Direct abdominal vein thrombus imaging (DATI): a contrast-free black-blood MR technique for the diagnosis of abdominal vein thrombosis.

OBJECTIVES: To develop and evaluate a direct abdominal vein thrombus imaging (DATI) technique, based on a respiratory navigating SPACE sequence with DANTE black-blood preparation, for diagnosing abdominal vein thrombosis (AVT) without the use of exogenous contrast agents. METHODS: We prospectively enrolled 10 healthy subjects and 28 suspected AVT patients who underwent DATI scans on 3.0 T MRI. Contrast-enhanced CT venography (CTV) was also conducted on the suspected AVT patients for comparison. All images were analyzed by two blinded radiologists who independently evaluated randomized images and gave image quality and diagnostic confidence scores (1-poor, 4-excellent) for DATI and CTV. The accuracy (ACC), sensitivity (SE), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) of CTV were calculated using CTV as a standard reference. The diagnostic agreement between DATI and CTV as well as the interobserver agreement were conducted using Cohen κ test. RESULTS: The patient study demonstrated that DATI can provide adequate thrombus signal intensity and the contrast between the thrombus to dark venous lumen for the diagnosis of AVT. It offers good to excellent image quality (reader1/reader2: 3.50 ± 0.64/3.42 ± 0.63, κ = 0.872) and diagnostic confidence (reader1/reader2: 3.71 ± 0.53/3.78 ± 0.42, κ = 0.804) for the diagnosis of AVT. Taking CTV as a reference, DATI has high accuracy (96.6%), SE (91.5%), SP (98.0%), PPV (92.3%), and NPV (97.8%). DATA CONCLUSION: DATI can provide good to excellent image quality, effective venous blood signal suppression, and definitive thrombus detection for the diagnosis of AVT without the use of exogenous contrast agents.

Advances in volumetric bioprinting.

The three-dimensional (3D) bioprinting technologies are suitable for biomedical applications owing to their ability to manufacture complex and high-precision tissue constructs. However, the slow printing speed of current layer-by-layer (bio)printing modality is the major limitation in biofabrication field. To overcome this issue, volumetric bioprinting (VBP) is developed. VBP changes the layer-wise operation of conventional devices, permitting the creation of geometrically complex, centimeter-scale constructs in tens of seconds. VBP is the next step onward from sequential biofabrication methods, opening new avenues for fast additive manufacturing in the fields of tissue engineering, regenerative medicine, personalized drug testing, and soft robotics, etc. Therefore, this review introduces the printing principles and hardware designs of VBP-based techniques; then focuses on the recent advances in VBP-based (bio)inks and their biomedical applications. Lastly, the current limitations of VBP are discussed together with future direction of research.

Black-Blood Magnetization Prepared 2 Rapid Acquisition Gradient Echoes: A Fast and Three-Dimensional MR Black-Blood T1 Mapping Technique for Quantitative Assessment of Atherosclerosis and Venous Thrombosis.

BACKGROUND: Blood flow signals may be a confounder in quantifying T1 values of plaque or thrombus and how to realize black-blood T1 mapping remains a challenge task. PURPOSE: To develop a fast and three-dimensional black-blood T1 mapping technique for quantitative assessment of atherosclerosis and venous thrombosis. STUDY TYPE: Sequence development and optimization via phantoms and volunteers as well as pilot prospective. PHANTOM AND SUBJECTS: Numerical simulations, a standard phantom, 8 healthy volunteers (mean age, 22 ± 1 years; 5 males), and 19 patients (mean age, 57 ± 14 years; 13 males) with atherosclerosis or venous thrombosis. FIELD STRENGTH/SEQUENCE: 3T/inversion recovery spin-echo sequence (IR-SE), magnetization prepared 2 rapid acquisition gradient echoes (MP2RAGE), and black-blood prepared MP2RAGE (BB-MP2RAGE). ASSESSMENT: The black-blood preparation (i.e., delay alternating with nutation for tailored excitation, DANTE) was incorporated into MP2RAGE for black-blood T1 mapping. The BB-MP2RAGE was optimized numerically based on the Bloch equation, and then the phantom study was performed to verify the accuracy of T1 mapping by BB-MP2RAGE against IR-SE and MP2RAGE. Preliminary clinical validation was prospectively performed to assess the flow suppression effect and its potential application in plaque and thrombosis identification. STATISTICAL TESTS: Pearson correlation test, Bland-Altman analysis, paired t-test, and intraclass correlation coefficient. A P value <0.05 indicates a statistically significant difference. RESULTS: Phantom experiments showed comparable accuracy of T1 maps by BB-MP2RAGE with IR-SE and MP2RAGE (all r2 > 0.99); Compared to MP2RAGE, BB-MP2RAGE effectively nulled the blood flow signals, and had a significant improvement in contrast-to-noise ratio between static tissue and blood (250.5 ± 66.6 vs. 91.9 ± 35.9). BB-MP2RAGE can quantify plaque or thrombus T1 relaxation time with blood flow signal suppression. DATA CONCLUSION: Accurate T1 mapping with sufficient blood flow suppression was achieved by BB-MP2RAGE. BB-MP2RAGE has the potential to quantitatively characterize atherosclerosis and venous thrombosis. LEVEL OF EVIDENCE: 1 TECHNICAL EFFICACY: Stage 1.

