Cine MRI Research Articles

Evaluation of the active contour and topographic watershed segmentation “assessment of the systolic ejection fraction in the left ventricular for medical assistance in 5D short axis cine MRI”

A comparative study has been depicted between the contour and topographic watershed segmentation approach for short-axis 5D cardiac sequences with MRI for medical decision. The fifth dimension has been defined as the excitation of pixels based on the gray scale around the myocardium without consideration of the morphological structure of the heart in 3D and fourth dimension (time). Three patients were performed the first is healthy, the second has a genetic disease, and the third had a heart failure syndrome for a dimension ROI = 150mm, average age is 54 years old, and mean of weight = 86 kg. A contouring and watershed segmentation algorithm for a sample of 63 Cine Fiesta MRI sequences for short-axis cuts with Matlab and its in-box toolbox complements was implemented.For a healthy patient 13.4% tolerance rate for the estimation of the stroke fraction, 6.4% for a patient with genetic disease, 8.7% error rate for a patient with heart failure symptom. The results show that the regurgitation fraction by the contour approach for a patient case with symptom of the presence of a genetic disease is 0.0335% for an aortic valve, 0.248% for a mitral valve, an error rate 0.16% for estimating this parameter for the aortic orifice with the watershed segmentation approach. In return, for a patient with suspected heart failure (stenosis or regurgitation) the regurgitation fraction is estimated by 0% for aortic valve, 1.49 e + 03% for a mitral valve, an error rate 11.76% compared to the watershed segmentation approach. The results are validated clinically. The Optimization of the topographic watershed approach with mutual information was simulated for the extraction of measurements (ejection fraction, regurgitation rate) within the left ventricle for three patient types (healthy, genetic pathology and heart failure). The results are considered interesting compared to the clinical assessment.

Haemodynamic changes in cirrhosis following terlipressin and induction of sepsis\u2014a preclinical study using caval subtraction phase-contrast and cardiac MRI

ObjectivesEffects of liver disease on portal venous (PV), hepatic arterial (HA), total liver blood flow (TLBF), and cardiac function are poorly understood. Terlipressin modulates PV flow but effects on HA, TLBF, and sepsis/acute-on-chronic liver failure (ACLF)-induced haemodynamic changes are poorly characterised. In this study, we investigated the effects of terlipressin and sepsis/ACLF on hepatic haemodynamics and cardiac function in a rodent cirrhosis model using caval subtraction phase-contrast (PC) MRI and cardiac cine MRI.MethodsSprague-Dawley rats (n = 18 bile duct–ligated (BDL), n = 16 sham surgery controls) underwent caval subtraction PCMRI to estimate TLBF and HA flow and short-axis cardiac cine MRI for systolic function at baseline, following terlipressin and lipopolysaccharide (LPS) infusion, to model ACLF.ResultsAll baseline hepatic haemodynamic/cardiac systolic function parameters (except heart rate and LV mass) were significantly different in BDL rats. Following terlipressin, baseline PV flow (sham 181.4 ± 12.1 ml/min/100 g; BDL 68.5 ± 10.1 ml/min/100 g) reduced (sham − 90.3 ± 11.1 ml/min/100 g, p < 0.0001; BDL − 31.0 ± 8.0 ml/min/100 g, p = 0.02), sham baseline HA flow (33.0 ± 11.3 ml/min/100 g) increased (+ 92.8 ± 21.3 ml/min/100 g, p = 0.0003), but BDL baseline HA flow (83.8 ml/min/100 g) decreased (− 34.4 ± 7.5 ml/min/100 g, p = 0.11). Sham baseline TLBF (214.3 ± 16.7 ml/min/100 g) was maintained (+ 2.5 ± 14.0 ml/min/100 g, p > 0.99) but BDL baseline TLBF (152.3 ± 18.7 ml/min/100 g) declined (− 65.5 ± 8.5 ml/min/100 g, p = 0.0004). Following LPS, there were significant differences between cohort and change in HA fraction (p = 0.03) and TLBF (p = 0.01) with BDL baseline HA fraction (46.2 ± 4.6%) reducing (− 20.9 ± 7.5%, p = 0.03) but sham baseline HA fraction (38.2 ± 2.0%) remaining unchanged (+ 2.9 ± 6.1%, p > 0.99). Animal cohort and change in systolic function interactions were significant only for heart rate (p = 0.01) and end-diastolic volume (p = 0.03).ConclusionsCaval subtraction PCMRI and cardiac MRI in a rodent model of cirrhosis demonstrate significant baseline hepatic haemodynamic/cardiac differences, failure of the HA buffer response post-terlipressin and an altered HA fraction response in sepsis, informing potential translation to ACLF patients.Key PointsCaval subtraction phase-contrast and cardiac MRI demonstrate:• Significant differences between cirrhotic/non-cirrhotic rodent hepatic blood flow and cardiac systolic function at baseline.• Failure of the hepatic arterial buffer response in cirrhotic rodents in response to terlipressin.• Reductions in hepatic arterial flow fraction in the setting of acute-on-chronic liver failure.

