Segmentation Of Left Ventricle Research Articles

Automated DeepLabV3+ based model for left ventricle segmentation on short-axis late gadolinium enhancement-magnetic cardiac resonance imaging images

Accurate segmentation of myocardial scar tissue on late gadolinium enhancement-magnetic cardiac resonance imaging (LGE-CMR) is exceptionally vital for clinical applications, enabling precise diagnosis and effective treatment of various cardiac diseases, such as myocardial infarction and cardiomyopathies. However, the ventricle (LV) variations in the size and shape, artifacts, and image resolution of LGE-CMR has made automatic segmentation of myocardial scar tissue more challenging. While many existing approaches delineate the LV myocardium region using multi-modal segmentation, these models may be computationally complex and suffer from misalignment. Therefore, this study proposed an automatic dual-stage DeepLabV3+ based approach tailored for myocardial scar segmentation on short-axis LGE-MRI exclusively. To segment myocardial scar tissue, the second stage employs the segmented LV chamber from the previous stage. The encoder part of the framework utilizes a MobileNetV2 and ResNet50 backbone for the first and second segmentation, respectively, aiming for optimal resolution of feature maps. Both stages tailor an improved Atrous Spatial Pyramid Pooling module in the DeepLabV3+ model with fine-tuned dilated atrous rates to effectively extract the LV chamber and myocardial scar from the complex LGE-MRI background. Based on the results, the proposed dual-stage network recorded an outstanding segmentation performance, with mean Dice score of 96.02% for LV chamber segmentation and 68.01% for scar tissue extraction.

Вибір блокатора ренін-ангіотензин-альдостеронової системи для лікування серцевої недостатності при гострому міокардиті

Вибір блокатора ренін-ангіотензин-альдостеронової системи для лікування серцевої недостатності при гострому міокардиті

Multi-granularity learning of explicit geometric constraint and contrast for label-efficient medical image segmentation and differentiable clinical function assessment

Automated segmentation is a challenging task in medical image analysis that usually requires a large amount of manually labeled data. However, most current supervised learning based algorithms suffer from insufficient manual annotations, posing a significant difficulty for accurate and robust segmentation. In addition, most current semi-supervised methods lack explicit representations of geometric structure and semantic information, restricting segmentation accuracy. In this work, we propose a hybrid framework to learn polygon vertices, region masks, and their boundaries in a weakly/semi-supervised manner that significantly advances geometric and semantic representations. Firstly, we propose multi-granularity learning of explicit geometric structure constraints via polygon vertices (PolyV) and pixel-wise region (PixelR) segmentation masks in a semi-supervised manner. Secondly, we propose eliminating boundary ambiguity by using an explicit contrastive objective to learn a discriminative feature space of boundary contours at the pixel level with limited annotations. Thirdly, we exploit the task-specific clinical domain knowledge to differentiate the clinical function assessment end-to-end. The ground truth of clinical function assessment, on the other hand, can serve as auxiliary weak supervision for PolyV and PixelR learning. We evaluate the proposed framework on two tasks, including optic disc (OD) and cup (OC) segmentation along with vertical cup-to-disc ratio (vCDR) estimation in fundus images; left ventricle (LV) segmentation at end-diastolic and end-systolic frames along with ejection fraction (LVEF) estimation in two-dimensional echocardiography images. Experiments on nine large-scale datasets of the two tasks under different label settings demonstrate our model’s superior performance on segmentation and clinical function assessment.

Changes in the cardiovascular profile in patients 3 and 12 months after COVID-19 pneumonia: parameters of arterial stiffness, global longitudinal strain, and diastolic function of the left ventricle

