‐Net, DeepVesselNet‐FCN, and Uception.Results: We achieved a competitive F1 = 0.8087 and recall = 0.8410 for blood vessel segmentation compared with F1 = (0.7604, 0.7573, 0.7651) and recall = (0.7791, 0.7570, 0.7774) obtained with Volumetric‐Net, DeepVesselNet‐FCN, and Uception. For the artery and vein separation stage, we achieved F1 = (0.8274/0.7863) in the calf region, which is the most challenging region in peripheral arteries and veins segmentation.Conclusion: Our pipeline is capable of fully automatic vessel segmentation based on FE‐MRA without need for human interaction in <4 min. This method improves upon manual segmentation by radiologists, which routinely takes several hours.
- Open Access
Evaluation of T2-Weighted MRI for Visualization and Sparing of Urethra with MR-Guided Radiation Therapy (MRgRT) On-Board MRI
Purpose: To evaluate urethral contours from two optimized urethral MRI sequences with an MR-guided radiotherapy system (MRgRT). Methods: Eleven prostate cancer patients were scanned on a MRgRT system using optimized urethral 3D HASTE and 3D TSE. A resident radiation oncologist contoured the prostatic urethra on the patients’ planning CT, diagnostic 3T T2w MRI, and both urethral MRIs. An attending radiation oncologist reviewed/edited the resident’s contours and additionally contoured the prostatic urethra on the clinical planning MRgRT MRI (bSSFP). For each image, the resident radiation oncologist, attending radiation oncologist, and a senior medical physicist qualitatively scored the prostatic urethra visibility. Using MRgRT 3D HASTE-based contouring workflow as baseline, prostatic urethra contours drawn on CT, diagnostic MRI, clinical bSSFP and 3D TSE were evaluated relative to the contour on 3D HASTE using 95th percentile Hausdorff distance (HD95), mean-distance-to-agreement (MDA), and DICE coefficient. Additionally, prostatic urethra contrast-to-noise-ratios (CNR) were calculated for all images. Results: For two out of three observers, the urethra visibility score for 3D HASTE was significantly higher than CT, and clinical bSSFP, but was not significantly different from diagnostic MRI. The mean HD95/MDA/DICE values were 11.35 ± 3.55 mm/5.77 ± 2.69 mm/0.07 ± 0.08 for CT, 7.62 ± 2.75 mm/3.83 ± 1.47 mm/0.12 ± 0.10 for CT + diagnostic MRI, 5.49 ± 2.32 mm/2.18 ± 1.19 mm/0.35 ± 0.19 for 3D TSE, and 6.34 ± 2.89 mm/2.65 ± 1.31 mm/0.21 ± 0.12 for clinical bSSFP. The CNR for 3D HASTE was significantly higher than CT, diagnostic MRI, and clinical bSSFP, but was not significantly different from 3D TSE. Conclusion: The urethra’s visibility scores showed optimized urethral MRgRT 3D HASTE was superior to the other tested methodologies. The prostatic urethra contours demonstrated significant variability from different imaging and workflows. Urethra contouring uncertainty introduced by cross-modality registration and sub-optimal imaging contrast may lead to significant treatment degradation when urethral sparing is implemented to minimize genitourinary toxicity.
Open AccessEvaluation of T2-Weighted MRI for Visualization and Sparing of Urethra with MR-Guided Radiation Therapy (MRgRT) On-Board MRI
Purpose: To evaluate urethral contours from two optimized urethral MRI sequences with an MR-guided radiotherapy system (MRgRT). Methods: Eleven prostate cancer patients were scanned on a MRgRT system using optimized urethral 3D HASTE and 3D TSE. A resident radiation oncologist contoured the prostatic urethra on the patients’ planning CT, diagnostic 3T T2w MRI, and both urethral MRIs. An attending radiation oncologist reviewed/edited the resident’s contours and additionally contoured the prostatic urethra on the clinical planning MRgRT MRI (bSSFP). For each image, the resident radiation oncologist, attending radiation oncologist, and a senior medical physicist qualitatively scored the prostatic urethra visibility. Using MRgRT 3D HASTE-based contouring workflow as baseline, prostatic urethra contours drawn on CT, diagnostic MRI, clinical bSSFP and 3D TSE were evaluated relative to the contour on 3D HASTE using 95th percentile Hausdorff distance (HD95), mean-distance-to-agreement (MDA), and DICE coefficient. Additionally, prostatic urethra contrast-to-noise-ratios (CNR) were calculated for all images. Results: For two out of three observers, the urethra visibility score for 3D HASTE was significantly higher than CT, and clinical bSSFP, but was not significantly different from diagnostic MRI. The mean HD95/MDA/DICE values were 11.35 ± 3.55 mm/5.77 ± 2.69 mm/0.07 ± 0.08 for CT, 7.62 ± 2.75 mm/3.83 ± 1.47 mm/0.12 ± 0.10 for CT + diagnostic MRI, 5.49 ± 2.32 mm/2.18 ± 1.19 mm/0.35 ± 0.19 for 3D TSE, and 6.34 ± 2.89 mm/2.65 ± 1.31 mm/0.21 ± 0.12 for clinical bSSFP. The CNR for 3D HASTE was significantly higher than CT, diagnostic MRI, and clinical bSSFP, but was not significantly different from 3D TSE. Conclusion: The urethra’s visibility scores showed optimized urethral MRgRT 3D HASTE was superior to the other tested methodologies. The prostatic urethra contours demonstrated significant variability from different imaging and workflows. Urethra contouring uncertainty introduced by cross-modality registration and sub-optimal imaging contrast may lead to significant treatment degradation when urethral sparing is implemented to minimize genitourinary toxicity.