Cardiac MRI Reveals Late Diastolic Changes in Left Ventricular Relaxation Patterns During Healthy Aging.

Cardiac MRI is an emerging modality for evaluating left ventricular (LV) diastolic dysfunction (LVDD), a pathological condition that is prevalent in aging populations. However, there is a lack of reports of MRI-derived LV diastolic properties in late diastole. To test the hypothesis that cine MRI-derived motion/deformation indices can be used to characterize age-related changes on LV relaxation patterns in late diastole. Retrospective. In all, 412 participants (72.5 ± 4.6 years old, range 65-84) without a documented history of cardiovascular diseases. Balanced steady-state free precession(bSSFP) acquired at 1.5T. Participants were divided into younger (65-74 years old, n = 275) and older (75-84 years old, n = 137) groups. Status of diabetes mellitus (DM), hypertension (HTN), and lipid disorders were recorded for each participant. Cine MRI datasets were analyzed by using heart deformation analysis (HDA). LV motion/deformation indices (displacement, velocity, strain, and strain rate) were measured through 22 phases within a cardiac cycle. The prevalence of traditional cardiovascular risk conditions, LV ejection fraction (LVEF), peak LV regional displacement, velocity, and strain rates at early and late diastole were compared between two participant groups using chi-square tests or t-tests. Older participants had a significantly lower peak early radial displacement (0.797 ± 0.249 cm vs. 0.876 ± 0.286 cm), radial velocity (19.3 ± 6.3 mm/s vs. 17.5 ± 5.2 mm/s), and circumferential strain rate (64.6 ± 15.7%/s vs. 70.1 ± 17%/s) but a higher peak late circumferential strain rate (69.8 ± 16.3 %/s vs. 66 ± 15.8 %/s) than their younger counterparts. Cine MRI can be used to characterize age-related LV relaxation patterns in late diastole. 3. 1.

A modular and scalable computational framework for interactive immersion into imaging data with a holographic augmented reality interface

Background and objectiveModern imaging scanners produce an ever-growing body of 3D/4D multimodal data requiring image analytics and visualization of fused images, segmentations, and information. For the latter, augmented reality (AR) with head-mounted displays (HMDs) has shown potential. This work describes a framework (FI3D) for interactive immersion with data, integration of image processing and analytics, and rendering and fusion with an AR interface. MethodsThe FI3D was designed and endowed with modules to communicate with peripherals, including imaging scanners and HMDs, and to provide computational power for data acquisition and processing. The core of FI3D is deployed to a dedicated computational unit that performs the computationally demanding processes in real-time, and the HMD is used as a display output peripheral and an input peripheral through gestures and voice commands. FI3D offers user-made processing and analysis dedicated modules. Users can customize and optimize these for a particular workflow while incorporating current or future libraries. ResultsThe FI3D framework was used to develop a workflow for processing, rendering, and visualization of CINE MRI cardiac sets. In this version, the data were loaded from a remote database, and the endocardium and epicardium of the left ventricle (LV) were segmented using a machine learning model and transmitted to a HoloLens HMD to be visualized in 4D. Performance results show that the system is capable of maintaining an image stream of one image per second with a resolution of 512 × 512. Also, it can modify visual properties of the holograms at 1 update per 16 milliseconds (62.5 Hz) while providing enough resources for the segmentation and surface reconstruction tasks without hindering the HMD. ConclusionsWe provide a system design and framework to be used as a foundation for medical applications that benefit from AR visualization, removing several technical challenges from the developmental pipeline.

