Heart Segmentation Research Articles

Plug-and-play segment anything model improves nnUNet performance.

The automatic segmentation of medical images has widespread applications in modern clinical workflows. The Segment Anything Model (SAM), a recent development of foundational models in computer vision, has become a universal tool for image segmentation without the need for specific domain training. However, SAM's reliance on prompts necessitates human-computer interaction during the inference process. Its performance on specific domains can also be limited without additional adaptation. In contrast, traditional models like nnUNet are designed to perform segmentation tasks automatically during inference and can work well for each specific domain, but they require extensive training on domain-specificdatasets. To leverage the advantages of both foundational and domain-specific models and achieve fully automated segmentation with limited training samples, we propose nnSAM, which combines the robust feature extraction capabilities of SAM with the automatic configuration abilities of nnUNet to enhance the accuracy and robustness of medical image segmentation on smalldatasets. We propose the nnSAM model for small sample medical image segmentation. We made optimizations for this goal via two main approaches: first, we integrated the feature extraction capabilities of SAM with the automatic configuration advantages of nnUNet, which enables robust feature extraction and domain-specific adaptation on small datasets. Second, during the training process, we designed a boundary shape supervision loss based on level set functions and curvature calculations, enabling the model to learn anatomical shape priors from limited annotationdata. We conducted quantitative and qualitative assessments on the performance of our proposed method on four segmentation tasks: brain white matter, liver, lung, and heart segmentation. Our method achieved the best performance across all tasks. Specifically, in brain white matter segmentation using 20 training samples, nnSAM achieved the highest DICE score of 82.77 ( 10.12) % and the lowest average surface distance (ASD) of 1.14 ( 1.03) mm, compared to nnUNet, which had a DICE score of 79.25 ( 17.24) % and an ASD of 1.36 ( 1.63) mm. A sample size study shows that the advantage of nnSAM becomes more prominent under fewer trainingsamples. A comprehensive evaluation of multiple small-sample segmentation tasks demonstrates significant improvements in segmentation performance by nnSAM, highlighting the vast potential of small-samplelearning.

Challenging endomyocardial biopsies' gold standard status in rejection screening: a prospective blinded study on cardiovascular magnetic resonance in heart transplant patients

Abstract Background The detection of rejection is crucial for graft survival after heart transplantation. Endomyocardial biopsies (EMBs) serve as the gold standard for rejection surveillance despite its limitations, while cardiac magnetic resonance (CMR), in conjunction with donor-derived cell-free DNA (dd-cfDNA) analysis, holds promise as a non-invasive approach. Purpose In this prospective blinded study, our aim was to evaluate feasibility of CMR for rejection screening, validating it against EMB, dd-cfDNA, and clinical status. Methods We blindly analyzed 250 studies involving CMR, EMB, dd-cfDNA, and clinical data in 58 heart transplant recipients (17 children and 41 adults). Of these 250 studies, 239 (96%) were assessed within 1 to 28 months after transplantation, and 11 studies (4%) 3-14 years post-transplant. In CMR T1- and T2-mapping analyses, acute rejection was defined by T1 or T2 times exceeding the specified thresholds (T1 ≥ 1060 ms, T2 ≥ 55 ms) in at least 2/16 heart segments. EMB findings were graded according to the ISHLT criteria (2005 and 2013), and acute cellular rejection (ACR) grade ≥ 2 or antibody-mediated rejection (AMR) grade ≥ 1 were considered significant. The cutoff for abnormal dd-cfDNA was set at 0.2%. Solely clinical rejection included patients with absent or normal EMB (ACR grade 0-1 and AMR grade 0), but with symptoms or ECG/echocardiographic findings indicative of acute rejection. Results Our study revealed 22 acute rejections, with 15 confirmed through EMB and 7 diagnosed solely based on clinical indicators. Among the EMB-confirmed cases, 3/15 exhibited AMR grade 1, 10/15 demonstrated ACR grade 2 and 2/15 ACR grade 3. CMR identified rejection in 77% (17/22) of cases. Specifically, it successfully identified rejection in 67% (10/15) of EMB-confirmed cases and 100% (7/7) of solely clinically diagnosed cases. Examining the concordance between CMR findings and dd-cfDNA levels, 35% (6/17) of the CMR-identified rejection cases exhibited abnormal dd-cfDNA, while 30% (5/17) showed normal levels. Notably, in one case, despite normal CMR results, both EMB and dd-cfDNA indicated rejection. Conversely, in four instances where EMB indicated acute rejection, both CMR and dd-cfDNA levels were normal. These findings highlight a potential discrepancy in the diagnostic accuracy of EMB, suggesting its tendency to both over- and underdiagnose acute heart transplant rejection. The integration of CMR and dd-cfDNA in the diagnostic process may provide a more comprehensive and reliable assessment of rejection episodes, offering valuable insights for clinical decision-making in heart transplant recipients. Conclusions EMB presents challenges in both under- and overdiagnosing rejection in heart transplant surveillance. A strategy based on CMR, combined with dd-cfDNA analyses, shows promise as a new gold standard for rejection screening.Flowchart