- 13 July 2021
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 86 (5) , 2666-2683
Temporally aware volumetric generative adversarial network‐based MR image reconstruction with simultaneous respiratory motion compensation: Initial feasibility in 3D dynamic cine cardiac MRI
Purpose: Develop a novel three‐dimensional (3D) generative adversarial network (GAN)‐based technique for simultaneous image reconstruction and respiratory motion compensation of 4D MRI. Our goal was to enable high‐acceleration factors 10.7X‐15.8X, while maintaining robust and diagnostic image quality superior to state‐of‐the‐art self‐gating (SG) compressed sensing wavelet (CS‐WV) reconstruction at lower acceleration factors 3.5X‐7.9X.Methods: Our GAN was trained based on pixel‐wise content loss functions, adversarial loss function, and a novel data‐driven temporal aware loss function to maintain anatomical accuracy and temporal coherence. Besides image reconstruction, our network also performs respiratory motion compensation for free‐breathing scans. A novel progressive growing‐based strategy was adapted to make the training process possible for the proposed GAN‐based structure. The proposed method was developed and thoroughly evaluated qualitatively and quantitatively based on 3D cardiac cine data from 42 patients.Results: Our proposed method achieved significantly better scores in general image quality and image artifacts at 10.7X‐15.8X acceleration than the SG CS‐WV approach at 3.5X‐7.9X acceleration (4.53 ± 0.540 vs. 3.13 ± 0.681 for general image quality, 4.12 ± 0.429 vs. 2.97 ± 0.434 for image artifacts, P < .05 for both). No spurious anatomical structures were observed in our images. The proposed method enabled similar cardiac‐function quantification as conventional SG CS‐WV. The proposed method achieved faster central processing unit‐based image reconstruction (6 s/cardiac phase) than the SG CS‐WV (312 s/cardiac phase).Conclusion: The proposed method showed promising potential for high‐resolution (1 mm3) free‐breathing 4D MR data acquisition with simultaneous respiratory motion compensation and fast reconstruction time.
- 13 July 2021
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 86 (5) , 2666-2683
Temporally aware volumetric generative adversarial network‐based MR image reconstruction with simultaneous respiratory motion compensation: Initial feasibility in 3D dynamic cine cardiac MRI
Purpose: Develop a novel three‐dimensional (3D) generative adversarial network (GAN)‐based technique for simultaneous image reconstruction and respiratory motion compensation of 4D MRI. Our goal was to enable high‐acceleration factors 10.7X‐15.8X, while maintaining robust and diagnostic image quality superior to state‐of‐the‐art self‐gating (SG) compressed sensing wavelet (CS‐WV) reconstruction at lower acceleration factors 3.5X‐7.9X.Methods: Our GAN was trained based on pixel‐wise content loss functions, adversarial loss function, and a novel data‐driven temporal aware loss function to maintain anatomical accuracy and temporal coherence. Besides image reconstruction, our network also performs respiratory motion compensation for free‐breathing scans. A novel progressive growing‐based strategy was adapted to make the training process possible for the proposed GAN‐based structure. The proposed method was developed and thoroughly evaluated qualitatively and quantitatively based on 3D cardiac cine data from 42 patients.Results: Our proposed method achieved significantly better scores in general image quality and image artifacts at 10.7X‐15.8X acceleration than the SG CS‐WV approach at 3.5X‐7.9X acceleration (4.53 ± 0.540 vs. 3.13 ± 0.681 for general image quality, 4.12 ± 0.429 vs. 2.97 ± 0.434 for image artifacts, P < .05 for both). No spurious anatomical structures were observed in our images. The proposed method enabled similar cardiac‐function quantification as conventional SG CS‐WV. The proposed method achieved faster central processing unit‐based image reconstruction (6 s/cardiac phase) than the SG CS‐WV (312 s/cardiac phase).Conclusion: The proposed method showed promising potential for high‐resolution (1 mm3) free‐breathing 4D MR data acquisition with simultaneous respiratory motion compensation and fast reconstruction time.