A mathematical model for predicting intracranial pressure based on noninvasively acquired PC-MRI parameters in communicating hydrocephalus.

To develop and validate a mathematical model for predicting intracranial pressure (ICP) noninvasively using phase-contrast cine MRI (PC-MRI). We performed a retrospective analysis of PC-MRI from patients with communicating hydrocephalus (n = 138). The patients were recruited from Shenzhen Second People’s Hospital between November 2017 and April 2020, and randomly allocated into training (n = 97) and independent validation (n = 41) groups. All participants underwent lumbar puncture and PC-MRI in order to evaluate ICP and cerebrospinal fluid (CSF) parameters (i.e., aqueduct diameter and flow velocity), respectively. A novel ICP-predicting model was then developed based on the nonlinear relationships between the CSF parameters, using the Levenberg–Marquardt and general global optimisation methods. There was no significant difference in baseline demographic characteristics between the training and independent validation groups. The accuracy of the model for predicting ICP was 0.899 in the training cohort (n = 97) and 0.861 in the independent validation cohort (n = 41). We obtained an ICP-predicting model that showed excellent performance in the noninvasive diagnosis of clinically significant communicating hydrocephalus.

Computational and Mechanical Modeling of SCG Signals

Introduction Seismocardiography (SCG) includes vibration signals propagated to the chest surface from cardiac valve movement, blood flow and myocardial muscle contraction. SCG feature analysis may prove useful in the early non-invasive detection of HF deterioration. In this study, the propagation and coupling of cardiac surface motion to the anterior chest surface was modeled using computer simulations. Objective Develop a computational model of SCG propagation to enhance understanding of SCG genesis, and SCG's spatial amplitude and spectrographic distribution over the anterior chest wall. Methods The computational geometry was extracted from short-axis cardiac cine MRI imaging of a healthy participant. Cardiac wall motion was captured by applying a 3D motion tracking algorithm to the cine MRI images. The propagation of cardiac motion to the chest surface was simulated using Finite Element Method (FEM). A mechanical model was also developed where the displacements at a simulated heart wall were coupled to a mechanical chest surface model using an array of linear actuators. Results Common SCG feature points (e.g., MC-mitral valve closure; IM- Isovolumic contraction; AO- aortic valve opening; AC- aortic valve closure; and MO- mitral valve opening) were predicted by the model. Furthermore, the timing of these feature points was validated against the left ventricle volume variation extracted from short-axis cine MRI images. The maximum SCG amplitude (in the dorso-ventral direction) was observed in the 4th intercoastal space near the left lower sternal border. The pilot results of the mechanical model were in a qualitative agreement. Conclusions The computer model predicted the feature points observed in real SCG measurements. The current model may be utilized in the future to study the distribution of SCG over precordium and to analyze the effect of the properties of cardiac and contiguous chest structures, including great vessels, chest wall ribs, muscles and articulations, lungs, and cutaneous soft tissue. Further model development may deepen understanding of SCG generation and propagation, increasing its potential utility for diagnosis and monitoring of cardiac pathologies.

Evaluation of cerebrospinal fluid flow dynamic changes in patients with idiopathic intracranial hypertension using phase contrast cine MR imaging