2D speckle tracking echocardiography and comparison with cardiac magnetic resonance in children with acute myocarditis.

Cardiac magnetic resonance (CMR) plays a major diagnostic role in acute myocarditis (AM) in children as biopsy is rarely performed in this age group. Contribution of standard echocardiography (ECHO) is limited in AM, but speckle tracking echocardiography (STE) quantitatively characterizes myocardial function, with good sensitivity for detecting subclinical left ventricular (LV) dysfunction and regional kinetics disorders beyond the site of inflammation. This work aimed to evaluate the diagnostic potential of STE as compared with CMR findings in pediatric patients with AM. The study was conducted during 2022-2023. Troponin, electrocardiography, ECHO with STE, and CMR with early and late enhancement were performed on each patient. Affected heart segments were analyzed by both STE and CMR, and the correlation of the two methods was assessed. During the study period, 20 children were diagnosed with AM [14 boys, 6 girls; mean age 12 years (median 14)]. On ECHO, three patients had a deviation in LV biometry, and four patients had a mild systolic function disorder. STE showed at least one affected cardiac segment in all patients, most often the inferolateral segment (16/20; 80%). Of the 20 patients, STE showed a reduction in LV global longitudinal strain in 13 (65%) patients. In all patients, CMR identified an inflammatory focus, most frequently inferolateral (15/20; 75%). The strongest accordance between STE and CMR was observed for the involvement of anterolateral segments (k = 0.88) and the weakest for inferoseptal damage (k = 0.4). STE can provide important diagnostic information in pediatric patients with AM. This modality supports the detection of early regional edema and subclinical myocardial dysfunction and can determine the impairment severity. STE is non-invasive and repeatable without the need for special patient preparation or for general anesthesia.

PCA-Net: a heart segmentation model based on the meta-learning method

PCA-Net: a heart segmentation model based on the meta-learning method

3D augmentation for volumetric whole heart segmentation

Data augmentation is a technique usually deployed to mitigate the possible performance limitation from training a neural network model on a limited dataset, especially in the medical domain. This paper presents a study on effects of applying different rotation settings to augment cardiac volumes from the Multi-modality Whole Heart Segmentation dataset, in order to improve the segmentation performance. This study presents a comparison between conventional 2D (slice-wise) rotation primarily on the axial axis, 3D (volume-wise) rotation, and our proposed rotation setting that takes into account possible cardiac alignment according to its anatomy. The study has suggested two key considerations: 2D slice-wise rotation should be avoided when using 3D data for segmentation, due to intrinsic structural correlation between subsequent slices, and that 3D rotations may help improve segmentation performance on data previously unseen to the model.

FPL+: Filtered Pseudo Label-Based Unsupervised Cross-Modality Adaptation for 3D Medical Image Segmentation.

Adapting a medical image segmentation model to a new domain is important for improving its cross-domain transferability, and due to the expensive annotation process, Unsupervised Domain Adaptation (UDA) is appealing where only unlabeled images are needed for the adaptation. Existing UDA methods are mainly based on image or feature alignment with adversarial training for regularization, and they are limited by insufficient supervision in the target domain. In this paper, we propose an enhanced Filtered Pseudo Label (FPL+)-based UDA method for 3D medical image segmentation. It first uses cross-domain data augmentation to translate labeled images in the source domain to a dual-domain training set consisting of a pseudo source-domain set and a pseudo target-domain set. To leverage the dual-domain augmented images to train a pseudo label generator, domain-specific batch normalization layers are used to deal with the domain shift while learning the domain-invariant structure features, generating high-quality pseudo labels for target-domain images. We then combine labeled source-domain images and target-domain images with pseudo labels to train a final segmentor, where image-level weighting based on uncertainty estimation and pixel-level weighting based on dual-domain consensus are proposed to mitigate the adverse effect of noisy pseudo labels. Experiments on three public multi-modal datasets for Vestibular Schwannoma, brain tumor and whole heart segmentation show that our method surpassed ten state-of-the-art UDA methods, and it even achieved better results than fully supervised learning in the target domain in some cases.

