Objective: Cardiovascular diseases (CVDs) are a leading global health threat. The automatic classification of phonocardiogram (PCG) signals is crucial for their early diagnosis, yet existing models are often limited by analyzing features from only a single domain (time or frequency), failing to fuse complementary information. This study aims to develop a model that overcomes this limitation by effectively integrating both time-domain and frequency-domain features to improve classification accuracy and robustness. Methods: We propose a novel end-to-end dual-branch deep learning model. The time-domain branch utilizes a 1D Convolutional Neural Network (CNN) with Transformer blocks to capture instantaneous dynamics and long-range dependencies. The frequency-domain branch uses a ResNet to extract robust spectral patterns from Mel-spectrograms. A key innovation is our bidirectional cross-attention fusion module, which facilitates deep interaction and mutual enhancement between the two feature modalities. Furthermore, we employ a transfer learning strategy to ensure robust performance on smaller or more challenging datasets. Results: Comprehensive evaluations on multiple public datasets demonstrate that our model achieves state-of-the-art (SOTA) performance. On the 2016 PhysioNet Challenge dataset, it reached an accuracy of 98.86% and an F1-score of 97.19%, significantly outperforming existing baseline methods. Conclusion and Significance: Our dual-branch fusion model provides a more effective and robust framework for heart sound classification. This work offers strong support for the development of highly accurate automated tools for the auxiliary diagnosis of CVDs, thereby holding the potential to enhance early detection and improve clinical outcomes.

Patients with sepsis-induced myocardial injury (SIMI) face a high risk of mortality. Although various biomarkers can be used to predict prognosis in SIMI patients, each has certain limitations. This study aimed to investigate the prognostic value of the lactate-to-calcium ratio (LCR) in patients with SIMI. This retrospective cohort study was conducted using data from the MIMIC-IV database. Patients diagnosed with SIMI who were admitted to the ICU were included. The LCR was calculated based on the first arterial blood gas analysis performed within 24 h of ICU admission. A restricted cubic spline (RCS) model was used to explore the nonlinear relationship between LCR and 28-day mortality. Patients were divided into high and low LCR groups based on the cutoff values, both overall and by sex. Kaplan–Meier survival curves were used to compare 28-day mortality between groups. Stratification analyses were conducted to assess the prognostic value of LCR across different age strata. A total of 1,631 patients were included. The RCS model revealed a positive association between higher LCR and increased 28-day mortality. The cut-off values for LCR were 2.96 for the overall population, 2.91 for females, and 2.16 for males. Cox regression analysis showed that high LCR was significantly associated with higher 28-day mortality (log-rank P < 0.001). Age-stratified analysis indicated that LCR had a higher predictive value in patients younger than 65 years. Among males, high LCR was associated with increased 28-day mortality only in those younger than 65. In females, the association was consistent regardless of age. A higher LCR is associated with increased 28-day mortality in ICU patients with SIMI. The sex-specific cut-off values (2.91 for females and 2.16 for males) suggest that LCR may serve as a useful prognostic indicator for identifying high-risk patients with sepsis-induced myocardial injury.

Background and objectiveConventional leadless pacemaker (LP) implantation relies on fluoroscopy, exposing patients and operators to ionizing radiation and contrast-related risks. Transthoracic echocardiography (TTE) is a radiation-free alternative, but complete TTE-guided LP implantation remains challenging due to poor ultrasound visibility of interventional devices. This study evaluated the short-term safety, technical feasibility, and procedural efficiency of completely TTE-guided LP implantation assisted by the ultrasound-optimized Panna guidewire.MethodsThis study utilized a prospectively protocolized, single-arm design for the TTE-guided cohort, with a retrospective comparative analysis against a historical fluoroscopy-guided control group. All safety and efficacy endpoints were formally predefined prior to patient enrollment. A total of 32 consecutive patients with LP implantation indications were screened during the study period (July 2024–July 2025), and 10 eligible patients underwent fluoroscopy/contrast-free, TTE-guided LP implantation using the Panna guidewire. Preoperative TTE acoustic window grading was performed, and standardized protocols (semi-quantitative “gooseneck” sign assessment, TTE-guided tug test) were applied during the procedure. A historical control group of 44 fluoroscopy-guided LP patients (January 2020–December 2023) was included, with propensity score overlap weighting-based comparative statistical analyses performed to balance baseline covariates and assess between-group differences. Procedural feasibility, short-term safety, pacing parameters, and skin-to-skin procedural duration were evaluated intraoperatively and during follow-up.ResultsAll 10 patients had optimal TTE acoustic windows (Grade 1). Procedural success was 100%, with no adverse events (median follow-up: 4.7 months) and stable device performance. Sensitivity analysis showed the TTE technique's effectiveness was not affected by operator experience. Compared with 44 propensity score-weighted controls, TTE-guided implantation had slightly longer but comparable procedural duration (62.78 ± 13.05 vs. 60.5 ± 19.1 min, P > 0.05) and comparable efficiency, eliminating radiation/contrast-related risks for high-risk patients (e.g., CKD, radiation sensitivity). Long-term follow-up (12/24 months) is ongoing per schedule.ConclusionsThis preliminary experience demonstrates the short-term safety and technical feasibility of completely TTE-guided LP implantation assisted by the Panna guidewire, which eliminates radiation/contrast risks while matching fluoroscopy-guided efficiency. As a hypothesis-generating proof-of-concept study (small sample, incomplete long-term follow-up), these findings require validation in larger multicenter registries (n ≥ 50) with ≥24-month follow-up to confirm long-term safety and generalizability.