Cross-Domain Unpaired Learning for Low-Dose CT Imaging.

Supervised deep-learning techniques with paired training datasets have been widely studied for low-dose computed tomography (LDCT) imaging with excellent performance. However, the paired training datasets are usually difficult to obtain in clinical routine, which restricts the wide adoption of supervised deep-learning techniques in clinical practices. To address this issue, a general idea is to construct a pseudo paired training dataset based on the widely available unpaired data, after which, supervised deep-learning techniques can be adopted for improving the LDCT imaging performance by training on the pseudo paired training dataset. However, due to the complexity of noise properties in CT imaging, the LDCT data are difficult to generate in order to construct the pseudo paired training dataset. In this article, we propose a simple yet effective cross-domain unpaired learning framework for pseudo LDCT data generation and LDCT image reconstruction, which is denoted as CrossDuL. Specifically, a dedicated pseudo LDCT sinogram generative module is constructed based on a data-dependent noise model in the sinogram domain, and then instead of in the sinogram domain, a pseudo paired dataset is constructed in the image domain to train an LDCT image restoration module. To validate the effectiveness of the proposed framework, clinical datasets are adopted. Experimental results demonstrate that the CrossDuL framework can obtain promising LDCT imaging performance in both quantitative and qualitative measurements.

Thrombus Signal on T1-Weighted Black-Blood MR Predicts Outcomes of Catheter-Directed Thrombolysis in Acute Deep Vein Thrombosis.

OBJECTIVES: Catheter-directed thrombolysis (CDT) is an effective therapy for acute deep vein thrombosis (DVT). However, predicting the CDT outcomes remains elusive. We hypothesized that the thrombus signal on T1-weighted black-blood magnetic resonance (MR) can provide insight into CDT outcomes in acute DVT patients. METHODS: A total of 117 patients with acute iliofemoral DVT were enrolled for T1-weighted black-blood MR before CDT in this prospective study. Based on the signal contrast between thrombus and adjacent muscle, patients were categorized into the iso-intense thrombus (Iso-IT), hyper-intense thrombus (Hyper-IT), and mixed iso-/hyper-intense thrombi (Mixed-IT) groups. Immediate treatment outcome (i.e., vein patency) and long-term treatment outcome (i.e., the incidence rate of postthrombotic syndrome) were accessed by the same expert. Histological analysis and iron quantification were performed on thrombus samples to characterize the content of fibrin, collagen, and the ratio of Fe3+ to total iron. RESULTS: Compared to Mixed-IT and Hyper-IT groups, the Iso-IT group had the best lytic effect (90.5 ± 1.6% vs. 78.4 ± 2.6% vs. 46.5 ± 3.3%, p < 0.001), lowest bleeding ratio (0.0 vs. 11.8 vs. 13.3, p < 0.001), and the lowest incidence rate of postthrombotic syndrome on 24 months (3.6 vs. 18.4 vs. 63.4%, p < 0.001) following CDT. The Iso-IT group had a significantly lower ratio of Fe3+ to total iron (93.1 ± 3.2% vs. 97.2 ± 2.1%, p = 0.034) and a higher content of fibrin (12.5 ± 5.3% vs. 4.76 ± 3.18%, p = 0.023) than Hyper-IT. CONCLUSION: Thrombus signal characteristics on T1-weighted black-blood MR is associated with CDT outcomes and possesses potential to serve as a noninvasive approach to guide treatment decision making in acute DVT patients. KEY POINTS: · Thrombus signal on T1-weighted black-blood MR is associated with lytic therapeutic outcome in acute DVT patients.. · Presence of iso-intense thrombus revealed by T1-weighted black-blood MRI is associated with successful thrombolysis, low bleeding ratio, and low incidence of the postthrombotic syndrome.. · T1-weighted thrombus signal characteristics may serve as a noninvasive imaging marker to predict CDT treatment outcomes and therefore guide treatment decision making in acute DVT patients..