BackgroundIdiopathic intracranial hypertension (IIH) mainly affects overweight women in the middle age period. The pathophysiology of IIH stays unclear, but suggested mechanisms include excess CSF production, reduced CSF absorption, increased brain water content, and increased cerebral venous pressureObjectivesTo assess the cerebrospinal fluid (CSF) flow dynamic changes in aqueduct of Sylvius in patients of idiopathic intracranial hypertension (IIH) with new MRI technique: phase contrast cine MRI (PCC-MRI).MethodsThirty patients diagnosed with idiopathic intracranial hypertension were divided into 3 groups according to treatment options (no treatment, medical treatment, and medical treatment with repeated lumbar tapping). CSF flow data were evaluated by phase contrast cine MRI.ResultsPCC-MRI parameters were significantly higher in group who was on medical treatment (group II) than other groups. The sensitivity of PCC MRI parameters ranged from 56.7 (stroke volume (SV) and mean flow (MF)) to 83.3% (peak systolic velocity (PSV)). A statistically significant difference was found for the mean flow value (p 0.039) between the control group and IIH patients.ConclusionThe most specific CSF flowmetry parameter detected to help diagnosis of IIH is mean flow especially among early discovered patients. PCC MRI can be used as non-invasive technique for diagnosis of IIH and treatment follow-up.

Reconstruction of 5D cardiac MRI through the blood flow registration: simulation of the fifth dimension and assessment of the left ventricular ejection fraction

In this paper, a medical decision approach for cardiac MRI by the registration algorithm and the reconstruction of 5D stacks for cine MRI sequences was developed. The estimation of the ejection fraction of the left ventricle (LVFE) was developed as an example. This study included 18 patients who underwent 1.5 T cardiac MRI for different etiologies: a total of 395 series and 18,483 scans was tested. The estimated values with the classical method of segmentation by contour with extreme (LVFE) in the sample was compared. A registration algorithm was implemented, with mean elapsed time registration = 0.5 s. The range of (LVFE) varied between [20 and 87%] shows that the results are satisfactory for the experts by comparing with the clinical assessment for the study of the anomalies of myocardial contractility and kinetic abnormalities and the error rate was significantly reduced. At baseline, the correlation between the clinical assessment of the LVFE with the 5D approach was consistent with 0.92, and with a 1.39% error rate mean for the sample studied, a mean = 60.34, the probability = 0.46 and the standard deviation was estimated to be 20.33 and about 21.78 for the LVFE with the 5D approach. Therefore, the 5D stacks approach applied to cine-cardiac MRI sequences for medical decision-making was implemented in the left ventricle with cardiac MRI. This approach shows its effectiveness in improving the accuracy of computation of LVFE. This approach seems to be very interesting, but needs to be validated in a larger sample study.

Rapid reconstruction of highly undersampled, non-Cartesian real-time cine k-space data using a perceptual complex neural network (PCNN).

Highly accelerated real-time cine MRI using compressed sensing (CS) is a promising approach to achieve high spatio-temporal resolution and clinically acceptable image quality in patients with arrhythmia and/or dyspnea. However, its lengthy image reconstruction time may hinder its clinical translation. The purpose of this study was to develop a neural network for reconstruction of non-Cartesian real-time cine MRI k-space data faster (<1 min per slice with 80 frames) than graphics processing unit (GPU)-accelerated CS reconstruction, without significant loss in image quality or accuracy in left ventricular (LV) functional parameters. We introduce a perceptual complex neural network (PCNN) that trains on complex-valued MRI signal and incorporates a perceptual loss term to suppress incoherent image details. This PCNN was trained and tested with multi-slice, multi-phase, cine images from 40 patients (20 for training, 20 for testing), where the zero-filled images were used as input and the corresponding CS reconstructed images were used as practical ground truth. The resulting images were compared using quantitative metrics (structural similarity index (SSIM) and normalized root mean square error (NRMSE)) and visual scores (conspicuity, temporal fidelity, artifacts, and noise scores), individually graded on a five-point scale (1, worst; 3, acceptable; 5, best), and LV ejection fraction (LVEF). The mean processing time per slice with 80 frames for PCNN was 23.7 ± 1.9 s for pre-processing (Step 1, same as CS) and 0.822 ± 0.004 s for dealiasing (Step 2, 166 times faster than CS). Our PCNN produced higher data fidelity metrics (SSIM = 0.88 ± 0.02, NRMSE = 0.014 ± 0.004) compared with CS. While all the visual scores were significantly different (P < 0.05), the median scores were all 4.0 or higher for both CS and PCNN. LVEFs measured from CS and PCNN were strongly correlated (R2 = 0.92) and in good agreement (mean difference = -1.4% [2.3% of mean]; limit of agreement = 10.6% [17.6% of mean]). The proposed PCNN is capable of rapid reconstruction (25 s per slice with 80 frames) of non-Cartesian real-time cine MRI k-space data, without significant loss in image quality or accuracy in LV functional parameters.

