Applications of artificial intelligence-powered prenatal diagnosis for congenital heart disease.

Artificial intelligence (AI) has made significant progress in the medical field in the last decade. The AI-powered analysis methods of medical images and clinical records can now match the abilities of clinical physicians. Due to the challenges posed by the unique group of fetuses and the dynamic organ of the heart, research into the application of AI in the prenatal diagnosis of congenital heart disease (CHD) is particularly active. In this review, we discuss the clinical questions and research methods involved in using AI to address prenatal diagnosis of CHD, including imaging, genetic diagnosis, and risk prediction. Representative examples are provided for each method discussed. Finally, we discuss the current limitations of AI in prenatal diagnosis of CHD, namely Volatility, Insufficiency and Independence (VII), and propose possible solutions.

Atomic Valence Reversal-Induced Polarization Resonance Spurs Highly Efficient Electromagnetic Wave Absorption in α-Fe2O3@Carbon Microtubes.

Variegation and complexity of polarization relaxation loss in many heterostructured materials provide available mechanisms to seek a strong electromagnetic wave (EMW) absorption performance. Here we construct a unique heterostructured compound that bonds α-Fe2O3 nanosheets of the (110) plane on carbon microtubes (CMTs). Through effective alignment between the Fermi energy level of CMTs and the conduction band position of α-Fe2O3 nanosheets at the interface, we attain substantial polarization relaxation loss via novel atomic valence reversal between Fe(III) ↔ Fe(III-) induced with periodic electron injection from conductive CMTs under EMW irradiation to give α-Fe2O3 nanosheets. Such heterostructured materials possess currently reported minimum reflection loss of -84.01 dB centered at 10.99 GHz at a thickness of 3.19 mm and an effective absorption bandwidth (reflection loss ≤ -10 dB) of 7.17 GHz (10.83-18 GHz) at 2.65 mm. This work provides an effective strategy for designing strong EMW absorbers by combining highly efficient electron injection and atomic valence reversal.

Improved Multi-Head Self-Attention Classification Network for Multi-View Fetal Echocardiography Recognition.

The accurate acquisition of multiview fetal cardiac ultrasound images is very important for the diagnosis of fetal congenital heart disease (FCHD). However, these manual clinical procedures have drawbacks, e.g., varying technical capabilities and inefficiency. Therefore, exploring automatic recognition method for multiview images of fetal heart ultrasound scans is highly desirable to improve prenatal diagnosis efficiency and accuracy. In this work, we propose an improved multi-head self-attention mechanism called IMSA combined with residual networks to stably solve the problem of multiview identification and anatomical structure localization. In details, IMSA can capture short- and long-range dependencies from different subspaces and merge them to extract more precise features, thus making use of the correlation between fetal heart structures to make view recognition more focused on anatomical structures rather than disturbing regions, such as artifacts and speckle noises. We validate our proposed method on fetal cardiac ultrasound imaging datasets from a single center and 38 multicenter studies and the results outperform other state-of-the-art networks by 3%-15% of F1 scores in fetal heart six standard view recognition.Clinical Relevance- This technology has great potential in assisting cardiologists to complete the automatic acquisition of multi-section fetal echocardiography images.

Rethinking Disentanglement in Unsupervised Domain Adaptation for Medical Image Segmentation.

Domain adaptation has become an important topic because the trained neural networks from the source domain generally perform poorly in the target domain due to domain shifts, especially for medical image analysis. Previous DA methods mainly focus on disentangling domain features. However, it is based on feature independence, which often can not be guaranteed in reality. In this work, we present a new DA approach called Dimension-based Disentangled Dilated Domain Adaptation (D4A) to disentangle the storage locations between the features to tackle the problem of domain shift for medical image segmentation tasks without the annotations of the target domain. We use Adaptive Instance Normalization (AdaIN) to encourage the content information to be stored in the spatial dimension, and the style information to be stored in the channel dimension. In addition, we apply dilated convolution to preserve anatomical information avoiding the loss of information due to downsampling. We validate the proposed method for cross-modality medical image segmentation tasks on two public datasets, and the comparison experiments and ablation studies demonstrate the effectiveness of our method, which outperforms the state-of-the-art methods.

Heart failure classifications via non-invasive pressure volume loops from echocardiography.