Heart and great vessels segmentation in congenital heart disease via CNN and conditioned energy function postprocessing.

The segmentation of the heart and great vessels in CT images of congenital heart disease (CHD) is critical for the clinical assessment of cardiac anomalies and the diagnosis of CHD. However, the diverse types and abnormalities inherent in CHD present significant challenges to comprehensive heart segmentation. We proposed a novel two-stage segmentation approach, integrating a Convolutional Neural Network (CNN) with a postprocessing method with conditioned energy function for pulmonary and aorta. The initial stage employs a CNN enhanced by a gated self-attention mechanism for the segmentation of five primary heart structures and two major vessels. Subsequently, the second stage utilizes a conditioned energy function specifically tailored to refine the segmentation of the pulmonary artery and aorta, ensuring vascular continuity. Our method was evaluated on a public dataset including 110 3D CT volumes, encompassing 16 CHD variants. Compared to prevailing segmentation techniques (U-Net, V-Net, Unetr, dynUnet), our approach demonstrated improvements of 1.02, 1.04, and 1.41% in Dice Coefficient (DSC), Intersection over Union (IOU), and the 95th percentile Hausdorff Distance (HD95), respectively, for heart structure segmentation. For the two great vessels, the enhancements were 1.05, 1.07, and 1.42% in these metrics. The outcomes on the public dataset affirm the efficacy of our proposed segmentation method. Precise segmentation of the entire heart and great vessels can significantly aid in the diagnosis and treatment of CHD, underscoring the clinical relevance of our findings.

Integrating myocardial metabolic imaging and stress myocardial contrast echocardiography to improve the diagnosis of coronary microvascular diseases in rabbits.

Persistent challenges associated with misdiagnosis and underdiagnosis of coronary microvascular disease (CMVD) necessitate the exploration of noninvasive imaging techniques to enhance diagnostic accuracy. Therefore, we aimed to integrate multimodal imaging approaches to achieve a higher diagnostic rate for CMVD using high-quality myocardial metabolism imaging (MMI) and myocardial contrast echocardiography (MCE). This combination diagnostic strategy may help address the urgent need for improved CMVD diagnosis. In this study, we established five distinct pretreatment groups, each consisting of nine male rabbit: a fasted group, a nonfasted group, a sugar load group, an acipimox group, and a combination group of nonfasted rabbits administered insulin. Moreover, positron emission tomography-computed tomography (PET/CT) scan windows were established at 30-, 60-, and 90-minute intervals. We developed 10 CMVD models and conducted a diagnosis of CMVD through an integrated analysis of MMI and MCE, including image acquisition and processing. For each heart segment, we calculated the standardized uptake value (SUV) based on body weight (SUVbw), as well as certain ratios of SUV including SUV of the heart (SUVheart) to that of the liver (SUVliver) and SUVheart to SUV of the lung (SUVlung). Additionally, we obtained three coronary SUVbw uptake values. To clarify the relationship between SUVbw uptake values and echocardiographic parameters of the myocardial contrast agent more thoroughly, we conducted a comprehensive analysis across different pretreatment protocols. Receiver operating characteristic (ROC) curve analysis was employed to evaluate the diagnostic accuracy of each parameter in the context of CMVD. In the context of MMI, the nonfasted-plus-insulin group, as observed during the 60-minute examination, exhibited a noteworthy total 18F-fluorodeoxyglucose (18F-FDG) uptake of 47.44±6.53 g/mL, which was found to be statistically different from the other groups. To ascertain the reliability of the results, two double-blind investigators independently assessed the data and achieved a good level of agreement, according to the intraclass correlation coefficient (ICC) (0.957). The SUVbw of the nonfasted-plus-insulin group exhibited a moderate correlation with the microvascular blood flow reserve (MBFR) parameters derived from the MCE examination, as evidenced by a r value of 0.686. For the diagnosis of CMVD disease, the diagnostic accuracy of the combined diagnostic method [area under the curve (AUC) =0.789; 95% confidence interval (CI): 0.705-0.873] was significantly higher than that of the MBFR (AUC =0.697; 95% CI: 0.597-0.797) and SUVbw (AUC =0.715; 95% CI: 0.622-0.807) methods (P<0.05). Our study demonstrated the feasibility of a simple premedication approach involving free feeding and intravenous insulin in producing high-quality gated heart 18F-FDG PET/CT images in adult male New Zealand white rabbits. This technique holds considerable potential for ischemic heart disease research in rabbits and can enhance CMVD diagnosis via the comprehensive assessment of myocardial metabolism and perfusion.