CINENet: deep learning-based 3D cardiac CINE MRI reconstruction with multi-coil complex-valued 4D spatio-temporal convolutions

Cardiac CINE magnetic resonance imaging is the gold-standard for the assessment of cardiac function. Imaging accelerations have shown to enable 3D CINE with left ventricular (LV) coverage in a single breath-hold. However, 3D imaging remains limited to anisotropic resolution and long reconstruction times. Recently deep learning has shown promising results for computationally efficient reconstructions of highly accelerated 2D CINE imaging. In this work, we propose a novel 4D (3D + time) deep learning-based reconstruction network, termed 4D CINENet, for prospectively undersampled 3D Cartesian CINE imaging. CINENet is based on (3 + 1)D complex-valued spatio-temporal convolutions and multi-coil data processing. We trained and evaluated the proposed CINENet on in-house acquired 3D CINE data of 20 healthy subjects and 15 patients with suspected cardiovascular disease. The proposed CINENet network outperforms iterative reconstructions in visual image quality and contrast (+ 67% improvement). We found good agreement in LV function (bias ± 95% confidence) in terms of end-systolic volume (0 ± 3.3 ml), end-diastolic volume (− 0.4 ± 2.0 ml) and ejection fraction (0.1 ± 3.2%) compared to clinical gold-standard 2D CINE, enabling single breath-hold isotropic 3D CINE in less than 10 s scan and ~ 5 s reconstruction time.

Initial clinical experience of Stereotactic Body Radiation Therapy (SBRT) for liver metastases, primary liver malignancy, and pancreatic cancer with 4D-MRI based online adaptation and real-time MRI monitoring using a 1.5 Tesla MR-Linac.

Purpose/ObjectivesRecently a 1.5 Tesla MR Linac has been FDA approved and is commercially available. Clinical series describing treatment methods and outcomes for upper abdominal tumors using a 1.5 Tesla MR Linac are lacking. We present the first clinical series of upper abdominal tumors treated using a 1.5 Tesla MR Linac along with the acquisition of intra-treatment quantitative imaging.Materials/Methods10 patients with abdominal tumors were treated at our institution. Each patient enrolled in an IRB approved advanced imaging protocol. Both daily real-time adaptive and non-adaptive methods were used, and selection criteria are described. Adaptive plans were based on pre-beam motion-averaged or mid-position images derived from respiratory-correlated 4D-MRI. Quantitative intravoxel incoherent motion diffusion-weighted imaging and T2 mapping were acquired during plan adaptation. Real-time motion monitoring using cine MRI was performed during beam-on.ResultsMedian patient age was 68.2, five patients were female. Tumor types included liver metastatic lesions from melanoma and sarcoma, primary liver hepatocellular carcinoma (HCC), and regional abdominal tumors included pancreatic metastatic lesions from renal cell carcinoma (RCC) along with two cases of recurrent pancreatic cancer. Doses included 30 Gy in 6 fractions, 33 Gy in 5 fractions, 50 Gy in 5 fractions, 45 Gy in 3 fractions, and 60 Gy in 3 fractions, depending on the location and clinical circumstances. Treatments were feasible and were successfully completed in all patients without significant acute toxicity, technical complications, or need for back up CT based treatment plans.ConclusionsWe present a first clinical series of patients treated for pancreatic tumors, primary liver tumors, and secondary liver tumors with a 1.5 Tesla MR Linear accelerator using adapt-to-position and adapt-to-shape strategies. Treatments were well tolerated by all patients. Acquisition of fully quantitative MR imaging was feasible during the course of the treatment delivery workflow without extending overall treatment times.

Amplified Flow Imaging (aFlow): A Novel MRI-Based Tool to Unravel the Coupled Dynamics Between the Human Brain and Cerebrovasculature.