BACKGROUND: Left ventricular pressure-volume (LV-PV) loops provide comprehensive characterization of cardiovascular system in both health and disease, which are the essential element of the hemodynamic evaluation of heart failure (HF). This study attempts to achieve more detailed HF classifications by non-invasive LV-PV loops from echocardiography and analyzes contribution of parameters to HF classifications. METHODS: Firstly, non-invasive PV loops are established by time-varying elastance model where LV volume curves were extracted from apical-four-chambers view of echocardiographic videos. Then, 16 parameters related to cardiac structure and functions are automatically acquired from PV loops. Next, we applied six machine learning (ML) methods to divide four categories. On this premise, we choose the best performing classifier among machine learning approaches for feature ranking. Finally, we compare the contributions of different parameters to HF classifications. RESULTS: By the experimental, the PV loops were successfully acquired in 1076 cases. When single left ventricular ejection fraction (LVEF) is used for HF classifications, the accuracy of the model is 91.67%. When added parameters extracted from ML-derived LV-PV loops, the classification accuracy is 96.57%, which improved by 5.1%. Especially, our parameters have a great improvement in the classification of non-HF controls and heart failure with preserved ejection fraction (HFpEF). CONCLUSIONS: We successfully presented the classification of HF by machine derived non-invasive LV-PV loops, which has the potential to improve the diagnosis and management of heart failure in clinic. Moreover, ventriculo-arterial (VA) coupling and ventricular efficiency were demonstrated important factors for ML-based HF classification model besides LVEF.

FVP: Fourier Visual Prompting for Source-Free Unsupervised Domain Adaptation of Medical Image Segmentation.

Medical image segmentation methods normally perform poorly when there is a domain shift between training and testing data. Unsupervised Domain Adaptation (UDA) addresses the domain shift problem by training the model using both labeled data from the source domain and unlabeled data from the target domain. Source-Free UDA (SFUDA) was recently proposed for UDA without requiring the source data during the adaptation, due to data privacy or data transmission issues, which normally adapts the pre-trained deep model in the testing stage. However, in real clinical scenarios of medical image segmentation, the trained model is normally frozen in the testing stage. In this paper, we propose Fourier Visual Prompting (FVP) for SFUDA of medical image segmentation. Inspired by prompting learning in natural language processing, FVP steers the frozen pre-trained model to perform well in the target domain by adding a visual prompt to the input target data. In FVP, the visual prompt is parameterized using only a small amount of low-frequency learnable parameters in the input frequency space, and is learned by minimizing the segmentation loss between the predicted segmentation of the prompted target image and reliable pseudo segmentation label of the target image under the frozen model. To our knowledge, FVP is the first work to apply visual prompts to SFUDA for medical image segmentation. The proposed FVP is validated using three public datasets, and experiments demonstrate that FVP yields better segmentation results, compared with various existing methods.

Improving the Quality of Fetal Heart Ultrasound Imaging With Multihead Enhanced Self-Attention and Contrastive Learning.

Fetal congenital heart disease (FCHD) is a common, serious birth defect affecting ∼1% of newborns annually. Fetal echocardiography is the most effective and important technique for prenatal FCHD diagnosis. The prerequisites for accurate ultrasound FCHD diagnosis are accurate view recognition and high-quality diagnostic view extraction. However, these manual clinical procedures have drawbacks such as, varying technical capabilities and inefficiency. Therefore, the automatic identification of high-quality multiview fetal heart scan images is highly desirable to improve prenatal diagnosis efficiency and accuracy of FCHD. Here, we present a framework for multiview fetal heart ultrasound image recognition and quality assessment that comprises two parts: a multiview classification and localization network (MCLN) and an improved contrastive learning network (ICLN). In the MCLN, a multihead enhanced self-attention mechanism is applied to construct the classification network and identify six accurate and interpretable views of the fetal heart. In the ICLN, anatomical structure standardization and image clarity are considered. With contrastive learning, the absolute loss, feature relative loss and predicted value relative loss are combined to achieve favorable quality assessment results. Experiments show that the MCLN outperforms other state-of-the-art networks by 1.52-13.61% when determining the F1 score in six standard view recognition tasks, and the ICLN is comparable to the performance of expert cardiologists in the quality assessment of fetal heart ultrasound images, reaching 97% on a test set within 2 points for the four-chamber view task. Thus, our architecture offers great potential in helping cardiologists improve quality control for fetal echocardiographic images in clinical practice.