Aim. To study changes in the brachial – ankle pulse wave velocity (baPWV), ankle – brachial index (ABI), diastolic function, and global longitudinal strain of the left ventricle (LV) 3 and 12 months after COVID-19 pneumonia.Materials and methods. The dynamics of vascular age and LV global longitudinal strain was studied in 154 patients 3 and 12 months after COVID-19 pneumonia (51 ± 12 years, 48% were women). The control group consisted of 55 sexand age-matched individuals.Results. During the follow-up, the average baPWV decreased (13.2 [11.8; 15.1] cm / sec vs. 13.0 [11.8; 14.1] cm/ s; p < 0.001), and the frequency of its elevated values declined (45.4 vs. 35.1%; p = 0.008). The average ABI increased (1.09 [1.04; 1.14] vs. 1.11 [1.06; 1.17]; p = 0.012), but remained within the normal range. LV global longitudinal strain (LV GLS) (–19.6 ± 2.2 and –19.7 ± 2.5%; p = 0.854) and the frequency of reduced LV GLS (21.4 and 26.6%; p = 0.268) did not change significantly and did not differ from values in the control group. Global longitudinal strain in the LV basal inferoseptal segment improved (–19.2 ± 3.6% vs. –20.1 ± 4.0%; p = 0.032). The early diastolic mitral annular velocity decreased (8.4 ± 3.0 cm / s vs. 8.0 ± 2.5 cm / s; p = 0.023). The LV isovolumic relaxation time was greater than in the control group (101.8 ± 22.3 ms at the 1st visit vs. 92.9 ± 21.5 ms; p = 0.012; 105.9 ± 21.9 ms vs. 92.9 ± 21.5 ms at the 2nd visit; p < 0.001). A positive correlation was found between baPWV (r = 0.209; p = 0.009) and ABI (r = 0.190; p = 0.021) and strain parameters of the LV basal segments 12 months after discharge.Conclusion. Patients with optimal visualization on echocardiography at 12 months after COVID-19 pneumonia, compared to the results of the examination 3 months after the disease, had deteriorated parameters of LV diastolic function. LV GLS was within the grey zone and did not change significantly. An improvement in arterial stiffness was noted, associated with an improvement in the strain of basal LV segments.

Dynamics of heart rhythm disorders in combatants with acute myocarditis during a 6-month follow-up

The main factors of an unfavorable prognosis, are life-threatening heart rhythm disorders which determine the severity of the clinical course of myocarditis, in addition to impaired contractility and dilatation of the left ventricle (LV). The aim of the study was to assess the frequency and nature of arrhythmias and establish predictors of their long-term persistence, assess the severity of anxiety and depression, changes in heart rate variability in combatants on the background of changes in the structural and functional state of the left ventricle during a 6-month follow-up. The study included 52 male servicemen (combatants) with acute myocarditis (AM) with an average age of 33.4±2.5 years. All patients were assessed by Questionnaire according to the Hospital Anxiety and Depression Scale (HADS) and underwent 24 hour ECG monitoring, transthoracic echocardiography and cardiac magnetic resonance (CMR) with late enhancement using gadovist. At the onset of AM frequent supraventricular (30.7% of cases) and frequent ventricular (42.3% of cases) extrasystolic arrhythmia, as well as paroxysms of non-sustained ventricular tachycardia (26.9% of cases) on the background of heart rhythm variability disorders were recorded. After 6 months, there was an improvement in the structural and functional state of the heart, which was characterized by a decrease in dilatation and an improvement of LV systolic function, a decrease in the volume of the inflammatory lesion and the total number of affected LV segments revealed on cardiac MRI, which was associated with a decrease in the number of ventricular rhythm disorders and reduction of clinical manifestations of anxiety according to the HADS scale. The obtained data indicate that the persistence of supraventricular rhythm disorders is associated with the long-term presence of anxiety and low values of standard deviation of RR intervals (SDNN) and root mean square of successive differences between normal heartbeats (RMSSD). Also predictors of persistence of non-sustained ventricular tachycardia paroxysms were established – SDNN value ≤80.0 ms, LV end-diastolic volume index ≥95 ml/m2, total number of affected LV segments ≥6 segments, presence of delayed contrasting in ≥3 LV segments, pre­sence of ≥12 points on the HADS scale and predictors of atrial fibrillation – SDNN value ≤80.0 ms, RMSSD value ≤12.0 ms, total number of affected segments of LV≥6 seg. and ≥12 points on the HADS anxiety scale, determined at the onset of AM.

Analysis of left ventricle regional myocardial motion for cardiac radioablation: Left ventricular motion analysis.