With each heartbeat, periodic variations in arterial blood pressure are transmitted along the vasculature, resulting in localized deformations of the arterial wall and its surrounding tissue. Quantification of such motions may help understand various cerebrovascular conditions, yet it has proven technically challenging thus far. We introduce a new image processing algorithm called amplified Flow (aFlow) which allows to study the coupled brain-blood flow motion by combining the amplification of cine and 4D flow MRI. By incorporating a modal analysis technique known as dynamic mode decomposition into the algorithm, aFlow is able to capture the characteristics of transient events present in the brain and arterial wall deformation. Validating aFlow, we tested it on phantom simulations mimicking arterial walls motion and observed that aFlow displays almost twice higher SNR than its predecessor amplified MRI (aMRI). We then applied aFlow to 4D flow and cine MRI datasets of 5 healthy subjects, finding high correlations between blood flow velocity and tissue deformation in selected brain regions, with correlation values r = 0.61 , 0.59, 0.52 for the pons, frontal and occipital lobe ( ). Finally, we explored the potential diagnostic applicability of aFlow by studying intracranial aneurysm dynamics, which seems to be indicative of rupture risk. In two patients, aFlow successfully visualized the imperceptible aneurysm wall motion, additionally quantifying the increase in the high frequency wall displacement after a one-year follow-up period (20%, 76%). These preliminary data suggest that aFlow may provide a novel imaging biomarker for the assessment of aneurysms evolution, with important potential diagnostic implications.

Accelerating cardiac cine MRI using a deep learning-based ESPIRiT reconstruction.

To propose a novel combined parallel imaging and deep learning-based reconstruction framework for robust reconstruction of highly accelerated 2D cardiac cine MRI data. We propose DL-ESPIRiT, an unrolled neural network architecture that utilizes an extended coil sensitivity model to address SENSE-related field-of-view (FOV) limitations in previously proposed deep learning-based reconstruction frameworks. Additionally, we propose a novel neural network design based on (2+1)D spatiotemporal convolutions to produce more accurate dynamic MRI reconstructions than conventional 3D convolutions. The network is trained on fully sampled 2D cardiac cine datasets collected from 11 healthy volunteers with IRB approval. DL-ESPIRiT is compared against a state-of-the-art parallel imaging and compressed sensing method known as -ESPIRiT. The reconstruction accuracy of both methods is evaluated on retrospectively undersampled datasets (R=12) with respect to standard image quality metrics as well as automatic deep learning-based segmentations of left ventricular volumes. Feasibility of DL-ESPIRiT is demonstrated on two prospectively undersampled datasets acquired in a single heartbeat per slice. The (2+1)D DL-ESPIRiT method produces higher fidelity image reconstructions when compared to -ESPIRiT reconstructions with respect to standard image quality metrics (P < .001). As a result of improved image quality, segmentations made from (2+1)D DL-ESPIRiT images are also more accurate than segmentations from -ESPIRiT images. DL-ESPIRiT synergistically combines a robust parallel imaging model and deep learning-based priors to produce high-fidelity reconstructions of retrospectively undersampled 2D cardiac cine data acquired with reduced FOV. Although a proof-of-concept is shown, further experiments are necessary to determine the efficacy of DL-ESPIRiT in prospectively undersampled data.

硬膜欠損の同定にphase-contrast cine MRIが有用であった脳表ヘモジデリン沈着症の1例

硬膜欠損の同定にphase-contrast cine MRIが有用であった脳表ヘモジデリン沈着症の1例

A Semi-Automatic Method To Segment The Left Atrium in MR Volumes With Varying Slice Numbers.

Atrial fibrillation (AF) is the most common sustained arrhythmia and is associated with dramatic increases in mortality and morbidity. Atrial cine MR images are increasingly used in the management of this condition, but there are few specific tools to aid in the segmentation of such data. Some characteristics of atrial cine MR (thick slices, variable number of slices in a volume) preclude the direct use of traditional segmentation tools. When combined with scarcity of labelled data and similarity of the intensity and texture of the left atrium (LA) to other cardiac structures, the segmentation of the LA in CINE MRI becomes a difficult task. To deal with these challenges, we propose a semi-automatic method to segment the left atrium (LA) in MR images, which requires an initial user click per volume. The manually given location information is used to generate a chamber location map to roughly locate the LA, which is then used as an input to a deep network with slightly over 0.5 million parameters. A tracking method is introduced to pass the location information across a volume and to remove unwanted structures in segmentation maps. According to the results of our experiments conducted in an in-house MRI dataset, the proposed method outperforms the U-Net [1] with a margin of 20 mm on Hausdorff distance and 0.17 on Dice score, with limited manual interaction.