SCCNet: Self-correction boundary preservation with a dynamic class prior filter for high-variability ultrasound image segmentation.

The highly ambiguous nature of boundaries and similar objects is difficult to address in some ultrasound image segmentation tasks, such as neck muscle segmentation, leading to unsatisfactory performance. Thus, this paper proposes a two-stage network called SCCNet (self-correction context network) using a self-correction boundary preservation module and class-context filter to alleviate these problems. The proposed self-correction boundary preservation module uses a dynamic key boundary point (KBP) map to increase the capability of iteratively discriminating ambiguous boundary points segments, and the predicted segmentation map from one stage is used to obtain a dynamic class prior filter to improve the segmentation performance at Stage 2. Finally, three datasets, Neck Muscle, CAMUS and Thyroid, are used to demonstrate that our proposed SCCNet outperforms other state-of-the art methods, such as BPBnet, DSNnet, and RAGCnet. Our proposed network shows at least a 1.2-3.7% improvement on the three datasets, Neck Muscle, Thyroid, and CAMUS. The source code is available at https://github.com/lijixing0425/SCCNet.

MSGAN: Multi-Stage Generative Adversarial Networks for Cross-Modality Domain Adaptation.

Domain adaptation has become an important topic because the trained neural networks from the source domain generally perform poorly in the target domain due to domain shifts, especially for cross-modality medical images. In this work, we present a new unsupervised domain adaptation approach called Multi-Stage GAN (MSGAN) to tackle the problem of domain shift for CT and MRI segmentation tasks. We adopt the multi-stage strategy in parallel to avoid information loss and transfer rough styles on low-resolution feature maps to the detailed textures on high-resolution feature maps. In detail, the style layers map the learnt style codes from the Gaussian noise to the input features in order to synthesize images with different styles. We validate the proposed method for cross-modality medical image segmentation tasks on two public datasets, and the results demonstrate the effectiveness of our method. Clinical relevance- This technique paves the way to translate cross-modality images (MRI and CT) and it can also mitigate the performance degradation when applying deep neural networks in a cross-domain scenario.

Diagnosis of fetal total anomalous pulmonary venous connection based on the post-left atrium space ratio using artificial intelligence.

OBJECTIVE: To explore whether the post-left atrium space (PLAS) ratio would be useful for prenatal diagnosis of total anomalous pulmonary venous connection (TAPVC) using echocardiography and artificial intelligence. METHODS: We retrospectively included 642 frames of four-chamber views from 319 fetuses (32 with TAPVC and 287 without TAPVC) in end-systolic and end-diastolic periods with multiple apex directions. The average gestational age was 25.6 ± 2.7 weeks. No other cardiac or extracardiac malformations were observed. The dataset was divided into a training set (n = 540; 48 with TAPVC and 492 without TAPVC) and test set (n = 102; 20 with TAPVC and 82 without TAPVC). The PLAS ratio was defined as the ratio of the epicardium-descending aortic distance to the center of the heart-descending aortic distance. Supervised learning was used in DeepLabv3+, FastFCN, PSPNet, and DenseASPP segmentation models. The area under the curve (AUC) was used on the test set. RESULTS: Expert annotations showed that this ratio was not related to the period or apex direction. It was higher in the TAPVC group than in the control group detected by the expert and the four models. The AUC of expert annotations, DeepLabv3+, FastFCN, PSPNet, and DenseASPP were 0.977, 0.941, 0.925, 0.856, and 0.887, respectively. CONCLUSION: Segmentation models achieve good diagnostic accuracy for TAPVC based on the PLAS ratio.

Strong Out-of-Plane Vibrations and Ultrasensitive Detection of Dopamine-like Neurotransmitters.