Water excitation with LIBRE pulses in three-dimensional variable flip angle fast spin echo for fat-free and large field of view imaging at 3 tesla.

PURPOSE: To develop and evaluate a sequence in which water excitation with lipid insensitive binomial off-resonant radio frequency excitation (LIBRE) pulses is incorporated into three-dimensional (3D) variable flip angle fast spin echo (LIBRE-vf-FSE) for fat-free and large field of view imaging at 3 Tesla (T). MATERIALS AND METHODS: Numerical simulation was conducted to optimize the parameters of LIBRE pulses, including the flip angle, pulse duration, and frequency offset, for maximizing the fat suppression effect of the proposed LIBRE-vf-FSE sequence. The sequence was then implemented at 3 T and assessed in phantoms, lower extremity imaging of 8 healthy volunteers, and head/neck imaging of 5 healthy volunteers. Conventional water excitation (WE) and fat saturation (FatSat) were also performed for comparison. Signal-to-noise ratio (SNR) of fat and contrast-to-noise ratio (CNR) between fat and water were used to evaluate the level of fat suppression. Standard deviation (SD) of SNR was used to evaluate the uniformity of fat suppression. RESULTS: The numerical simulation demonstrated that LIBRE-vf-FSE enables large volume imaging with uniform fat suppression, which was further confirmed by phantom and healthy volunteer experiments. LIBRE provided the lowest fat SNR and offered more uniform fat suppression compared with the WE and FatSat. Specifically, average oil SNRs obtained by LIBRE (1.10 ms, 360 Hz, and 60°), WE, and FatSat were (180.1 vs. 280.2 vs. 811.2) in phantom experiments, and average fat SNRs and SDs in legs obtained by LIBRE (1.10 ms, 360 Hz, and 60°), WE, and FatSat were (85.1 vs. 105.0 vs. 105.1) and (22.4 vs. 27.4 vs. 56.4) in vivo experiments, respectively. CONCLUSION: The proposed LIBRE-vf-FSE sequence allows for fat suppression and large field of view imaging at 3 T. It could be an alternative approach for fat-free vf-FSE scan.

Black-blood Venous Imaging (BBVI): A Contrast-Free and High-Resolution Magnetic Resonance Approach for Diagnosing IVCS - a Proof of Concept Study.

BACKGROUND: Iliac vein compression syndrome (IVCS) diagnosis heavily relies on an imaging test. However, non-invasive and contrast-free imaging test for the diagnosis of IVCS remains a big challenge. To address this issue, this prospective study aimed to assess the image quality and diagnostic performance of a magnetic resonance imaging technique, black-blood venous imaging (BBVI), in detecting IVCS by comparing it with contrast-enhanced computed tomography venography (CTV) and using invasive digital subtraction angiography (DSA) as the reference. METHODS: We enrolled 105 patients, and all patients underwent BBVI, CTV, and DSA examinations. We compared the consistency of CTV and BBVI image quality and their consistency in diagnosing the rate of iliac vein stenosis in IVCS patients. Using the consensus DSA as a reference, the sensitivity, specificity, positive and negative predictive values, and accuracy of BBVI and CTV and their diagnostic agreement with DSA were calculated. RESULTS: BBVI demonstrated high sensitivity, specificity, and accuracy for the diagnosis of IVCS, without contrast agents. BBVI and CTV are quite in diagnosis IVCS. Quite SE (67.8% vs 68.3%), SP (94.8% vs 94.8%), PPV (98.0% vs 98.0%), NPV (46.2% vs 46.9%) and ACC (75.3% vs 75.7%) were obtained by BBVI in comparison with CTV. CONCLUSION: BBVI has comparable diagnostic performance with CTV. It may be a viable alternative to CTV techniques in screening the IVCS without contrast agents and free of ionizing radiation.

A Cascaded Multi-Task Generative Framework for Detecting Aortic Dissection on 3-D Non-Contrast-Enhanced Computed Tomography.