Left ventricle (LV) regional myocardial displacement due to cardiac motion was assessed using cardiovascular magnetic resonance (CMR) cine images to establish region-specific margins for cardiac radioablation treatments. CMR breath-hold cine images and LV myocardial tissue contour points were analyzed for 200 subjects, including controls (n=50) and heart failure (HF) patients with preserved ejection fraction (HFpEF, n=50), mid-range ejection fraction (HFmrEF, n=50), and reduced ejection fraction (HFrEF, n=50). Contour points were divided into segments according to the 17-segment model. For each patient, contour point displacements were determined for the long-axis (all 17 segments) and short-axis (segments 1-12) directions. Mean overall, tangential (longitudinal or circumferential), and normal (radial) displacements were calculated for the 17 segments and for each segment level. The greatest overall motion was observed in the control group-long axis: 4.5±1.2mm (segment 13 [apical anterior] epicardium) to 13.8±3.0mm (segment 6 [basal anterolateral] endocardium), short axis: 4.3±0.8mm (segment 9 [mid inferoseptal] epicardium) to 11.5±2.3mm (segment 1 [basal anterior] endocardium). HF patients exhibited lesser motion, with the smallest overall displacements observed in the HFrEF group-long axis: 4.3±1.7mm (segment 13 [apical anterior] epicardium) to 10.6±3.4mm (segment 6 [basal anterolateral] endocardium), short axis: 3.9±1.3mm (segment 8 [mid anteroseptal] epicardium) to 7.4±2.8mm (segment 1 [basal anterior] endocardium). This analysis provides an estimate of epicardial and endocardial displacement for the 17 segments of the LV for patients with normal and impaired LV function. This reference data can be used to establish treatment planning margin guidelines for cardiac radioablation. Smaller margins may be used for patients with higher degree of impaired heart function, depending on the LV segment.

Deconfounded multi-organ weakly-supervised semantic segmentation via causal intervention

In weakly-supervised semantic segmentation, obtaining the class activation maps for pseudo masks is crucial. Since multiple organs appear in the same medical image, it is reasonable to obtain the activation maps of each organ by the organ-level features instead of the image-level features. The image-level features are decomposed into the organ-level features, yet the prior anatomical knowledge makes a spurious association between the image-level and organ-level features. To this end, we apply the causal intervention to cut off the spurious association and propose a novel deconfounded multi-organ weakly-supervised semantic segmentation (DeMos) method. Based on the original class activation mapping (CAM) method, the model is retrained to learn the deconfounded features of each organ via cross-attention, and we approximate the expectation of the intervention instead of the traditional likelihood. When the model converges, we extract the activation maps by CAM. Our method not only generates high-quality pseudo masks on the CHAOS, ACDC and ProMRI datasets, but is also applicable to other CAM variants. Furthermore, with the refinement, DeMos achieves the dice similarity coefficient of 93.26% on the task of the left ventricle segmentation, which outperforms the state-of-the-art methods.

Left atrial strain in patients with β-thalassemia major: a cross-sectional CMR study.

The aim of this cross-sectional study was to investigate the association of left atrial (LA) strain parameters with demographics, clinical data, cardiovascular magnetic resonance (CMR) findings, and cardiac complications (heart failure and arrhythmias) in a cohort of patients with β-thalassemia major (β-TM). We considered 264 β-TM patients (133 females, 36.79 ± 11.95 years) consecutively enrolled in the Extension-Myocardial Iron Overload in Thalassemia (E-MIOT) project. Moreover, we included 35 sex- and age-matched healthy controls (14 females, mean age 37.36 ± 17.52 years). Reservoir, conduit, and booster LA functions were analysed by CMR feature tracking using dedicated software. Compared to the healthy control group, β-TM patients demonstrated lower LA reservoir strain and booster strains, as well as LA reservoir and booster strain rates. However, no differences were found in LA conduit deformation parameters. In β-TM patients, ageing, sex, and left ventricle (LV) volume indexes were independent determinants of LA strain parameters. The number of segments with late gadolinium enhancement (LGE) significantly correlated with all LA strain parameters, with the exception of the LA conduit rate. Patients with cardiac complications exhibited significantly impaired strain parameters compared to patients without cardiac complications. In patients with β-TM, LA strain parameters were impaired compared to control subjects, and they exhibited a significant correlation with the number of LV segments with LGE. Furthermore, patients with cardiac complications had impaired left atrial strain parameters. Clinical relevance statement In patients with β-thalassemia major, left atrial strain parameters were impaired compared to control subjects and emerged as a sensitive marker of cardiac complications, stronger than cardiac iron levels. • Compared to healthy subjects, β-thalassemia major patients demonstrated significantly lower left atrial reservoir strain and booster strains, as well as left atrial reservoir and booster strain rates. • In β-thalassemia major, ageing, sex, and left ventricular volume indexes were independent determinants of left atrial strain parameters, while left atrial strain parameters were not correlated with myocardial iron overload. • An independent association between reduced left atrial strain parameters and a history of cardiac complications was found in β-thalassemia major patients.