- 31 May 2021
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 86 (4) , 2034-2048
Slice encoding for the reduction of outflow signal artifacts in cine balanced SSFP imaging
Purpose Standard balanced SSFP (bSSFP) cine MRI often suffers from blood outflow artifacts. We propose a method that spatially encodes these outflowing spins to reduce their effects in the intended slice. Methods Bloch simulations were performed to characterize through-plane flow and to investigate how the use of phase encoding along the slice select’s direction (“slice encoding”) could alleviate its issues. Phantom scans and in vivo cines were acquired on a 3T system, comparing the standard 2D acquisition to the proposed slice-encoding method. Nineteen healthy volunteers were recruited for short-axis and horizontal long-axis oriented scans. An expert radiologist evaluated each slice-encoded/standard cine pairs in a rank comparison test and graded their quality on a 1-5 scale. The grades were used for a nonparametric paired evaluation for independent samples with a null hypothesis that there was no statistical difference between the two quality-grade distributions for α = 0.05 significance. Results Bloch simulation results demonstrated this technique’s feasibility, showing a fully resolved slice profile given a sufficient number of slice encodes. These results were confirmed with the phantom experiments. Each in vivo slice-encoded cine had a higher quality than its corresponding standard acquisition. The nonparametric paired evaluation came to 0.01 significance, encouraging us to reject the null hypothesis and conclude that slice-encoding effectively works in reducing outflow effects. Conclusion The slice-encoding balanced SSFP technique is helpful in mitigating outflow effects and is achievable within a single breath hold, being a useful alternative for cases in which the flow artifacts are significant.
- 31 May 2021
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 86 (4) , 2034-2048
Slice encoding for the reduction of outflow signal artifacts in cine balanced SSFP imaging
Purpose Standard balanced SSFP (bSSFP) cine MRI often suffers from blood outflow artifacts. We propose a method that spatially encodes these outflowing spins to reduce their effects in the intended slice. Methods Bloch simulations were performed to characterize through-plane flow and to investigate how the use of phase encoding along the slice select’s direction (“slice encoding”) could alleviate its issues. Phantom scans and in vivo cines were acquired on a 3T system, comparing the standard 2D acquisition to the proposed slice-encoding method. Nineteen healthy volunteers were recruited for short-axis and horizontal long-axis oriented scans. An expert radiologist evaluated each slice-encoded/standard cine pairs in a rank comparison test and graded their quality on a 1-5 scale. The grades were used for a nonparametric paired evaluation for independent samples with a null hypothesis that there was no statistical difference between the two quality-grade distributions for α = 0.05 significance. Results Bloch simulation results demonstrated this technique’s feasibility, showing a fully resolved slice profile given a sufficient number of slice encodes. These results were confirmed with the phantom experiments. Each in vivo slice-encoded cine had a higher quality than its corresponding standard acquisition. The nonparametric paired evaluation came to 0.01 significance, encouraging us to reject the null hypothesis and conclude that slice-encoding effectively works in reducing outflow effects. Conclusion The slice-encoding balanced SSFP technique is helpful in mitigating outflow effects and is achievable within a single breath hold, being a useful alternative for cases in which the flow artifacts are significant.