Contrast-enhanced computed tomography (CE-CT) is the gold standard for diagnosing aortic dissection (AD). However, contrast agents can cause allergic reactions or renal failure in some patients. Moreover, AD diagnosis by radiologists using non-contrast-enhanced CT (NCE-CT) images has poor sensitivity. To address this issue, we propose a novel cascaded multi-task generative framework for AD detection using NCE-CT volumes. The framework includes a 3D nnU-Net and a 3D multi-task generative architecture (3D MTGA). Specifically, the 3D nnU-Net was employed to segment aortas from NCE-CT volumes. The 3D MTGA was then employed to simultaneously synthesize CE-CT volumes, segment true & false lumen, and classify the patient as AD or non-AD. A theoretical formulation demonstrated that the 3D MTGA could increase the Jensen-Shannon Divergence (JSD) between AD and non-AD for each NCE-CT volume, thus indirectly improving the AD detection performance. Experiments also showed that the proposed framework could achieve an average accuracy of 0.831, a sensitivity of 0.938, and an F1-score of 0.847 in comparison with seven state-of-the-art classification models used by three radiologists with junior, intermediate, and senior experiences, respectively. The experimental results indicate that the proposed framework obtains superior performance to state-of-the-art models in AD detection. Thus, it has great potential to reduce the misdiagnosis of AD using NCE-CT in clinical practice. The source codes and supplementary materials for our framework are available at https://github.com/yXiangXiong/CMTGF.

Knee Bone and Cartilage Segmentation Based on a 3D Deep Neural Network Using Adversarial Loss for Prior Shape Constraint.

Fast and accurate segmentation of knee bone and cartilage on MRI images is becoming increasingly important in the orthopaedic area, as the segmentation is an essential prerequisite step to a patient-specific diagnosis, optimising implant design and preoperative and intraoperative planning. However, manual segmentation is time-intensive and subjected to inter- and intra-observer variations. Hence, in this study, a three-dimensional (3D) deep neural network using adversarial loss was proposed to automatically segment the knee bone in a resampled image volume in order to enlarge the contextual information and incorporate prior shape constraints. A restoration network was proposed to further improve the bone segmentation accuracy by restoring the bone segmentation back to the original resolution. A conventional U-Net-like network was used to segment the cartilage. The ultimate results were the combination of the bone and cartilage outcomes through post-processing. The quality of the proposed method was thoroughly assessed using various measures for the dataset from the Grand Challenge Segmentation of Knee Images 2010 (SKI10), together with a comparison with a baseline network U-Net. A fine-tuned U-Net-like network can achieve state-of-the-art results without any post-processing operations. This method achieved a total score higher than 76 in terms of the SKI10 validation dataset. This method showed to be robust to extract bone and cartilage masks from the MRI dataset, even for the pathological case.

Carotid arterial wall MRI of apolipoprotein e-deficient mouse at 7 T using DANTE-prepared variable-flip-angle rapid acquisition with relaxation enhancement.

PURPOSE: To optimize a sequence combining the delay alternating with nutation for tailored excitation (DANTE) preparative module with the variable-flip-angle rapid acquisition with relaxation enhancement (VF-RARE) sequence (DANTE-VF-RARE) and to investigate its feasibility for vessel wall imaging in Apolipoprotein E-Deficient (ApoE-/-) mouse at 7 Tesla (T). MATERIALS AND METHODS: Specific T1/T2 values were used for producing a sharper vessel wall in the variable-flip-angle optimization scheme. The DANTE RF pulse flip angle and pulse train length were optimized for maximizing the wall-lumen contrast. ApoE-/- (fed high fat diet for 20/40/ 60 weeks, n = 9/4/4) and wild-type mice (controls, n = 3) were imaged at 7 T using VF-RARE, DANTE-VF-RARE, time-of-flight (TOF) angiography, and multi-slice T1-weighted 2D RARE coupled with inflow outflow saturation bands (IOSB-RARE). Wall-lumen contrast-to-noise-ratio efficiency (CNReff), lumen area (LA), and wall area (WA) were compared between DANTE-VF-RARE and 2D IOSB-RARE sequences. Additionally, linear regression analysis was conducted between MR measurements and histomorphometric planimetry results. RESULTS: Residual blood signal was observed in the four out of eighteen carotids on VF-RARE images, whereas it was significantly suppressed on DANTE-VF-RARE images. Compared with IOSB-RARE, DANTE-VF-RARE offered significantly improved CNReff (P < 0.001). The LA and WA were both comparable (P = 0.085 and 0.112, respectively) and showed excellent agreement between DANTE-VF-RARE and IOSB-RARE (ICC = 0.96 and 0.95, respectively). The luminal stenosis identified by DANTE-VF-RARE was in consistent with the results of TOF. Strong correlations were found between MR measurements and histopathological analysis for both WA (DANTE-VF-RARE: r = 0.92, slope = 0.94, P < 0.001; IOSB-RARE: r = 0.93, slope = 0.94, P < 0.001) and LA (DANTE-VF-RARE: r = 0.82, slope = 0.54, P < 0.001; IOSB-RARE: r = 0.78, slope = 0.50, P < 0.001). CONCLUSION: DANTE-VF-RARE achieves effective blood signal suppression and is a feasible approach for the 3D carotid arterial wall imaging of ApoE-/- mouse at 7 T.