FORENSIC-MEDICAL FEATURES OF THE HEART IN CASES OF ACUTE CORONARY INSUFFICIENCY WITH SUDDEN CARDIAC DEATH

The article provides an assessment of changes in the endocardium of the atria and ventricles of the heart in cases of sudden cardiac death due to acute coronary insufficiency. In cases of sudden death due to acute coronary insufficiency, distinct focal pathological changes were observed in the connective tissue of the endocardium of the left atrium and ventricle, manifested by loss of fibrillar structures and fragmentation of its non-cellular elements. This corresponded to local contractile changes in the cardiomyocytes, localised in the subendocardial space of the affected parts of the heart, leading to local fragmentation and an increase in the thickness of the endocardial folds. In addition, the loss of their characteristic waviness and the acquisition of linearity were observed. These changes are considered to be manifestations of morpho-functional disorders of the endocardium of the specified heart segments. The aim of the study was to determine the characteristics of endocardial changes in the atria and ventricles in cases of sudden death due to acute ischaemic heart disease. Materials and Methods. The study material included 12 hearts from young and middle-aged individuals without cardiovascular disease (control group) and 12 hearts from working-age individuals who died suddenly of acute ischaemic heart disease. Conclusions: At the time of death from acute ischaemic heart disease, there were pronounced focal pathological changes in cellular and non-cellular components in the connective tissue of the endocardium of the left atrium and ventricle. These changes caused focal contractile changes in cardiomyocytes localised in the subendocardial layer of the myocardium of the left ventricle. The identified changes in the subendocardial layer of the myocardium caused focal thickening, fragmentation, and increased thickness of the endocardial folds, their undulation, which can be considered as manifestations of morpho-functional disorders of the endocardium.

Fetal heart segmentation in a virtual reality environment.

This study presents the initial results of a pilot project using the Elucis Virtual Reality (VR) platform for fetal heart segmentation. Twelve fetal heart cases, ranging in gestational age from 24 to 30 weeks, including various cardiac conditions, were reconstructed using 3D models facilitated by the Elucis platform's integration of automated algorithms and manual adjustments. The models, which were evaluated by four experts in virtual and 3D printed formats, were of high quality and offered improved visuospatial visualization and detailed anatomical insights. This research highlights the potential of VR technology to improve prenatal diagnosis and planning for complex cardiac conditions, suggesting significant implications for continuing medical education and clinical practice in fetal cardiology.

MSMHSA-DeepLab V3+: An Effective Multi-Scale, Multi-Head Self-Attention Network for Dual-Modality Cardiac Medical Image Segmentation.

The automatic segmentation of cardiac computed tomography (CT) and magnetic resonance imaging (MRI) plays a pivotal role in the prevention and treatment of cardiovascular diseases. In this study, we propose an efficient network based on the multi-scale, multi-head self-attention (MSMHSA) mechanism. The incorporation of this mechanism enables us to achieve larger receptive fields, facilitating the accurate segmentation of whole heart structures in both CT and MRI images. Within this network, features extracted from the shallow feature extraction network undergo a MHSA mechanism that closely aligns with human vision, resulting in the extraction of contextual semantic information more comprehensively and accurately. To improve the precision of cardiac substructure segmentation across varying sizes, our proposed method introduces three MHSA networks at distinct scales. This approach allows for fine-tuning the accuracy of micro-object segmentation by adapting the size of the segmented images. The efficacy of our method is rigorously validated on the Multi-Modality Whole Heart Segmentation (MM-WHS) Challenge 2017 dataset, demonstrating competitive results and the accurate segmentation of seven cardiac substructures in both cardiac CT and MRI images. Through comparative experiments with advanced transformer-based models, our study provides compelling evidence that despite the remarkable achievements of transformer-based models, the fusion of CNN models and self-attention remains a simple yet highly effective approach for dual-modality whole heart segmentation.