The detection of monoamine neurotransmitters has become a vital research subject due to their high correlations with nervous system diseases, but insufficient detection precisions have obstructed diagnosis of some related diseases. Here, we focus on four monoamine neurotransmitters, dopamine, norepinephrine, epinephrine, and serotonin, to conduct their rapid and ultrasensitive detection. We find that the low-frequency (<200 cm-1) Raman vibrations of these molecules show some sharp peaks, and their intensities are significantly stronger than those of the high-frequency side. Theoretical calculations identify these peaks to be from strong out-of-plane vibrations of the C-C single bonds at the joint point of the ring-like molecule and its side chain. Using our surface enhanced low-frequency Raman scattering substrates, we show that the detection limit of dopamine as an example can reach 10 nM in artificial cerebrospinal fluid. This work provides a useful way for ultrasensitive and rapid detection of some neurotransmitters.

Fetal Heart and Descending Aorta Detection in Four-Chamber View of Fetal Echocardiography.

Automatic analysis of fetal heart and related components in fetal echocardiography can help cardiologists to reach a diagnosis for Congenital Heart Disease (CHD). Previous studies mainly focused on cardiac chamber segmentation, while few researches deal with the cardiac component detection. In this paper, we tackle the task of simultaneous detection of the fetal heart and descending aorta in four-chamber view of fetal echocardiography, which is useful to analyze some kinds of CHD, such as left/right atrial isomerism, dextroversion of heart, etc. Several CNN-based object detection methods with different backbones are thoroughly evaluated, and finally, the Hybrid Task Cascade method with HRNet is selected as the detection method. Experiments on a fetal echocardiography dataset show that the method can achieve superior performance according to common-used evaluation metrics.Clinical relevance-This can be used to help the cardiologists to estimate the position of the fetal heart and the descending aorta, which is also useful to estimate the direction of the cardiac axis and apex and analyze some kinds of CHD, such as left/right atrial isomerism, dextroversion of heart, etc.

Simultaneous Segmentation of Four Cardiac Chambers in Fetal Echocardiography.

Accurate segmentation of cardiac chambers is helpful for the diagnosis of Congenital Heart Disease (CHD) in fetal echocardiography. Previous studies mainly focused on single cardiac chamber segmentation, which cannot provide sufficient information for the cardiologists. In this paper, we present an instance segmentation approach capable of segmenting four cardiac chambers accurately and simultaneously. A novel object proposal recovery strategy is further deployed to retrieve possible missing objects. To alleviate the shortage of medical data and further improve the segmentation performance, we utilize a rotation and distortion method for data augmentation. Experiments on a fetal echocardiography dataset of 319 fetuses demonstrate that the proposed approach can achieve superior performance according to common-used evaluation metrics.Clinical relevance-This can be used to help the cardiologists to better analyze the structure and function of the fetal heart.

Strong histamine torsion Raman spectrum enables direct, rapid, and ultrasensitive detection of allergic diseases.

Allergic diseases are closely related to degranulation and release of histamine and difficult to diagnose because non-allergic diseases also exhibit the same clinical symptoms as allergy. Here, we report direct, rapid, and ultrasensitive detection of histamine using low-frequency molecular torsion Raman spectroscopy. We show that the low-frequency (<200 cm-1) Raman spectral intensities are stronger by one order of magnitude than those of the high-frequency Raman ones. Density functional theory calculation and nuclear magnetic resonance spectroscopy identify the strong spectral feature to be from torsions of carbon-carbon single bonds, which produce large variations of the polarizability densities in the imidazole ring and ethyl amino side chain. Using an omniphobic substrate and surface plasmonic effect of Au@SiO2 nanoparticles, the detection limit (signal-noise ratio >3) of histamine reaches 10-8 g/L in water and 10-6 g/L in serum. This scheme thus opens new lines of inquiry regarding the clinical diagnosis of allergic diseases.

A category attention instance segmentation network for four cardiac chambers segmentation in fetal echocardiography.

Fetal echocardiography is an essential and comprehensive examination technique for the detection of fetal heart anomalies. Accurate cardiac chambers segmentation can assist cardiologists to analyze cardiac morphology and facilitate heart disease diagnosis. Previous research mainly focused on the segmentation of single cardiac chambers, such as left ventricle (LV) segmentation or left atrium (LA) segmentation. We propose a generic framework based on instance segmentation to segment the four cardiac chambers accurately and simultaneously. The proposed Category Attention Instance Segmentation Network (CA-ISNet) has three branches: a category branch for predicting the semantic category, a mask branch for segmenting the cardiac chambers, and a category attention branch for learning category information of instances. The category attention branch is used to correct instance misclassification of the category branch. In our collected dataset, which contains echocardiography images with four-chamber views of 319 fetuses, experimental results show our method can achieve superior segmentation performance against state-of-the-art methods. Specifically, using fivefold cross-validation, our model achieves Dice coefficients of 0.7956, 0.7619, 0.8199, 0.7470 for the four cardiac chambers, and with an average precision of 45.64%.