A Cascaded Deep Learning Framework for Detecting Aortic Dissection Using Non-contrast Enhanced Computed Tomography.

Aortic dissection (AD) is a rare but potentially fatal disease with high mortality. The aim of this study is to synthesize contrast enhanced computed tomography (CE-CT) images from non-contrast CT (NCE-CT) images for detecting aortic dissection. In this paper, a cascaded deep learning framework containing a 3D segmentation network and a synthetic network was proposed and evaluated. A 3D segmentation network was firstly used to segment aorta from NCE-CT images and CE-CT images. A conditional generative adversarial network (CGAN) was subsequently employed to map the NCE-CT images to the CE-CT images non-linearly for the region of aorta. The results of the experiment suggest that the cascaded deep learning framework can be used for detecting the AD and outperforms CGAN alone.

Deep Learning for Accurate Segmentation of Venous Thrombus from Black-Blood Magnetic Resonance Images: A Multicenter Study.

OBJECTIVE: Deep vein thrombosis (DVT) is the third-largest cardiovascular disease, and accurate segmentation of venous thrombus from the black-blood magnetic resonance (MR) images can provide additional information for personalized DVT treatment planning. Therefore, a deep learning network is proposed to automatically segment venous thrombus with high accuracy and reliability. METHODS: In order to train, test, and external test the developed network, total images of 110 subjects are obtained from three different centers with two different black-blood MR techniques (i.e., DANTE-SPACE and DANTE-FLASH). Two experienced radiologists manually contoured each venous thrombus, followed by reediting, to create the ground truth. 5-fold cross-validation strategy is applied for training and testing. The segmentation performance is measured on pixel and vessel segment levels. For the pixel level, the dice similarity coefficient (DSC), average Hausdorff distance (AHD), and absolute volume difference (AVD) of segmented thrombus are calculated. For the vessel segment level, the sensitivity (SE), specificity (SP), accuracy (ACC), and positive and negative predictive values (PPV and NPV) are used. RESULTS: The proposed network generates segmentation results in good agreement with the ground truth. Based on the pixel level, the proposed network achieves excellent results on testing and the other two external testing sets, DSC are 0.76, 0.76, and 0.73, AHD (mm) are 4.11, 6.45, and 6.49, and AVD are 0.16, 0.18, and 0.22. On the vessel segment level, SE are 0.95, 0.93, and 0.81, SP are 0.97, 0.92, and 0.97, ACC are 0.96, 0.94, and 0.95, PPV are 0.97, 0.82, and 0.96, and NPV are 0.97, 0.96, and 0.94. CONCLUSIONS: The proposed deep learning network is effective and stable for fully automatic segmentation of venous thrombus on black blood MR images.

Time-efficient and contrast-free magnetic resonance imaging approach to the diagnosis of deep vein thrombosis on black-blood gradient-echo sequence: a pilot study.