Active learning for left ventricle segmentation in echocardiography

Background and ObjectiveTraining deep learning models for medical image segmentation require large annotated datasets, which can be expensive and time-consuming to create. Active learning is a promising approach to reduce this burden by strategically selecting the most informative samples for segmentation. This study investigates the use of active learning for efficient left ventricle segmentation in echocardiography with sparse expert annotations. MethodsWe adapt and evaluate various sampling techniques, demonstrating their effectiveness in judiciously selecting samples for segmentation. Additionally, we introduce a novel strategy, Optimised Representativeness Sampling, which combines feature-based outliers with the most representative samples to enhance annotation efficiency. ResultsOur findings demonstrate a substantial reduction in annotation costs, achieving a remarkable 99% upper bound performance while utilising only 20% of the labelled data. This equates to a reduction of 1680 images needing annotation within our dataset. When applied to a publicly available dataset, our approach yielded a remarkable 70% reduction in required annotation efforts, representing a significant advancement compared to baseline active learning strategies, which achieved only a 50% reduction. Our experiments highlight the nuanced performance of diverse sampling strategies across datasets within the same domain. ConclusionsThe study provides a cost-effective approach to tackle the challenges of limited expert annotations in echocardiography. By introducing a distinct dataset, made publicly available for research purposes, our work contributes to the field's understanding of efficient annotation strategies in medical image segmentation.

Fetal Left Ventricle Function Evaluated by Two-Dimensional Speckle-Tracking Echocardiography across Clinical Stages of Severity in Growth-Restricted Fetuses.

Fetal growth restriction (FGR) can result in adverse perinatal outcomes due to cardiac dysfunction. This study used 2D speckle-tracking echocardiography to assess left ventricle (LV) longitudinal strain across FGR severity stages. A prospective longitudinal cohort study measured global (GLS) and segmental LV longitudinal strain in FGR fetuses, with evaluations conducted at various time points. FGR was classified into subtypes based on published criteria using fetal weight centile and Doppler parameters. A linear mixed model was employed to analyze repeated measures and compare Z-score measurements between groups throughout gestational age. The study included 40 FGR fetuses and a total of 107 evaluations were performed: 21 from small for gestational age (SGA), 74 from the FGR stage I, and 12 from the FGR stage ≥ II. The results indicate that SGA and stage I FGR fetuses exhibit higher LV GLS than stages ≥ II. Throughout gestation, SGA and FGR stage I fetuses showed similar behavior with consistently better LV GLS values when compared to FGR stages ≥ II. No significant differences were observed in LV GLS strain behavior between SGA and FGR stage I. In conclusion, all FGRs show signs of early cardiac dysfunction, with severe cases demonstrating significantly a lower LV GLS when compared to mild cases, suggesting deterioration of cardiac dysfunction with progression of fetal compromise.

Redundant co-training: Semi-supervised segmentation of medical images using informative redundancy

Pseudo-labeling, consistency regularization, and co-training are common paradigms for semi-supervised learning. In this paper, we propose a novel method based on co-training and pseudo-labeling for the semi-supervised segmentation of the left ventricle. Our co-training strategy is novel and unlike most previous works does not rely on using multiple-view datasets, performing weak/strong augmentations on the input images or perturbations on the networks. We proposed creating redundant labels by utilizing the provided ground-truths and training networks segmenting different overlapping regions corresponding to the created labels. Although the new labels seem to be redundant, we demonstrated that they provide valuable information to the networks. The predictions of the redundant networks (which are trained on the redundant labels) can be used in the pixels where the primary network’s predictions are not reliable. This enables extracting a secondary source of information without requiring any additional ground-truths. The common practice in pseudo-labeling is using the reliable predictions of the unlabeled data and discarding the unreliable ones. However, we proposed utilizing predictions from the redundant networks to generate pseudo-labels for the unreliable pixels in the primary network’s predictions, rather than simply discarding them. We validated our method on two left ventricle segmentation datasets, and it surpassed the state-of-the-art semi-supervised learning approaches. Furthermore, we conducted extensive studies to analyze the proposed method from different aspects. Implementation of our work is available at https://github.com/behnam-rahmati/redundant-cotraining.