Fabrication of large size individual octahedral tungsten oxide hydrate and Au NPs as SERS platforms for sensitive detection of cytochrome C.

Surface-enhanced Raman scattering (SERS) has attracted much attention with its powerful trace detection and analysis capabilities, especially biological and environmental molecules. However, building a protein SERS detection platform based on semiconductor devices is a huge challenge. Herein, through the synergy of NH3 and nickel foam, a large-sized semiconductor tungsten oxide hydrate platform (WOHP) was synthesized. The crystal plane of a single WOHP particle is larger than the excitation spot. As a SERS substrate, WOHP can make full use of the excitation light without destroying the structure during the protein molecules detection process. Through the synergy of WOHP and Au NPs, the enhancement factor is 1.5 × 104. Raman peaks of WOHP can be used as references for the detection of typical protein cytochrome C (Cyt C). As the Cyt C concentration decreases, the ICyt C/IWOHP ratio decreases, and the signal can still be obtained when the concentration is as low as 5 × 10-9 mol L-1. More importantly, the method does not affect the catalytic activity of Cyt C and can be applied to the detection of Cyt C concentration in serum.

Brain Age Estimation From MRI Using Cascade Networks With Ranking Loss.

Chronological age of healthy people is able to be predicted accurately using deep neural networks from neuroimaging data, and the predicted brain age could serve as a biomarker for detecting aging-related diseases. In this paper, a novel 3D convolutional network, called two-stage-age-network (TSAN), is proposed to estimate brain age from T1-weighted MRI data. Compared with existing methods, TSAN has the following improvements. First, TSAN uses a two-stage cascade network architecture, where the first-stage network estimates a rough brain age, then the second-stage network estimates the brain age more accurately from the discretized brain age by the first-stage network. Second, to our knowledge, TSAN is the first work to apply novel ranking losses in brain age estimation, together with the traditional mean square error (MSE) loss. Third, densely connected paths are used to combine feature maps with different scales. The experiments with 6586 MRIs showed that TSAN could provide accurate brain age estimation, yielding mean absolute error (MAE) of 2.428 and Pearson's correlation coefficient (PCC) of 0.985, between the estimated and chronological ages. Furthermore, using the brain age gap between brain age and chronological age as a biomarker, Alzheimer's disease (AD) and Mild Cognitive Impairment (MCI) can be distinguished from healthy control (HC) subjects by support vector machine (SVM). Classification AUC in AD/HC and MCI/HC was 0.904 and 0.823, respectively. It showed that brain age gap is an effective biomarker associated with risk of dementia, and has potential for early-stage dementia risk screening. The codes and trained models have been released on GitHub: https://github.com/Milan-BUAA/TSAN-brain-age-estimation.

Neural Network-Based Retinal Nerve Fiber Layer Profile Compensation for Glaucoma Diagnosis in Myopia: Model Development and Validation.