BACKGROUND: Black-blood thrombus imaging (BTI) has shown to be advantageous for the diagnosis of deep vein thrombosis (DVT). However, previous techniques using fast spin echo have a high specific absorption rate. As DANTE (delay alternating with nutation for tailored excitation) black-blood preparation can suppress blood flows over a broad range of velocities, we hypothesized that a DANTE black-blood preparation combined with a fast low-angle shot (FLASH) gradient-echo readout-DANTE-FLASH could be used to diagnose DVT. METHODS: Eleven healthy volunteers and 30 suspected DVT patients were recruited to undergo DANTE-FLASH and magnetic resonance direct thrombus imaging (MRDTI). The suspected DVT patients were also examined by ultrasound (US). For the segment level, a total of 1,066 venous vessel segments were analyzed. Using US and MRDTI as the references, the sensitivity (SE), specificity (SP), positive and negative predictive values (PPV and NPV), and accuracy (ACC) of DANTE-FLASH were calculated. To quantitatively compare image quality between DANTE-FLASH and MRDTI, image signal-to-noise ratio (SNR), apparent contrast-to-noise ratio (CNR) between muscle and the venous lumen, and the apparent CNR between the thrombus and venous lumen were measured. Additionally, diagnostic confidence, image quality, and clot burden were also evaluated. RESULTS: Using the consensus results of US and MRDTI as a standard reference, the diagnostic SE, SP, PPV, NPV, and ACC of DANTE-FLASH for the 2 readers were 97.0% and 93.2%, 99.0% and 98.2%, 93.4% and 87.9%, 99.6% and 99.0%, and 98.8% and 97.6%, respectively. According to the image quantitative analysis results, DANTE-FLASH demonstrated higher image SNR and CNR than MRDTI. The image quality and diagnostic confidence scores of DANTE-FLASH were higher than MRDTI (3.66±0.44 vs. 3.52±0.52, P<0.001, and 3.84±0.36 vs. 3.76±0.41, P<0.001). There was excellent agreement between DANTE-FLASH and MRDTI on clot burden evaluation. CONCLUSIONS: DANTE-FLASH provided better image quality than MRDTI and accurately detected thrombi. It may, therefore, serve as a safe and convenient alternative for the diagnosis of DVT.

Positive Contrast Susceptibility MR Imaging Using GPU-based Primal-Dual Algorithm.

The susceptibility-based positive contrast MR technique was applied to estimate arbitrary magnetic susceptibility distributions of the metallic devices using a kernel deconvolution algorithm with a regularized L-1 minimization. Previously, the first-order primal-dual (PD) algorithm could provide a faster reconstruction time to solve the L-1 minimization, compared with other methods. Here, we propose to accelerate the PD algorithm of the positive contrast image using the multi-core multi-thread feature of graphics processor units (GPUs). The some experimental results showed that the GPU-based PD algorithm could achieve comparable accuracy of the metallic interventional devices in positive contrast imaging with less computational time. And the GPU-based PD approach was 4~15 times faster than the previous CPU-based scheme.Clinical Relevance-This can estimate arbitrary magnetic susceptibility distributions of the metallic devices with the processing efficacy of 4~15 times faster than before.

Evaluation of left ventricular strain in patients with arrhythmia based on the 3T MR temporal parallel acquisition technique.

Most of the current studies on myocardial strain are mainly applied in patients with sinus rhythm because the image quality of arrhythmias obtained with conventional scanning sequences does not meet diagnostic needs. Here, we intend to assess left ventricular (LV) global myocardial strain in patients with arrhythmias with 3 Tesla magnetic resonance (MR) and a new cine sequence. Thirty-three patients with arrhythmia and forty-eight subjects with sinus rhythm were enrolled in the study. LV myocardial thickness, cardiac function, myocardial strain and the apparent contrast-to-noise ratio (CNR) were all measured and compared using images generated by the real-time temporal parallel acquisition technique (TPAT) and the conventional cine sequence. In the arrhythmia group, the image quality of real-time TPAT was significantly better than that of the conventional cine sequence. In the arrhythmia group, the LV global peak radial strain and global peak circumferential strain values of real-time TPAT were significantly different from those of the conventional technique (radial strain, conventional: 20.27 ± 15.39 vs. TPAT: 24.14 ± 15.85, p = 0.007; circumferential strain, conventional:-12.06 ± 6.60 vs. TPAT: -13.71 ± 6.31, p = 0.015). There was no significant difference in global peak longitudinal strain between real-time TPAT and the conventional technique (-10.94 ± 4.66 vs. -10.70 ± 5.96, p = 0.771). There was no significant difference in the cardiac function parameters between the two techniques (p > 0.05), but there was a significant difference in 12 segments of the LV wall thickness between the two sequences (p < 0.05). In the sinus rhythm group, image quality using real-time TPAT was comparable to that using the conventional technique, and there was no significant difference in any of the indices (p > 0.05). Real-time TPAT is an effective method for detection of left ventricular myocardial deformation in patients with arrhythmia.

Multicenter Study on the Diagnostic Performance of Native-T1 Cardiac Magnetic Resonance of Chronic Myocardial Infarctions at 3T.