- 30 November 2020
- Published by Wiley in NMR in Biomedicine
- Vol. 34 (2) , e4433
Retrospective respiratory motion correction in cardiac cine MRI reconstruction using adversarial autoencoder and unsupervised learning
The aim of this study was to develop a deep neural network for respiratory motion compensation in free‐breathing cine MRI and evaluate its performance. An adversarial autoencoder network was trained using unpaired training data from healthy volunteers and patients who underwent clinically indicated cardiac MRI examinations. A U‐net structure was used for the encoder and decoder parts of the network and the code space was regularized by an adversarial objective. The autoencoder learns the identity map for the free‐breathing motion‐corrupted images and preserves the structural content of the images, while the discriminator, which interacts with the output of the encoder, forces the encoder to remove motion artifacts. The network was first evaluated based on data that were artificially corrupted with simulated rigid motion with regard to motion‐correction accuracy and the presence of any artificially created structures. Subsequently, to demonstrate the feasibility of the proposed approach in vivo, our network was trained on respiratory motion‐corrupted images in an unpaired manner and was tested on volunteer and patient data. In the simulation study, mean structural similarity index scores for the synthesized motion‐corrupted images and motion‐corrected images were 0.76 and 0.93 (out of 1), respectively. The proposed method increased the Tenengrad focus measure of the motion‐corrupted images by 12% in the simulation study and by 7% in the in vivo study. The average overall subjective image quality scores for the motion‐corrupted images, motion‐corrected images and breath‐held images were 2.5, 3.5 and 4.1 (out of 5.0), respectively. Nonparametric‐paired comparisons showed that there was significant difference between the image quality scores of the motion‐corrupted and breath‐held images (P < .05); however, after correction there was no significant difference between the image quality scores of the motion‐corrected and breath‐held images. This feasibility study demonstrates the potential of an adversarial autoencoder network for correcting respiratory motion‐related image artifacts without requiring paired data.
- 30 November 2020
- Published by Wiley in NMR in Biomedicine
- Vol. 34 (2) , e4433
Retrospective respiratory motion correction in cardiac cine MRI reconstruction using adversarial autoencoder and unsupervised learning
The aim of this study was to develop a deep neural network for respiratory motion compensation in free‐breathing cine MRI and evaluate its performance. An adversarial autoencoder network was trained using unpaired training data from healthy volunteers and patients who underwent clinically indicated cardiac MRI examinations. A U‐net structure was used for the encoder and decoder parts of the network and the code space was regularized by an adversarial objective. The autoencoder learns the identity map for the free‐breathing motion‐corrupted images and preserves the structural content of the images, while the discriminator, which interacts with the output of the encoder, forces the encoder to remove motion artifacts. The network was first evaluated based on data that were artificially corrupted with simulated rigid motion with regard to motion‐correction accuracy and the presence of any artificially created structures. Subsequently, to demonstrate the feasibility of the proposed approach in vivo, our network was trained on respiratory motion‐corrupted images in an unpaired manner and was tested on volunteer and patient data. In the simulation study, mean structural similarity index scores for the synthesized motion‐corrupted images and motion‐corrected images were 0.76 and 0.93 (out of 1), respectively. The proposed method increased the Tenengrad focus measure of the motion‐corrupted images by 12% in the simulation study and by 7% in the in vivo study. The average overall subjective image quality scores for the motion‐corrupted images, motion‐corrected images and breath‐held images were 2.5, 3.5 and 4.1 (out of 5.0), respectively. Nonparametric‐paired comparisons showed that there was significant difference between the image quality scores of the motion‐corrupted and breath‐held images (P < .05); however, after correction there was no significant difference between the image quality scores of the motion‐corrected and breath‐held images. This feasibility study demonstrates the potential of an adversarial autoencoder network for correcting respiratory motion‐related image artifacts without requiring paired data.
- 16 May 2020
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 84 (5) , 2831-2845
Fast and accurate calculation of myocardial T1 and T2 values using deep learning Bloch equation simulations (DeepBLESS)
Purpose: To propose and evaluate a deep learning model for rapid and accurate calculation of myocardial T1/T2 values based on a previously proposed Bloch equation simulation with slice profile correction (BLESSPC) method.Methods: Deep learning Bloch equation simulations (DeepBLESS) models are proposed for rapid and accurate T1 estimation for the MOLLI T1 mapping sequence with balanced SSFP readouts and T1/T2 estimation for a radial simultaneous T1 and T2 mapping (radial T1‐T2) sequence. The DeepBLESS models were trained separately based on simulated radial T1‐T2 and MOLLI data, respectively. The DeepBLESS T1‐T2 estimation accuracy was evaluated based on simulated data with different noise levels. The DeepBLESS model was compared with BLESSPC in simulation, phantom, and in vivo studies for the MOLLI sequence at 1.5 T and radial T1‐T2 sequence at 3 T.Results: After DeepBLESS was trained, in phantom studies, DeepBLESS and BLESSPC achieved similar accuracy and precision in T1‐T2 estimations for both MOLLI and radial T1‐T2 (P > .05). For in vivo, DeepBLESS and BLESSPC generated similar myocardial T1/T2 values for radial T1‐T2 at 3 T (T1: 1366 ± 31 ms for both methods, P > .05; T2: 37.4 ms ± 0.9 ms for both methods, P > .05), and similar myocardial T1 values for the MOLLI sequence at 1.5 T (1044 ± 20 ms for both methods, P > .05). DeepBLESS generated a T1/T2 map in less than 1 second.Conclusion: The DeepBLESS model offers an almost instantaneous approach for estimating accurate T1/T2 values, replacing BLESSPC for both MOLLI and radial T1‐T2 sequences, and is promising for multiparametric mapping in cardiac MRI.