A Motion-Aware DNN Model with Edge Focus Loss and Quality Control for Short-Axis Left Ventricle Segmentation of Cine MR Sequences.

Accurate segmentation of the left ventricle myocardium is the key step of automatic assessment of cardiac function. However, the current methods mainly focus on the end-diastolic and the end-systolic frames in cine MR sequences and lack the attention to myocardial motion in the cardiac cycle. Additionally, due to the lack of fine segmentation tools, the simplified approach, excluding papillary muscles and trabeculae from myocardium, is applied in clinical practice. To solve these problems, we propose a motion-aware DNN model with edge focus loss and quality control in this paper. Specifically, the bidirectional ConvLSTM layer and a new motion attention layer are proposed to encode motion-aware feature maps, and an edge focus loss function is proposed to train the model to generate the fine segmentation results. Additionally, a quality control method is proposed to filter out the abnormal segmentations before subsequent analyses. Compared with state-of-the-art segmentation models on the public dataset and the in-house dataset, the proposed method has obtained high segmentation accuracy. On the 17-segment model, the proposed method has obtained the highest Pearson correlation coefficient at 14 of 17 segments, and the mean PCC of 85%. The experimental results highlight the segmentation accuracy of the proposed method as well as its availability to substitute for the manually annotated boundaries for the automatic assessment of cardiac function.

Motion correction of wide-detector 4DCT images for cardiac resynchronization therapy planning

BackgroundLead placement at the latest mechanically activated left ventricle (LV) segments is strongly correlated with response to cardiac resynchronization therapy (CRT). We demonstrate the feasibility of a cardiac 4DCT motion correction algorithm (ResyncCT) in estimating LV mechanical activation for guiding lead placement in CRT. MethodsSubjects with full cardiac cycle 4DCT images acquired using a wide-detector CT scanner for CRT planning/upgrade were included. 4DCT images exhibited motion artifact-induced false-dyssynchrony, hindering LV mechanical activation time estimation. Motion-corrupted images were processed with ResyncCT to yield motion-corrected images. Time to onset of shortening (TOS) was estimated in each of 72 endocardial segments. A false-dyssynchrony index (FDI) was used to quantify the extent of motion artifacts in the uncorrected and the ResyncCT images. After motion correction, the change in classification of LV free-wall segments as optimal target sites for lead placement was investigated. ResultsTwenty subjects (70.7 ​± ​13.9 years, 6 female) were analyzed. Motion artifacts in the ResyncCT-processed images were significantly reduced (FDI: 28.9 ​± ​9.3 ​% vs 47.0 ​± ​6.0 ​%, p ​< ​0.001). In 10 (50 ​%) subjects, ResyncCT motion correction yielded statistically different TOS estimates (p ​< ​0.05). Additionally, 43 ​% of LV free-wall segments were reclassified as optimal target sites for lead placement after motion correction. ConclusionsResyncCT significantly reduced motion artifacts in wide-detector cardiac 4DCT images, yielded statistically different time to onset of shortening estimates, and changed the location of optimal target sites for lead placement. These results highlight the potential utility of ResyncCT motion correction in CRT planning when using wide-detector 4DCT imaging.

Deep learning based automated left ventricle segmentation and flow quantification in 4D flow cardiac MRI.