Background: Congenital heart disease (CHD) is a major cause of infant mortality worldwide. Phonocardiogram (PCG) analysis offers a low-cost, non-invasive screening solution, yet existing feature extraction methods struggle to reliably distinguish pathological from normal heart sounds. Novel techniques capable of capturing subtle time–frequency characteristics are essential for accurate computer-aided diagnosis of CHD. Method: This study proposes Mel-scaled Frequency-domain Polynomial Chirplet Transform (Mel-FPCT), a novel time–frequency representation for heart sound classification. Mel-FPCT effectively models the complex frequency modulation patterns in heart sounds. A 12-layer CNN was trained on Mel-FPCT features to classify four heart sound types: normal, ASD, VSD, and PDA. To assess feature generalizability, we also evaluated Mel-FPCT with LSTM and ResNet-18. Performance was measured using hold-out and 5-fold cross-validation, and compared against multiple time–frequency features (STFT, MFCC, SFFT, FCT, FPCT). Results: With CNN, Mel-FPCT achieved accuracies of 0.9536 (hold-out) and 0.9439 (5-fold CV), outperforming the best baseline (FPCT) by 2.5 percentage points. When transferred to LSTM and ResNet-18, Mel-FPCT maintained stable accuracies above 85%, demonstrating strong cross-model generalizability. Mel-FPCT consistently outperformed all comparative methods on the evaluated heart sound dataset. The method proposed in this paper warrants validation in broader clinical settings.

BACKGROUND:The acquisition of large-scale phonocardiogram (PCG) data is crucial for clinical research and has significantly advanced the application of data-driven heart sound classification models. Owing to the complexity of clinical environments, the quality of collected PCG data cannot be guaranteed. The exclusion of low-quality signals is essential for conducting a reliable PCG analysis. METHODS:This study introduces a novel quality assessment model, the Parallel Channel Sequence Convolutional Network (PCSCN). The PCSCN automatically and accurately detects and removes low-quality PCG signals, thereby improving dataset reliability and usability. Unlike previous methods that rely on manually extracting numerous statistical features, the PCSCN employs low-complexity features and a multi-channel sequence architecture, offering greater accuracy and efficiency. In addition, this study applies the PCSCN in a controlled experiment for database cleaning and enhancement to verify the impact of this process on downstream data-driven PCG classification models. RESULTS:When tested on a public dataset, PCSCN achieved an accuracy of 95.45% and an F1-score of 95.44%. In a database cleaning task involving PCG data from 7220 subjects, PCSCN completed the task in 515.76s. Furthermore, in the controlled experiment, the PCG classification model trained with PCSCN-enhanced data demonstrated enhanced performance across multiple metrics. CONCLUSIONS:The PCSCN is both reliable and efficient, and is poised to play a pivotal role in the cleaning of large-scale PCG databases. By enhancing and improving the quality of the training data, the PCSCN significantly strengthens the clinical decision-making capabilities of PCG classification models, thereby elevating their value for clinical research and application.

The global prevalence of heart failure is still growing, which imposes a heavy economic burden. The role of microRNA-146b (miR-146b) in HF remains largely unknown. This study aims to explore the role and mechanism of miR-146b in HF. Method: We applied reverse transcription-polymerase chain reaction to search for differential microRNAs between myocardial tissues of heart failure patients and controls. We also used reverse transcription-polymerase chain reaction to detect the miR-146b expression in primary neonatal mouse cardiomyocytes and mice models of doxorubicin-induced HF. In vivo experiments, echocardiography was performed at baseline and weeks 6. After that we harvested mice's heart and evaluated the cardiomyocyte with hematoxylin and eosin (HE), Masson trichrome staining, and TUNEL staining. Through bioinformatics analysis, we found HIF-1 alpha might be the target gene of miR-146b, which validated by luciferase reporter gene assay. Subsequently, mRNA and protein expression levels of HIF-1 alpha were detected by overexpression or inhibition of miR-146b in primary neonatal mouse cardiomyocytes. Results: We found that miR-146b expression was decreased in myocardial tissues of HF patients compared with controls (P < 0.01). MiR-146b levels were notably downregulated in HF models. MiR-146b knockout mice showed a more pronounced decrease in cardiac function and more severe myocardial fibrosis and apoptosis than wild type. Meanwhile, over expression or repression of miR-146b in primary neonatal mouse cardiomyocytes could inhibit or upregulate HIF-1 alpha mRNA and protein expression. Conclusion: Our study shows that miR-146b may be a protective factor for cardiomyocytes by modulating HIF-1 alpha.