Segmentation and Volumetric Analysis of Heart from Cardiac CT Images.

Cardiac CT is a valuable diagnostic tool in evaluating cardiovascular diseases. Accurate segmentation of the heart and its structures from cardiac CT and MRI images is essential for diagnosing functional abnormalities, treatment plans and cardiovascular diseases management. Accurate segmentation and quantitative assessments are still a challenge. Manual delineation of the heart from the scan images is labour-intensive, time-consuming, and error prone as it depends on the radiologist's experience. Thus, automated techniques are highly desirable as they can significantly improve the efficiency and accuracy of image analysis. This work addresses the above problems. A new, image-driven, fast, and fully automatic segmentation method was developed to segment the heart from CT images using a processing pipeline of adaptive median filter, multi-level thresholding, active contours, mathematical morphology, and the knowledge of human anatomy to delineate the regions of interest. The algorithm proposed is simple to implement and validate and requires no human intervention. The method is tested on the 'Image CHD' DICOM images (multi-centre, clinically approved single-phase de-identified images), and the results obtained were validated against the ground truths provided with the dataset. The results show an average Dice score, Jaccard score, and Hausdorff distance of 0.866, 0.776, and 33.29mm, respectively, for the segmentation of the heart's chambers, aorta, and blood vessels. The results and the ground truths were compared using Bland-Altmon plots. The heart was correctly segmented from the CT images using the proposed method. Further this segmentation technique can be used to develop AI based solutions for segmentation.

Computer based method for identification of fibrotic scars from electrograms and local activation times on the epi- and endocardial surfaces of the ventricles.

Cardiac fibrosis stands as one of the most critical conditions leading to lethal cardiac arrhythmias. Identifying the precise location of cardiac fibrosis is crucial for planning clinical interventions in patients with various forms of ventricular and atrial arrhythmias. As fibrosis impedes and alters the path of electrical waves, detecting fibrosis in the heart can be achieved through analyzing electrical signals recorded from its surface. In current clinical practices, it has become feasible to record electrical activity from both the endocardial and epicardial surfaces of the heart. This paper presents a computational method for reconstructing 3D fibrosis using unipolar electrograms obtained from both surfaces of the ventricles. The proposed method calculates the percentage of fibrosis in various ventricular segments by analyzing the local activation times and peak-to-peak amplitudes of the electrograms. Initially, the method was tested using simulated data representing idealized fibrosis in a heart segment; subsequently, it was validated in the left ventricle with fibrosis obtained from a patient with nonischemic cardiomyopathy. The method successfully determined the location and extent of fibrosis in 204 segments of the left ventricle model with an average error of 0.0±4.3% (N = 204). Moreover, the method effectively detected fibrotic scars in the mid-myocardial region, a region known to present challenges in accurate detection using electrogram amplitude as the primary criterion.

Multi-residual 2D network integrating spatial correlation for whole heart segmentation

Whole heart segmentation (WHS) has significant clinical value for cardiac anatomy, modeling, and analysis of cardiac function. This study aims to address the WHS accuracy on cardiac CT images, as well as the fast inference speed and low graphics processing unit (GPU) memory consumption required by practical clinical applications. Thus, we propose a multi-residual two-dimensional (2D) network integrating spatial correlation for WHS. The network performs slice-by-slice segmentation on three-dimensional cardiac CT images in a 2D encoder–decoder manner. In the network, a convolutional long short-term memory skip connection module is designed to perform spatial correlation feature extraction on the feature maps at different resolutions extracted by the sub-modules of the pre-trained ResNet-based encoder. Moreover, a decoder based on the multi-residual module is designed to analyze the extracted features from the perspectives of multi-scale and channel attention, thereby accurately delineating the various substructures of the heart. The proposed method is verified on a dataset of the multi-modality WHS challenge, an in-house WHS dataset, and a dataset of the abdominal organ segmentation challenge. The dice, Jaccard, average symmetric surface distance, Hausdorff distance, inference time, and maximum GPU memory of the WHS are 0.914, 0.843, 1.066 mm, 15.778 mm, 9.535 s, and 1905 MB, respectively. The proposed network has high accuracy, fast inference speed, minimal GPU memory consumption, strong robustness, and good generalization. It can be deployed to clinical practical applications for WHS and can be effectively extended and applied to other multi-organ segmentation fields. The source code is publicly available at https://github.com/nancy1984yan/MultiResNet-SC.