4D flow MRI enables assessment of cardiac function and intra-cardiac blood flow dynamics from a single acquisition. However, due to the poor contrast between the chambers and surrounding tissue, quantitative analysis relies on the segmentation derived from a registered cine MRI acquisition. This requires an additional acquisition and is prone to imperfect spatial and temporal inter-scan alignment. Therefore, in this work we developed and evaluated deep learning-based methods to segment the left ventricle (LV) from 4D flow MRI directly. We compared five deep learning-based approaches with different network structures, data pre-processing and feature fusion methods. For the data pre-processing, the 4D flow MRI data was reformatted into a stack of short-axis view slices. Two feature fusion approaches were proposed to integrate the features from magnitude and velocity images. The networks were trained and evaluated on an in-house dataset of 101 subjects with 67,567 2D images and 3030 3D volumes. The performance was evaluated using various metrics including Dice, average surface distance (ASD), end-diastolic volume (EDV), end-systolic volume (ESV), LV ejection fraction (LVEF), LV blood flow kinetic energy (KE) and LV flow components. The Monte Carlo dropout method was used to assess the confidence and to describe the uncertainty area in the segmentation results. Among the five models, the model combining 2D U-Net with late fusion method operating on short-axis reformatted 4D flow volumes achieved the best results with Dice of 84.52% and ASD of 3.14mm. The best averaged absolute and relative error between manual and automated segmentation for EDV, ESV, LVEF and KE was 19.93ml (10.39%), 17.38ml (22.22%), 7.37% (13.93%) and 0.07mJ (5.61%), respectively. Flow component results derived from automated segmentation showed high correlation and small average error compared to results derived from manual segmentation. Deep learning-based methods can achieve accurate automated LV segmentation and subsequent quantification of volumetric and hemodynamic LV parameters from 4D flow MRI without requiring an additional cine MRI acquisition.

Improving a Deep Learning Model to Accurately Diagnose LVNC.

Accurate diagnosis of Left Ventricular Noncompaction Cardiomyopathy (LVNC) is critical for proper patient treatment but remains challenging. This work improves LVNC detection by improving left ventricle segmentation in cardiac MR images. Trabeculated left ventricle indicates LVNC, but automatic segmentation is difficult. We present techniques to improve segmentation and evaluate their impact on LVNC diagnosis. Three main methods are introduced: (1) using full 800 × 800 MR images rather than 512 × 512; (2) a clustering algorithm to eliminate neural network hallucinations; (3) advanced network architectures including Attention U-Net, MSA-UNet, and U-Net++.Experiments utilize cardiac MR datasets from three different hospitals. U-Net++ achieves the best segmentation performance using 800 × 800 images, and it improves the mean segmentation Dice score by 0.02 over the baseline U-Net, the clustering algorithm improves the mean Dice score by 0.06 on the images it affected, and the U-Net++ provides an additional 0.02 mean Dice score over the baseline U-Net. For LVNC diagnosis, U-Net++ achieves 0.896 accuracy, 0.907 precision, and 0.912 F1-score outperforming the baseline U-Net. Proposed techniques enhance LVNC detection, but differences between hospitals reveal problems in improving generalization. This work provides validated methods for precise LVNC diagnosis.

Dual-Branch TransV-Net for 3-D Echocardiography Segmentation

Segmentation of left and right ventricles from 3D echocardiographic images is premised and key for quantitative analysis of cardiac function, which is important for pediatric cardiac diagnosis. Compared with 2D echocardiography, 3D echocardiography can fully represent ventricular structure without geometric inference. However, it usually takes experts several hours to obtain a 3D segmentation mask of the left and right ventricles. Therefore, a fast and automatic segmentation method is highly desired. Unfortunately, 3D echocardiography suffers from low contrast, unclear left and right ventricles borders, and blind zone. Existing segmentation methods usually have poor performance at ventricular boundaries. To deal with these problems, we propose a novel dual-branch TransV-Net (DBTV). DBTV comprises two parallel, interleaved, and relatively independent V-shaped encoder-decoder branches. The main branch acts on the original data to extract image features, and the auxiliary branch acts on edge maps to extract the additional edge features. To suppress noise and enhance the edge information, extra concatenations are added to bridge the features from the main and auxiliary branches. To reduce object missing caused by blind zone, a 3D transformer-based module is proposed in the bottom layer of the dual-branch structure to extract the global contexts. We do experiments on a self-collected dataset with 120 3D echocardiographic images from 60 cardiac sequences, and the dice score of 0.913 and 0.880 are obtained in the left and right ventricle segments, respectively. Each inference takes about two-thirds of a second for a single 3D frame.