- 16 May 2020
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 84 (5) , 2831-2845
Fast and accurate calculation of myocardial T1 and T2 values using deep learning Bloch equation simulations (DeepBLESS)
Purpose: To propose and evaluate a deep learning model for rapid and accurate calculation of myocardial T1/T2 values based on a previously proposed Bloch equation simulation with slice profile correction (BLESSPC) method.Methods: Deep learning Bloch equation simulations (DeepBLESS) models are proposed for rapid and accurate T1 estimation for the MOLLI T1 mapping sequence with balanced SSFP readouts and T1/T2 estimation for a radial simultaneous T1 and T2 mapping (radial T1‐T2) sequence. The DeepBLESS models were trained separately based on simulated radial T1‐T2 and MOLLI data, respectively. The DeepBLESS T1‐T2 estimation accuracy was evaluated based on simulated data with different noise levels. The DeepBLESS model was compared with BLESSPC in simulation, phantom, and in vivo studies for the MOLLI sequence at 1.5 T and radial T1‐T2 sequence at 3 T.Results: After DeepBLESS was trained, in phantom studies, DeepBLESS and BLESSPC achieved similar accuracy and precision in T1‐T2 estimations for both MOLLI and radial T1‐T2 (P > .05). For in vivo, DeepBLESS and BLESSPC generated similar myocardial T1/T2 values for radial T1‐T2 at 3 T (T1: 1366 ± 31 ms for both methods, P > .05; T2: 37.4 ms ± 0.9 ms for both methods, P > .05), and similar myocardial T1 values for the MOLLI sequence at 1.5 T (1044 ± 20 ms for both methods, P > .05). DeepBLESS generated a T1/T2 map in less than 1 second.Conclusion: The DeepBLESS model offers an almost instantaneous approach for estimating accurate T1/T2 values, replacing BLESSPC for both MOLLI and radial T1‐T2 sequences, and is promising for multiparametric mapping in cardiac MRI.
- 23 April 2020
- Published by Wiley in Medical Physics
- Vol. 47 (8) , 3511-3519
3D isotropic resolution diffusion‐prepared magnitude‐stabilized bSSFP imaging with high geometric fidelity at 1.5 Tesla
Introduction: MRI has been increasingly used in radiation therapy to facilitate tumor and organ delineation and assess treatment response. Diffusion MRI can provide cellularity information and may enable functional‐based treatment planning and adaptation. However, strong distortion associated with the conventional diffusion‐weighted single‐shot echo‐planar imaging (DW‐ssEPI) sequence is problematic for accurate target delineation. The goal of this work is to propose a 3D diffusion sequence with minimal distortion for radiation therapy applications.Methods: A 3D diffusion‐prepared magnitude‐stabilized balanced steady‐state free precession sequence (DP‐MS‐bSSFP) was developed. A 2D navigator was acquired during the linear catalyzation stage of the bSSFP readout to estimate the phase, which was then used in a plane‐by‐plane low‐rank constrained reconstruction to correct the shot‐to‐shot k‐space inconsistency. A diffusion phantom was scanned to evaluate and compare the geometric reliability and apparent diffusion coefficient (ADC) accuracy with the conventional DW‐ssEPI. Eight landmarks were selected on each slice of the images to calculate the target registration error (TRE), which was used as a surrogate for geometric fidelity. The phantom was scanned under both 0℃ and room temperature. Brain scans were performed on five healthy volunteers. In the first volunteer, protocols of 1, 2, and 4 shots per Kz plane were compared. In vivo geometric fidelity and ADC accuracy were evaluated on the remaining four volunteers using the protocol of four shots per Kz plane. In the geometric fidelity study, 8–10 landmarks were picked on each slice to calculate the TRE. Regions of interest were placed on the white matter, the cerebellum, and the cerebrospinal fluid region to evaluate the ADC agreement between DW‐ssEPI and DP‐MS using the Bland‐Altman plot. All scans were performed at 1.5 mm isotropic resolution to meet the high‐resolution requirement of many radiotherapy applications.Results: The DP‐MS had drastically improved geometric accuracy compared with DW‐ssEPI on the phantom. The mean TRE decreased from 2.09 mm to 0.70 mm. The percentage difference of the ADC values between the two diffusion sequences were less than 5.5% and 7% for the 0℃ and room temperature study, respectively. The DW‐ssEPI had strong distortion and susceptibility‐related artifacts at tissue air boundary, whereas distortion was minimal in DP‐MS images. Overall, the mean/max TRE was over 2 mm/7 mm in the volunteers for DW‐ssEPI, whereas less than 0.8 mm/2 mm for DP‐MS. Good ADC agreement was observed for the white matter, the cerebellum, and the CSF based on the Bland‐Altman plots.Conclusion: A 3D diffusion sequence was developed and validated. It provided high‐resolution diffusion imaging with mean distortion less than 1 mm at 1.5 T, and is a promising imaging technique for treatment planning and adaptive radiotherapy.