Unsupervised Domain Adaptation for Medical Image Segmentation Using Transformer With Meta Attention.

Image segmentation is essential to medical image analysis as it provides the labeled regions of interest for the subsequent diagnosis and treatment. However, fully-supervised segmentation methods require high-quality annotations produced by experts, which is laborious and expensive. In addition, when performing segmentation on another unlabeled image modality, the segmentation performance will be adversely affected due to the domain shift. Unsupervised domain adaptation (UDA) is an effective way to tackle these problems, but the performance of the existing methods is still desired to improve. Also, despite the effectiveness of recent Transformer-based methods in medical image segmentation, the adaptability of Transformers is rarely investigated. In this paper, we present a novel UDA framework using a Transformer for building a cross-modality segmentation method with the advantages of learning long-range dependencies and transferring attentive information. To fully utilize the attention learned by the Transformer in UDA, we propose Meta Attention (MA) and use it to perform a fully attention-based alignment scheme, which can learn the hierarchical consistencies of attention and transfer more discriminative information between two modalities. We have conducted extensive experiments on cross-modality segmentation using three datasets, including a whole heart segmentation dataset (MMWHS), an abdominal organ segmentation dataset, and a brain tumor segmentation dataset. The promising results show that our method can significantly improve performance compared with the state-of-the-art UDA methods.

Self-supervised spatial–temporal transformer fusion based federated framework for 4D cardiovascular image segmentation

Availability of high-quality large annotated data is indeed a significant challenge in healthcare. In addition, privacy concerns and data-sharing restrictions often hinder access to large and diverse medical image datasets. To reduce the requirement for annotated training data, self-supervised pre-training strategies on nonannotated data have been extensively used, whereas collaborative algorithm training without the need to exchange the underlying data. In this paper, we introduce a novel federated learning-based self-supervised spatial–temporal transformer’s fusion (SSFL) for cardiovascular image segmentation. The integration of spatial–temporal swin transformer is used to extract the features from 3D SAX multiple phases (full cycle of cardiac heart). An efficient self-supervised contrastive framework consisting of a spatial–temporal transformer network with 25 encoders is used to model the temporal features. The spatial and temporal features are fused and forwarded to the decoder for cardiac heart segmentation using cine MRI images. To further improve segmentation, we use an attention-based unpaired GAN model to map or transfer the style from ACDC to M&Ms and use synthetically generated volumes in the proposed self-supervised approach. Experiments with three different cardiovascular image segmentation tasks, such as segmentation of the right ventricle, left ventricle, and myocardium, showed significant improvement compared to the state-of-the-art segmentation framework.

Automatic cardiothoracic ratio calculation based on lung fields abstracted from chest X-ray images without heart segmentation.

The cardiothoracic ratio (CTR) based on postero-anterior chest X-rays (P-A CXR) images is one of the most commonly used cardiac measurement methods and an indicator for initially evaluating cardiac diseases. However, the hearts are not readily observable on P-A CXR images compared to the lung fields. Therefore, radiologists often manually determine the CTR's right and left heart border points of the adjacent left and right lung fields to the heart based on P-A CXR images. Meanwhile, manual CTR measurement based on the P-A CXR image requires experienced radiologists and is time-consuming and laborious. Based on the above, this article proposes a novel, fully automatic CTR calculation method based on lung fields abstracted from the P-A CXR images using convolutional neural networks (CNNs), overcoming the limitations to heart segmentation and avoiding errors in heart segmentation. First, the lung field mask images are abstracted from the P-A CXR images based on the pre-trained CNNs. Second, a novel localization method of the heart's right and left border points is proposed based on the two-dimensional projection morphology of the lung field mask images using graphics. The results show that the mean distance errors at the x-axis direction of the CTR's four key points in the test sets T1 (21 × 512 × 512 static P-A CXR images) and T2 (13 × 512 × 512 dynamic P-A CXR images) based on various pre-trained CNNs are 4.1161 and 3.2116 pixels, respectively. In addition, the mean CTR errors on the test sets T1 and T2 based on four proposed models are 0.0208 and 0.0180, respectively. Our proposed model achieves the equivalent performance of CTR calculation as the previous CardioNet model, overcomes heart segmentation, and takes less time. Therefore, our proposed method is practical and feasible and may become an effective tool for initially evaluating cardiac diseases.