BACKGROUND: Preclinical studies and pilot patient studies have shown that chronic infarctions can be detected and characterized from cardiac magnetic resonance without gadolinium-based contrast agents using native-T1 maps at 3T. We aimed to investigate the diagnostic capacity of this approach for characterizing chronic myocardial infarctions (MIs) in a multi-center setting. METHODS: Patients with a prior MI (n=105) were recruited at 3 different medical centers and were imaged with native-T1 mapping and late gadolinium enhancement (LGE) at 3T. Infarct location, size, and transmurality were determined from native-T1 maps and LGE. Sensitivity, specificity, receiver-operating characteristic metrics, and inter- and intraobserver variabilities were assessed relative to LGE. RESULTS: Across all subjects, T1 of MI territory was 1621±110 ms, and remote territory was 1225±75 ms. Sensitivity, specificity, and area under curve for detecting MI location based on native-T1 mapping relative to LGE were 88%, 92%, and 0.93, respectively. Native-T1 maps were not different for measuring infarct size (native-T1 maps: 12.1±7.5%; LGE: 11.8±7.2%, P=0.82) and were in agreement with LGE (R2=0.92, bias, 0.09±2.6%). Corresponding inter- and intraobserver assessments were also highly correlated (interobserver: R2=0.90, bias, 0.18±2.4%; and intraobserver: R2=0.91, bias, 0.28±2.1%). Native T1 maps were not different for measuring MI transmurality (native-T1 maps: 49.1±15.8%; LGE: 47.2±19.0%, P=0.56) and showed agreement (R2=0.71; bias, 1.32±10.2%). Corresponding inter- and intraobserver assessments were also in agreement (interobserver: R2=0.81, bias, 0.1±9.4%; and intraobserver: R2=0.91, bias, 0.28±2.1%, respectively). While the overall accuracy for detecting MI with native-T1 maps at 3T was high, logistic regression analysis showed that MI location was a prominent confounder. CONCLUSIONS: Native-T1 mapping can be used to image chronic MI with high degree of accuracy, and as such, it is a viable alternative for scar imaging in patients with chronic MI who are contraindicated for LGE. Technical advancements may be needed to overcome the imaging confounders that currently limit native-T1 mapping from reaching equivalent detection levels as LGE.

Comparison Between the Diagnostic Performance of 1.5 T and 3.0 T field Strengths for Detecting Deep Vein Thrombosis Using Magnetic Resonance Black-Blood Thrombus Imaging.

BACKGROUND: Magnetic resonance (MR) black-blood thrombus imaging (BTI) is an accurate diagnostic technique for detecting deep vein thrombosis (DVT) but to date there have been no studies comparing the diagnostic performance and consistency of this technique at different field strengths. In this study, we evaluated and compared the diagnostic performance of BTI for detecting DVT at 1.5 T and 3.0 T field strengths. METHODS: A total of 40 patients with DVT were enrolled in this study from November 2015 up to October 2018. All patients underwent BTI, a contrast-free T1-weighted MR imaging technique for detecting DVT, and contrast-enhanced MR venography (CE-MRV) at 1.5 T or 3.0 T field strengths. The MR data analyses used 1160 segments from the venous lumen of the 40 patients. The signal-to-noise ratio and contrast-to-noise ratio between thrombus and muscle/lumen were calculated to compare BTI at 1.5 T or 3.0 T to determine the image performance for thrombus detection at 1.5 T or 3.0 T. Two physicians blinded to the study evaluated all BTI images and calculated the overall sensitivity (SE), specificity (SP), positive predictive value (PPV) and negative predictive value (NPV), accuracy, and diagnostic consistency at 1.5 T and 3.0 T. These images and values were compared to control CE-MRV images that had been obtained by 2 senior physicians and used as reference standards. In addition, the reliability and consistency of diagnoses between observers were also evaluated. RESULTS: Two study-blind physicians reviewed all BTI images to diagnose thrombus and to determine SE, SP, PPV, NPV, and accuracy. There were no statistical differences in SE, SP, PPV, NPV, or accuracy between the 1.5 T and 3.0 T groups. CONCLUSIONS: Black-blood thrombus imaging has high SE, SP, and accuracy for DVT diagnosis both at 1.5 T and 3.0 T field strengths. This noninvasive diagnostic technique, which does not require the use of contrast agents, can be widely used in the clinical screening of DVT and follow-up after treatment.

High spatial resolution BOLD fMRI using simultaneous multislice excitation with echo-shifting gradient echo at 7 Tesla.