- 23 April 2020
- Published by Wiley in Medical Physics
- Vol. 47 (8) , 3511-3519
3D isotropic resolution diffusion‐prepared magnitude‐stabilized bSSFP imaging with high geometric fidelity at 1.5 Tesla
Introduction: MRI has been increasingly used in radiation therapy to facilitate tumor and organ delineation and assess treatment response. Diffusion MRI can provide cellularity information and may enable functional‐based treatment planning and adaptation. However, strong distortion associated with the conventional diffusion‐weighted single‐shot echo‐planar imaging (DW‐ssEPI) sequence is problematic for accurate target delineation. The goal of this work is to propose a 3D diffusion sequence with minimal distortion for radiation therapy applications.Methods: A 3D diffusion‐prepared magnitude‐stabilized balanced steady‐state free precession sequence (DP‐MS‐bSSFP) was developed. A 2D navigator was acquired during the linear catalyzation stage of the bSSFP readout to estimate the phase, which was then used in a plane‐by‐plane low‐rank constrained reconstruction to correct the shot‐to‐shot k‐space inconsistency. A diffusion phantom was scanned to evaluate and compare the geometric reliability and apparent diffusion coefficient (ADC) accuracy with the conventional DW‐ssEPI. Eight landmarks were selected on each slice of the images to calculate the target registration error (TRE), which was used as a surrogate for geometric fidelity. The phantom was scanned under both 0℃ and room temperature. Brain scans were performed on five healthy volunteers. In the first volunteer, protocols of 1, 2, and 4 shots per Kz plane were compared. In vivo geometric fidelity and ADC accuracy were evaluated on the remaining four volunteers using the protocol of four shots per Kz plane. In the geometric fidelity study, 8–10 landmarks were picked on each slice to calculate the TRE. Regions of interest were placed on the white matter, the cerebellum, and the cerebrospinal fluid region to evaluate the ADC agreement between DW‐ssEPI and DP‐MS using the Bland‐Altman plot. All scans were performed at 1.5 mm isotropic resolution to meet the high‐resolution requirement of many radiotherapy applications.Results: The DP‐MS had drastically improved geometric accuracy compared with DW‐ssEPI on the phantom. The mean TRE decreased from 2.09 mm to 0.70 mm. The percentage difference of the ADC values between the two diffusion sequences were less than 5.5% and 7% for the 0℃ and room temperature study, respectively. The DW‐ssEPI had strong distortion and susceptibility‐related artifacts at tissue air boundary, whereas distortion was minimal in DP‐MS images. Overall, the mean/max TRE was over 2 mm/7 mm in the volunteers for DW‐ssEPI, whereas less than 0.8 mm/2 mm for DP‐MS. Good ADC agreement was observed for the white matter, the cerebellum, and the CSF based on the Bland‐Altman plots.Conclusion: A 3D diffusion sequence was developed and validated. It provided high‐resolution diffusion imaging with mean distortion less than 1 mm at 1.5 T, and is a promising imaging technique for treatment planning and adaptive radiotherapy.