An open source auto-segmentation algorithm for delineating heart and substructures – Development and validation within a multicenter lung cancer cohort

Background and purposeIrradiation of the heart in thoracic cancers raises toxicity concerns. For accurate dose estimation, automated heart and substructure segmentation is potentially useful. In this study, a hybrid automatic segmentation is developed. The accuracy of delineation and dose predictions were evaluated, testing the method's potential within heart toxicity studies. Materials and MethodsThe hybrid segmentation method delineated the heart, four chambers, three large vessels, and the coronary arteries. The method consisted of a nnU-net heart segmentation and partly atlas- and model-based segmentation of the substructures. The nnU-net training and atlas segmentation was based on lung cancer patients and was validated against a national consensus dataset of 12 patients with breast cancer. The accuracy of dose predictions between manual and auto-segmented heart and substructures was evaluated by transferring the dose distribution of 240 previously treated lung cancer patients to the consensus data set. ResultsThe hybrid auto-segmentation method performed well with a heart dice similarity coefficient (DSC) of 0.95, with no statistically significant difference between the automatic and manual delineations. The DSC for the chambers varied from 0.78-0.86 for the automatic segmentation and was comparable with the inter-observer variability. Most importantly, the automatic segmentation was as precise as the clinical experts in predicting the dose distribution to the heart and all substructures. ConclusionThe hybrid segmentation method performed well in delineating the heart and substructures. The prediction of dose by the automatic segmentation was aligned with the manual delineations, enabling measurement of heart and substructure dose in large cohorts. The delineation algorithm will be available for download.

A Drosophila heart optical coherence microscopy dataset for automatic video segmentation

The heart of the fruit fly, Drosophila melanogaster, is a particularly suitable model for cardiac studies. Optical coherence microscopy (OCM) captures in vivo cross-sectional videos of the beating Drosophila heart for cardiac function quantification. To analyze those large-size multi-frame OCM recordings, human labelling has been employed, leading to low efficiency and poor reproducibility. Here, we introduce a robust and accurate automated Drosophila heart segmentation algorithm, called FlyNet 2.0+, which utilizes a long short-term memory (LSTM) convolutional neural network to leverage time series information in the videos, ensuring consistent, high-quality segmentation. We present a dataset of 213 Drosophila heart videos, equivalent to 604,000 cross-sectional images, containing all developmental stages and a wide range of beating patterns, including faster and slower than normal beating, arrhythmic beating, and periods of heart stop to capture these heart dynamics. Each video contains a corresponding ground truth mask. We expect this unique large dataset of the beating Drosophila heart in vivo will enable new deep learning approaches to efficiently characterize heart function to advance cardiac research.

GLRP: Global and local contrastive learning based on relative position for medical image segmentation on cardiac MRI

AbstractContrastive learning, as an unsupervised technique, is widely employed in image segmentation to enhance segmentation performance even when working with small labeled datasets. However, generating positive and negative data pairs for medical image segmentation poses a challenge due to the presence of similar tissues and organs across different slices in datasets. To tackle this issue, we propose a novel contrastive learning strategy that leverages the relative position differences between image slices. Additionally, we combine global and local features to address this problem effectively. In order to enhance segmentation accuracy and reduce isolated mis‐segmented regions, we employ a two‐dimensional fully connected conditional random field for iterative optimization of the segmentation results. With only 10 labeled samples, our proposed method is able to achieve average dice scores of 0.876 and 0.899 on the public and private dataset heart segmentation tasks, surpassing the PCL method's 0.801 and 0.852. Experimental results on both public and private MRI datasets demonstrate that our proposed method yields significant improvements in medical segmentation tasks with limited annotated samples, outperforming existing semi‐supervised and self‐supervised techniques.