Fully Automated Quantification of Cardiac Indices from Cine MRI Using a Combination of Convolution Neural Networks.

Cardiovascular magnetic resonance imaging (CMRI) is one of the most accurate non-invasive modalities for evaluation of cardiac function, especially the left ventricle (LV). In this modality, the manual or semi-automatic delineation of LV by experts is currently the standard clinical practice for chambers segmentation. Despite these efforts, global quantification of LV remains a challenge. In this work, a combination of two convolutional neural network (CNN) architectures for quantitative evaluation of the LV is described, which estimates the cavity and the myocardium areas, endocardial cavity dimensions in three directions, and the myocardium regional wall thickness in six radial directions. The method was validated in CMRI exams of 56 patients (LVQuan19 dataset) and evaluated by metrics Dice Index, Mean Absolute Error, and Correlation with superior performance compared to the state-of-the-art methods. The combination of the CNN architectures provided a simpler yet fully automated approach, requiring no specialist interaction.Clinical Relevance- With the proposed method, it is possible to perform automatically the full quantification of regional clinically relevant parameters of the left ventricle in short-axis CMRI images with superior performance compared to state-of-the-art methods.

Strain maps of the left atrium imaged with a novel high-resolution CINE MRI protocol.

To date, regional atrial strains have not been imaged in vivo, despite their potential to provide useful clinical information. To address this gap, we present a novel CINE MRI protocol capable of imaging the entire left atrium at an isotropic 2-mm resolution in one single breath-hold. As proof of principle, we acquired data in 10 healthy volunteers and 2 cardiovascular patients using this technique. We also demonstrated how regional atrial strains can be estimated from this data following a manual segmentation of the left atrium using automatic image tracking techniques. The estimated principal strains vary smoothly across the left atrium and have a similar magnitude to estimates reported in the literature.

MRI Manufacturer Shift and Adaptation: Increasing the Generalizability of Deep Learning Segmentation for MR Images Acquired with Different Scanners.

To quantitatively evaluate the generalizability of a deep learning segmentation tool to MRI data from scanners of different MRI manufacturers and to improve the cross-manufacturer performance by using a manufacturer-adaptation strategy. This retrospective study included 150 cine MRI datasets from three MRI manufacturers, acquired between 2017 and 2018 (n = 50 for manufacturer 1, manufacturer 2, and manufacturer 3). Three convolutional neural networks (CNNs) were trained to segment the left ventricle (LV), using datasets exclusively from images from a single manufacturer. A generative adversarial network (GAN) was trained to adapt the input image before segmentation. The LV segmentation performance, end-diastolic volume (EDV), end-systolic volume (ESV), LV mass, and LV ejection fraction (LVEF) were evaluated before and after manufacturer adaptation. Paired Wilcoxon signed rank tests were performed. The segmentation CNNs exhibited a significant performance drop when applied to datasets from different manufacturers (Dice reduced from 89.7% ± 2.3 [standard deviation] to 68.7% ± 10.8, P < .05, from 90.6% ± 2.1 to 59.5% ± 13.3, P < .05, from 89.2% ± 2.3 to 64.1% ± 12.0, P < .05, for manufacturer 1, 2, and 3, respectively). After manufacturer adaptation, the segmentation performance was significantly improved (from 68.7% ± 10.8 to 84.3% ± 6.2, P < .05, from 72.4% ± 10.2 to 85.7% ± 6.5, P < .05, for manufacturer 2 and 3, respectively). Quantitative LV function parameters were also significantly improved. For LVEF, the manufacturer adaptation increased the Pearson correlation from 0.005 to 0.89 for manufacturer 2 and from 0.77 to 0.94 for manufacturer 3. A segmentation CNN well trained on datasets from one MRI manufacturer may not generalize well to datasets from other manufacturers. The proposed manufacturer adaptation can largely improve the generalizability of a deep learning segmentation tool without additional annotation.Supplemental material is available for this article.© RSNA, 2020.