- 1 September 2019
- Published by AME Publishing Company in Quantitative Imaging in Medicine and Surgery
- Vol. 9 (9) , 1516-1527
MR image reconstruction using deep learning: evaluation of network structure and loss functions
Background: To review and evaluate approaches to convolutional neural network (CNN) reconstruction for accelerated cardiac MR imaging in the real clinical context. Methods: Two CNN architectures, Unet and residual network (Resnet) were evaluated using quantitative and qualitative assessment by radiologist. Four different loss functions were also considered: pixel-wise (L1 and L2), patch-wise structural dissimilarity (Dssim) and feature-wise (perceptual loss). The networks were evaluated using retrospectively and prospectively under-sampled cardiac MR data. Results: Based on our assessments, we find that Resnet and Unet achieve similar image quality but that former requires only 100,000 parameters compared to 1.3 million parameters for the latter. The perceptual loss function performed significantly better than L1, L2 or Dssim loss functions as determined by the radiologist scores. Conclusions: CNN image reconstruction using Resnet yields comparable image quality to Unet with 10X the number of parameters. This has implications for training with significantly lower data requirements. Network training using the perceptual loss function was found to better agree with radiologist scoring compared to L1, L2 or Dssim loss functions.
- 1 September 2019
- Published by AME Publishing Company in Quantitative Imaging in Medicine and Surgery
- Vol. 9 (9) , 1516-1527
MR image reconstruction using deep learning: evaluation of network structure and loss functions
Background: To review and evaluate approaches to convolutional neural network (CNN) reconstruction for accelerated cardiac MR imaging in the real clinical context. Methods: Two CNN architectures, Unet and residual network (Resnet) were evaluated using quantitative and qualitative assessment by radiologist. Four different loss functions were also considered: pixel-wise (L1 and L2), patch-wise structural dissimilarity (Dssim) and feature-wise (perceptual loss). The networks were evaluated using retrospectively and prospectively under-sampled cardiac MR data. Results: Based on our assessments, we find that Resnet and Unet achieve similar image quality but that former requires only 100,000 parameters compared to 1.3 million parameters for the latter. The perceptual loss function performed significantly better than L1, L2 or Dssim loss functions as determined by the radiologist scores. Conclusions: CNN image reconstruction using Resnet yields comparable image quality to Unet with 10X the number of parameters. This has implications for training with significantly lower data requirements. Network training using the perceptual loss function was found to better agree with radiologist scoring compared to L1, L2 or Dssim loss functions.
- 28 August 2019
- Published by Wiley in NMR in Biomedicine
- Vol. 32 (11) , e4165
Accurate, precise, simultaneous myocardial T1 and T2 mapping using a radial sequence with inversion recovery and T2 preparation
We propose a simultaneous myocardial T1 and T2 mapping technique using a radial sequence with inversion recovery and T2 preparation, which achieves high accuracy and precision, with T1 and T2 reproducibility similar to the Modified Look‐Locker Inversion recovery (MOLLI) sequence and the conventional bright blood T2 mapping technique, respectively. The sequence was developed by incorporating gold angle radial fast low angle shot (FLASH) readout combined with an inversion pulse and T2prep pulses. The extended Bloch equation simulation with slice profile correction (BLESSPC) algorithm was proposed to reconstruct T1 and T2 maps at the same time in a few seconds, while maintaining good T1 and T2 estimation accuracy. Accuracy and precision were compared among the proposed technique, MOLLI and conventional T2 mapping techniques using phantom studies, 10 healthy volunteers and three patients. In phantom studies, the proposed technique was more accurate than MOLLI (P < 0.05) while achieving similar precision (P = 0.3) in T1 estimation, and was more accurate (P < 0.05) and precise (P < 0.001) than conventional T2 mapping (two‐parameter fitting) in T2 estimation. In vivo, the proposed technique achieved significantly higher T1 values (P < 0.001) and similar reproducibility (P = 0.3) compared with MOLLI, with significantly lower T2 values (P < 0.001) and similar reproducibility (P = 0.6) compared with the conventional T2 mapping technique. Thus, the proposed radial T1‐T2 mapping technique allows for accurate, precise, simultaneous myocardial T1 and T2 mapping in an 11‐heartbeat single breath‐hold acquisition.