BACKGROUND: Due to the axial elongation-associated changes in the optic nerve and retina in high myopia, traditional methods like optic disc evaluation and visual field are not able to correctly differentiate glaucomatous lesions. It has been clinically challenging to detect glaucoma in highly myopic eyes. OBJECTIVE: This study aimed to develop a neural network to adjust for the dependence of the peripapillary retinal nerve fiber layer (RNFL) thickness (RNFLT) profile on age, gender, and ocular biometric parameters and to evaluate the network's performance for glaucoma diagnosis, especially in high myopia. METHODS: RNFLT with 768 points on the circumferential 3.4-mm scan was measured using spectral-domain optical coherence tomography. A fully connected network and a radial basis function network were trained for vertical (scaling) and horizontal (shift) transformation of the RNFLT profile with adjustment for age, axial length (AL), disc-fovea angle, and distance in a test group of 2223 nonglaucomatous eyes. The performance of RNFLT compensation was evaluated in an independent group of 254 glaucoma patients and 254 nonglaucomatous participants. RESULTS: By applying the RNFL compensation algorithm, the area under the receiver operating characteristic curve for detecting glaucoma increased from 0.70 to 0.84, from 0.75 to 0.89, from 0.77 to 0.89, and from 0.78 to 0.87 for eyes in the highest 10% percentile subgroup of the AL distribution (mean 26.0, SD 0.9 mm), highest 20% percentile subgroup of the AL distribution (mean 25.3, SD 1.0 mm), highest 30% percentile subgroup of the AL distribution (mean 24.9, SD 1.0 mm), and any AL (mean 23.5, SD 1.2 mm), respectively, in comparison with unadjusted RNFLT. The difference between uncompensated and compensated RNFLT values increased with longer axial length, with enlargement of 19.8%, 18.9%, 16.2%, and 11.3% in the highest 10% percentile subgroup, highest 20% percentile subgroup, highest 30% percentile subgroup, and all eyes, respectively. CONCLUSIONS: In a population-based study sample, an algorithm-based adjustment for age, gender, and ocular biometric parameters improved the diagnostic precision of the RNFLT profile for glaucoma detection particularly in myopic and highly myopic eyes.

The Clouclip, a wearable device for measuring near-work and outdoor time: validation and comparison of objective measures with questionnaire estimates.

PURPOSE: To validate a novel wearable device that can measure both viewing distance and light exposure, Clouclip, and compare questionnaire estimates regarding near-work and outdoor time with the objective measures obtained using Clouclip. METHODS: Fifteen Clouclips were selected to measure different distances and levels of illuminance. With each Clouclip, five measurements at different distances and light intensities were measured and recorded. Eighty participants wore Clouclips for a week and completed an activity questionnaire afterwards. RESULTS: The intra- and inter-Clouclip coefficients were 1.00 and 0.99 for measuring distance and 1.00 and 1.00 for illuminance, respectively. Within the measurement limit, the maximum relative error was 2.07% for distance and 2.23% for illuminance. Assuming that <30 cm was the typical distance for near-work activities and >1000 Lux was the typical cut-off for outdoor environments, the questionnaire showed a trend of overestimation for both. The greatest overestimation of near-work occurred during the school period [Questionnaire: 4.73 hr (4.73, 5.07) versus Clouclip: 2.16 hr (1.74, 2.78); p < 0.01]. The greatest overestimation of outdoor activity also occurred during the school period [Questionnaire: 1.60 hr (1.33, 1.85) versus Clouclip: 1.21 hr (0.96, 1.50); p < 0.01]. Based on Clouclip, the total time spent outdoors was estimated to be 1.55 hr on school days, of which 0.34 hr occurred after school. For weekend days, however, the duration was only 0.17 hr. CONCLUSIONS: Clouclip had excellent precision and accuracy. Although the agreement between the questionnaire and Clouclip was relatively poor, they were able to complement each other to provide a more logical and feasible assessment of exposure to near-work and outdoor activity. Indoor-oriented lifestyles were found to predominate in Chinese children.

Trail-Traced Threshold Test (T4) With a Weighted Binomial Distribution for a Psychophysical Test.

Clinical visual field testing is performed with commercial perimetric devices and employs psychophysical techniques to obtain thresholds of the differential light sensitivity (DLS) at multiple retinal locations. Current thresholding algorithms are relatively inefficient and tough to get satisfied test accuracy, stability concurrently. Thus, we propose a novel Bayesian perimetric threshold method called the Trail-Traced Threshold Test (T4), which can better address the dependence of the initial threshold estimation and achieve significant improvement in the test accuracy and variability while also decreasing the number of presentations compared with Zippy Estimation by Sequential Testing (ZEST) and FT. This study compares T4 with ZEST and FT regarding presentation number, mean absolute difference (MAD between the real Visual field result and the simulate result), and measurement variability. T4 uses the complete response sequence with the spatially weighted neighbor responses to achieve better accuracy and precision than ZEST, FT, SWeLZ, and with significantly fewer stimulus presentations. T4 is also more robust to inaccurate initial threshold estimation than other methods, which is an advantage in subjective methods, such as in clinical perimetry. This method also has the potential for using in other psychophysical tests.