We introduce an accelerated gradient echo (GRE) sequence combining simultaneous multislice excitation (SMS) with echo-shifting technique for high spatial resolution blood oxygen level dependent (BOLD) functional MRI (fMRI). The simulation was conducted to optimize scan parameters. To validate the feasibility of the proposed technique, the visual and motor task experiments were performed at 7.0 Tesla (T). The single-shot EPI sequence was also applied in comparison with the proposed technique. The simulation results showed that an optimized flip angle of 9° provided maximal BOLD contrast for our scanning scheme, allowing low power deposition and SMS acceleration factor of 5. Additionally, parallel acquisition imaging with acceleration factor of 2 was utilized, which allowed a total acceleration factor of 10 in volunteer study. The experiment results showed that geometric distortion-free BOLD images with voxel size of 1.0 × 1.0 × 2.5 mm3 were obtained. Significant brain activation was identified in both visual and motor task experiments, which were in accordance with previous investigations. The proposed technique has potential for high spatial resolution fMRI at ultra-high field because of its sufficient BOLD sensitivity as well as improved acquisition speed over conventional GRE-based techniques.

Grading of hepatocellular carcinoma based on diffusion weighted images with multiple b-values using convolutional neural networks.

PURPOSE: To effectively grade hepatocellular carcinoma (HCC) based on deep features derived from diffusion weighted images (DWI) with multiple b-values using convolutional neural networks (CNN). MATERIALS AND METHODS: Ninety-eight subjects with 100 pathologically confirmed HCC lesions from July 2012 to October 2018 were included in this retrospective study, including 47 low-grade and 53 high-grade HCCs. DWI was performed for each subject with a 3.0T MR scanner in a breath-hold routine with three b-values (0,100, and 600 s/mm2 ). First, logarithmic transformation was performed on original DWI images to generate log maps (logb0, logb100, and logb600). Then, a resampling method was performed to extract multiple 2D axial planes of HCCs from the log map to increase the dataset for training. Subsequently, 2D CNN was used to extract deep features of the log map for HCCs. Finally, fusion of deep features derived from three b-value log maps was conducted for HCC malignancy classification. Specifically, a deeply supervised loss function was devised to further improve the performance of lesion characterization. The data set was split into two parts: the training and validation set (60 HCCs) and the fixed test set (40 HCCs). Four-fold cross validation with 10 repetitions was performed to assess the performance of deep features extracted from single b-value images for HCC grading using the training and validation set. Receiver operating characteristic curve (ROC) and area under the curve (AUC) values were used to assess the characterization performance of the proposed deep feature fusion method to differentiate low-grade and high-grade in the fixed test set. RESULTS: The proposed fusion of deep features derived from logb0, logb100, and logb600 with deeply supervised loss function generated the highest accuracy for HCC grading (80%), thus outperforming the method of deep feature derived from the ADC map directly (72.5%), the original b0 (65%), b100 (68%), and b600 (70%) images. Furthermore, AUC values of the deep features of the ADC map, the deep feature fusion with concatenation, and the proposed deep feature fusion with deeply supervised loss function were 0.73, 0.78, and 0.83, respectively. CONCLUSION: The proposed fusion of deep features derived from the logarithm of the three b-value images yields high performance for HCC grading, thus providing a promising approach for the assessment of DWI in lesion characterization.

Positive-contrast susceptibility imaging based on first-order primal-dual optimization.

PURPOSE: To achieve faster reconstruction and better imaging quality of positive-contrast MRI based on the susceptibility mapping by incorporating a primal-dual (PD) formulation. METHODS: The susceptibility-based positive contrast MR technique was applied to estimate arbitrary magnetic susceptibility distributions of the metallic devices using a kernel deconvolution algorithm with a regularized ℓ1 minimization. The regularized positive-contrast inversion problem and its PD formulation were derived. The visualization of the positive contrast and convergence behavior of the PD algorithm were compared with those of the nonlinear conjugate gradient algorithm, fast iterative soft-thresholding algorithm, and alternating direction method of multipliers. These methods were tested and validated on computer simulations and phantom experiments. RESULTS: The PD approach could provide a faster reconstruction time compared with other methods. Experimental results showed that the PD algorithm could achieve comparable or even better visualization and accuracy of the metallic interventional devices in positive-contrast imaging with different SNRs and orientations to the B0 field. CONCLUSION: A susceptibility-based positive-contrast imaging technique by PD algorithm was proposed. The PD approach has more superior performance than other algorithms in terms of reconstruction time and accuracy for imaging the metallic interventional devices.