- 28 August 2019
- Published by Wiley in NMR in Biomedicine
- Vol. 32 (11) , e4165
Accurate, precise, simultaneous myocardial T1 and T2 mapping using a radial sequence with inversion recovery and T2 preparation
We propose a simultaneous myocardial T1 and T2 mapping technique using a radial sequence with inversion recovery and T2 preparation, which achieves high accuracy and precision, with T1 and T2 reproducibility similar to the Modified Look‐Locker Inversion recovery (MOLLI) sequence and the conventional bright blood T2 mapping technique, respectively. The sequence was developed by incorporating gold angle radial fast low angle shot (FLASH) readout combined with an inversion pulse and T2prep pulses. The extended Bloch equation simulation with slice profile correction (BLESSPC) algorithm was proposed to reconstruct T1 and T2 maps at the same time in a few seconds, while maintaining good T1 and T2 estimation accuracy. Accuracy and precision were compared among the proposed technique, MOLLI and conventional T2 mapping techniques using phantom studies, 10 healthy volunteers and three patients. In phantom studies, the proposed technique was more accurate than MOLLI (P < 0.05) while achieving similar precision (P = 0.3) in T1 estimation, and was more accurate (P < 0.05) and precise (P < 0.001) than conventional T2 mapping (two‐parameter fitting) in T2 estimation. In vivo, the proposed technique achieved significantly higher T1 values (P < 0.001) and similar reproducibility (P = 0.3) compared with MOLLI, with significantly lower T2 values (P < 0.001) and similar reproducibility (P = 0.6) compared with the conventional T2 mapping technique. Thus, the proposed radial T1‐T2 mapping technique allows for accurate, precise, simultaneous myocardial T1 and T2 mapping in an 11‐heartbeat single breath‐hold acquisition.
- 17 June 2019
- Published by Wiley in Medical Physics
- Vol. 46 (8) , 3399-3413
Parallel imaging and convolutional neural network combined fast MR image reconstruction: Applications in low‐latency accelerated real‐time imaging
Purpose: To develop and evaluate a parallel imaging and convolutional neural network combined image reconstruction framework for low‐latency and high‐quality accelerated real‐time MR imaging.Methods: Conventional Parallel Imaging reconstruction resolved as gradient descent steps was compacted as network layers and interleaved with convolutional layers in a general convolutional neural network. All parameters of the network were determined during the offline training process, and applied to unseen data once learned. The proposed network was first evaluated for real‐time cardiac imaging at 1.5 T and real‐time abdominal imaging at 0.35 T, using threefold to fivefold retrospective undersampling for cardiac imaging and threefold retrospective undersampling for abdominal imaging. Then, prospective undersampling with fourfold acceleration was performed on cardiac imaging to compare the proposed method with standard clinically available GRAPPA method and the state‐of‐the‐art L1‐ESPIRiT method.Results: Both retrospective and prospective evaluations confirmed that the proposed network was able to images with a lower noise level and reduced aliasing artifacts in comparison with the single‐coil based and L1‐ESPIRiT reconstructions for cardiac imaging at 1.5 T, and the GRAPPA and L1‐ESPIRiT reconstructions for abdominal imaging at 0.35 T. Using the proposed method, each frame can be reconstructed in less than 100 ms, suggesting its clinical compatibility.Conclusion: The proposed Parallel Imaging and convolutional neural network combined reconstruction framework is a promising technique that allows low‐latency and high‐quality real‐time MR imaging.
- 17 June 2019
- Published by Wiley in Medical Physics
- Vol. 46 (8) , 3399-3413
Parallel imaging and convolutional neural network combined fast MR image reconstruction: Applications in low‐latency accelerated real‐time imaging
Purpose: To develop and evaluate a parallel imaging and convolutional neural network combined image reconstruction framework for low‐latency and high‐quality accelerated real‐time MR imaging.Methods: Conventional Parallel Imaging reconstruction resolved as gradient descent steps was compacted as network layers and interleaved with convolutional layers in a general convolutional neural network. All parameters of the network were determined during the offline training process, and applied to unseen data once learned. The proposed network was first evaluated for real‐time cardiac imaging at 1.5 T and real‐time abdominal imaging at 0.35 T, using threefold to fivefold retrospective undersampling for cardiac imaging and threefold retrospective undersampling for abdominal imaging. Then, prospective undersampling with fourfold acceleration was performed on cardiac imaging to compare the proposed method with standard clinically available GRAPPA method and the state‐of‐the‐art L1‐ESPIRiT method.Results: Both retrospective and prospective evaluations confirmed that the proposed network was able to images with a lower noise level and reduced aliasing artifacts in comparison with the single‐coil based and L1‐ESPIRiT reconstructions for cardiac imaging at 1.5 T, and the GRAPPA and L1‐ESPIRiT reconstructions for abdominal imaging at 0.35 T. Using the proposed method, each frame can be reconstructed in less than 100 ms, suggesting its clinical compatibility.Conclusion: The proposed Parallel Imaging and convolutional neural network combined reconstruction framework is a promising technique that allows low‐latency and high‐quality real‐time MR imaging.