Parametrization of subsegmental infarcts using high spatial resolution 2DSTE and synthetic ultrasonic data

Abstract Introduction: The purpose of this study was to assess the performance of a set of parameters in characterizing simulated infarcts in left ventricular (LV) models with variations in size and transmural extent. Material and methods: The deformation of the LV models with different infarct sizes was simulated using the Finite Element Method. These simulations provided meshes that were used to generate synthetic ultrasonic data within the FIELD II package. The strain components (longitudinal and circumferential) were then estimated over small subsegments of the of segments 7 and 12 (according to 17-segment left ventricle segmentation standard proposed by the American Heart Association - AHA17), using a hierarchical block matching method. The strain maps obtained were utilized to calculate the Strain Drop Factor (SDF) maps, which represent the percentage ratio of strain observed in the subsegments of the studied model to that observed in the healthy model. Infarct segmentation was performed using these maps, and various parameters were derived, including Infarct Cross-Section Area (ICSA), relative ICSA, Transmurality Ratio (TR), Mean Infarct Transmurality (MIT), strain drop factor in the infarcted region (SDFi), and Strain Contrast (SC). Results: The estimates of ICSA, SC, MIT, and SDFI showed good repeatability and demonstrated the ability to provide a quantitative assessment of the size and transmural extent of the infarcts. Conclusions: The study findings suggest that the evaluated parameters, including ICSA, SC, MIT, and SDFI, can be reliably used to assess the size and transmural extent of infarcts. These parameters offer a quantitative approach for characterizing infarcts based on strain analysis and have the potential to contribute to the diagnosis and evaluation of myocardial infarctions.

CAT-Seg: cascaded medical assistive tool integrating residual attention mechanisms and Squeeze-Net for 3D MRI biventricular segmentation

Cardiac image segmentation is a critical step in the early detection of cardiovascular disease. The segmentation of the biventricular is a prerequisite for evaluating cardiac function in cardiac magnetic resonance imaging (CMRI). In this paper, a cascaded model CAT-Seg is proposed for segmentation of 3D-CMRI volumes. CAT-Seg addresses the problem of biventricular confusion with other regions and localized the region of interest (ROI) to reduce the scope of processing. A modified DeepLabv3+ variant integrating SqueezeNet (SqueezeDeepLabv3+) is proposed as a part of CAT-Seg. SqueezeDeepLabv3+ handles the different shapes of the biventricular through the different cardiac phases, as the biventricular only accounts for small portion of the volume slices. Also, CAT-Seg presents a segmentation approach that integrates attention mechanisms into 3D Residual UNet architecture (3D-ResUNet) called 3D-ARU to improve the segmentation results of the three major structures (left ventricle (LV), Myocardium (Myo), and right ventricle (RV)). The integration of the spatial attention mechanism into ResUNet handles the fuzzy edges of the three structures. The proposed model achieves promising results in training and testing with the Automatic Cardiac Diagnosis Challenge (ACDC 2017) dataset and the external validation using MyoPs. CAT-Seg demonstrates competitive performance with state-of-the-art models. On ACDC 2017, CAT-Seg is able to segment LV, Myo, and RV with an average minimum dice symmetry coefficient (DSC) performance gap of 1.165%, 4.36%, and 3.115% respectively. The average maximum improvement in terms of DSC in segmenting LV, Myo and RV is 4.395%, 6.84% and 7.315% respectively. On MyoPs external validation, CAT-Seg outperformed the state-of-the-art in segmenting LV, Myo, and RV with an average minimum performance gap of 6.13%, 5.44%, and 2.912% respectively.