Neural networks-based regularization for large-scale medical image reconstruction

In this paper we present a generalized Deep Learning-based approach for solving ill-posed large-scale inverse problems occuring in medical image reconstruction. Recently, Deep Learning methods using iterative neural networks (NNs) and cascaded NNs have been reported to achieve state-of-the-art results with respect to various quantitative quality measures as PSNR, NRMSE and SSIM across different imaging modalities. However, the fact that these approaches employ the application of the forward and adjoint operators repeatedly in the network architecture requires the network to process the whole images or volumes at once, which for some applications is computationally infeasible. In this work, we follow a different reconstruction strategy by strictly separating the application of the NN, the regularization of the solution and the consistency with the measured data. The regularization is given in the form of an image prior obtained by the output of a previously trained NN which is used in a Tikhonov regularization framework. By doing so, more complex and sophisticated network architectures can be used for the removal of the artefacts or noise than it is usually the case in iterative NNs. Due to the large scale of the considered problems and the resulting computational complexity of the employed networks, the priors are obtained by processing the images or volumes as patches or slices. We evaluated the method for the cases of 3D cone-beam low dose CT and undersampled 2D radial cine MRI and compared it to a total variation-minimization-based reconstruction algorithm as well as to a method with regularization based on learned overcomplete dictionaries. The proposed method outperformed all the reported methods with respect to all chosen quantitative measures and further accelerates the regularization step in the reconstruction by several orders of magnitude.

Cardiac MRI Myocardial Functional and Tissue Characterization Detects Early Cardiac Dysfunction in a Mouse Model of Chemotherapy-Induced Cardiotoxicity.

Doxorubicin and doxorubicin-trastuzumab combination chemotherapy have been associated with cardiotoxicity that eventually leads to heart failure and may limit dose-effective cancer treatment. Current diagnostic strategies rely on decreased ejection fraction (EF) to diagnose cardiotoxicity. The aim of this study is to explore the potential of cardiac MR (CMR) imaging to identify imaging biomarkers in a mouse model of chemotherapy-induced cardiotoxicity. A cumulative dose of 25 mg/kg doxorubicin was administered over three weeks using subcutaneous pellets (n = 9, Dox). Another group (n = 9) received same dose of Dox and a total of 10 mg/kg trastuzumab (DT). Mice were imaged at baseline, 5/6 weeks and 10 weeks post-treatment on a 7T MRI system. The protocol included short-axis cine MRI covering the left ventricle (LV) and mid-ventricular short-axis tissue phase mapping (TPM), pre- and post-contrast T1 mapping, T2 mapping and Displacement Encoding with Stimulated Echoes (DENSE) strain encoded MRI. EF, peak myocardial velocities, native T1, T2, extracellular volume (ECV), and myocardial strain were quantified. N = 7 mice were sacrificed for histopathologic assessment of apoptosis at 5/6 weeks. Global peak systolic longitudinal velocity was reduced at 5/6 weeks in Dox (0.6 ± 0.3 vs 0.9 ± 0.3, p = 0.02). In the Dox group, native T1 was reduced at 5/6 weeks (1.3 ± 0.2 ms vs 1.6 ± 0.2 ms, p = 0.02), and relatively normalized at week 10 (1.4 ± 0.1 ms vs 1.6 ± 0.2 ms, p > 0.99). There was no change in EF and other MRI parameters and histopathologic results demonstrated minimal apoptosis in all mice (~1-2 apoptotic cell/high power field), suggesting early-stage cardiotoxicity. In a mouse model of chemotherapy-induced cardiotoxicity using doxorubicin and trastuzumab, advanced CMR shows promise in identifying treatment-related decrease in myocardial velocity and native T1 prior to the onset of cardiomyocyte apoptosis and reduction of EF.