Assessment of global and segmental strain in heterozygous APE1/Ref-1 transgenic mice using cardiac magnetic resonance imaging

Abstract Background Apurinic/apyrimidinic-endonuclease1(APE1/Ref-1) is a redox factor necessary for DNA binding and transcription regulator in ROS generation. The association between APE1/Ref-1 and heart failure(HF) has been observed in previous studies. Heterozygous APE1/Ref-1 transgenic mice displayed endothelial cell dysfunction with systemic hypertension. Global strain(GS) measures the myocardium's deformation along its axis and is known to be a sensitive marker of subclinical HF in hypertensive patients with preserved ejection fraction. Purpose This study aims to investigate the functional dynamics and GS of the heart in heterozygous APE1/Ref-1 mice to help identify the association between APE1/Ref-1 and early signs of HF and translate to clinical applications. Methods A study was conducted to assess the functional parameters of both the left ventricle(LV) and right ventricle (RV) and perform LV global and segmental strain analysis using cardiac magnetic resonance imaging(CMR) in heterozygous APE1/Ref-1 transgenic mice(n=4) and wild type mice (n=4) of age 10-12 weeks. A 7T CMR scanner and FSE T1 sequence were used to acquire short-axis views of mouse cardiac images. LV and RV functional parameters including ejection fraction(EF), stroke volume(SV), cardiac output, peak filling rate, peak emptying rate and strain indices including global circumferential strain, global radial strain, segmental circumferential strain, and segmental radial strain were calculated with a software dedicated to analyzing CMR images. Global LV and RV function was quantified both manually and using an AI-based algorithm of LV and RV segmentation. The strain analysis was performed using the strain medical image tracking toolbox module. Additionally, reproducibility testing was done to validate the accuracy of AI-driven calculations. Results We found that heterozygous APE1/Ref-1 mice displayed a decrease in SV(8.00±0.01ul vs 9.75±0.96ul, p=0.01) and LVEF(58.75±2.63% vs 66.50±2.89%, p=0.007) compared to controls while the functional parameters of RV were similar. Interestingly, the circumferential strain rate of LV was significantly reduced in APE1/Ref-1 mice compared to controls(p&amp;lt;0.001), particularly in the systolic phase. In addition, the study showed that manually-calculated and AI-derived indices were within an acceptable margin of error, rendering the AI system reproducible and reliable in calculating volumetric parameters and strain of the mice's hearts. Conclusions This study demonstrated that LV global circumferential strain rate, along with SV and LVEF decreased in heterozygous APE1/Ref-1 transgenic mice. These parameters could serve as markers for assessing early functional changes of mice hearts in a non-invasive manner. Volumetric and strain indices measured by AI-based analysis of mouse CMR images are reproducible parameters for cardiovascular research in small animals.LV strain analysis of mice CMRTable of volumetric parameters

Assessment of left ventricular myocardial fibrosis in adult patients with Ebstein's anomaly: a retrospective cohort study based on cardiac magnetic resonance and histopathological samples

Abstract Background The relationship between Ebstein's Anomaly (EA) and myocardial fibrosis in the left ventricle (LV) has been debated. This study aimed to determine the prevalence of replacement fibrosis in the LV using cardiac magnetic resonance (CMR) imaging and correlate it with histopathological findings. The study also investigated whether LV fibrosis is an independent risk factor for cardiovascular disease (CVD) mortality and developed a risk score for this purpose. Methods We performed a twelve-years (2009-2021) retrospective cohort study of adult patients with EA who underwent CMR. The CMR evaluation included a comprehensive assessment of myocardial fibrosis by late gadolinium enhancement (LGE). Four post-mortem samples were obtained from our cohort and stained using Masson trichrome to characterize LV-fibrosis. We used Cox-Regression analysis to identify and derive a prediction score that associates LV-fibrosis with CVD-mortality. Results We included 57 adults with EA (52% male, median age 29.52, IQR: 21.24-39.17 years). LGE prevalence by CMR was observed in 52.6% (95% CI: 39.9%-66.0%) in any chamber; LGE-LV in 29.8% (95% CI: 19.5%-42.7%). Histopathological findings revealed a mid-wall pattern with predominantly interstitial fibrosis and minimal replacement fibrosis. LV-fibrosis was associated with increased risk for CVD mortality (HR: 6.02, 95% CI: 1.22-19.91) attributable to lateral and mid-wall LV segment involvement. Our mortality score achieved an overall good prediction capacity (C-statistic: 0.93, Dxy= 0.86). Conclusions There is a high prevalence of replacement LV fibrosis in adults with EA, characterized by specific CMR and histological patterns. Furthermore, LV fibrosis is an independent predictor of CVD mortality, which could be integrated into a risk assessment in clinical management.Figure 1.CMR and pathological imagesFigure 1.Kaplan-Meier analysis