Activated astrocytes and their associated inflammatory responses play critical roles in the pathogenesis of hypoxic-ischemic brain damage (HIBD). Gastrodin (GAS), an anti-inflammatory herbal agent, is known to suppress microglial activation. Here, we investigate whether it exerts a similar effect on activated astrocytes and whether it acts through S100B/RAGE-Smad3 signaling. The expression changes of S100B/RAGE-Smad3 signaling pathway-related proteins, inflammatory factors and A1/A2 astrocyte markers were detected by ELISA, western blot analysis, immunofluorescence and immunohistochemistry. The results show that GAS decreases the expression of sRAGE in the brain tissue and S100B in the serum and brain tissue of HIBD mice. However, it promotes the expression of sRAGE in the serum of HIBD mice. Moreover, GAS inhibits the expressions of RAGE, p-Smad3, TNF-α, and C3 (A1 astrocyte marker), and promotes the expressions of S100A10 (A2 astrocyte marker) and BDNF in HIBD model mice, as well as in oxygen glucose deprivation (OGD)-treated TNC-1 astrocytes. The immunofluorescence and immunohistochemical results of RAGE and p-Smad3, as well as the immunofluorescence results of C3 and S100A10, reveal the same trend. Interestingly, FPS-ZM1 (a specific inhibitor of RAGE) inhibits the expressions of p-Smad3, TNF-α, C3, and S100A10, but promotes that of BDNF compared with those in the OGD group. The combination of GAS and FPS-ZM1 further decreases the expression of C3. These results indicate that GAS can inhibit the activation of Smad3 through S100B/RAGE signaling and regulate the expression of A1/A2-type astrocytes.

Congenital heart disease (CHD) is a serious cardiovascular disease in infants. Timely diagnosis is essential to cure CHD, especially in infants under 3. To bridge the gap of multi-classification algorithms for CHD in infant, a novel Gated Swin-Transformer (G-Swin-T) with Multi-scale Feature Fusion algorithm has been proposed. In the preprocessing stage, data enhancement was used to address the inherent imbalance problem of biomedical signals. In the feature extraction stage, G-Swin-T dynamically adjusts the influence of each input on the final result according to the classification effect. Multi-scale Feature Fusion combines different scale features obtained from the hierarchical structure to avoid information loss. Compared with the original Swin-Transformer, the accuracy and stability were further improved. During model training, Entropy Regularization was innovatively integrated into the loss function to encourage the model to make different predictions for each category, thereby improving category discrimination. Given the limited infant data, Transfer Learning was applied. Based on the pre-trained model on the esc-50 dataset, Layer-wise Learning Rate was set to train infant data. Finally, the algorithm achieves an accuracy of 96.66% for six-classification (VSD\ASD\PDA\PAH\Other types of CHD\Normal) on self-constructed dataset of 423 infants. The five-classification (VSD\ASD\PDA\PFO\Normal) accuracy was 96.62 % on the publicly available ZCHSound high-quality dataset (including 941 children), while 85.92 % on the low-quality dataset (including 318 newborns). Compared with the original paper, there is an improvement of 3.22 % and 30.22 %. The model proposed in this paper can effectively and accurately classify common CHDs automatically, thus serving as an aid for initial diagnosis during screening in remote areas.

Coronary heart disease (CHD) affects a large number of patients, and their survival period with the condition is relatively long. Cardiac shock wave therapy (CSWT) is an anti-myocardial ischemic therapy. This study aims to investigate the rehabilitative effects and underlying mechanisms of CSWT combined with exercise rehabilitation in patients undergoing percutaneous coronary intervention (PCI) for CHD. The study was a prospective randomized controlled trial conducted at Yunnan Fuwai Cardiovascular Hospital from 2020 to 2022, enrolling 74 patients who underwent PCI for CHD. They were randomly assigned to two groups: a control group of 37 patients and a study group of 37 patients. The intervention consisted of CSWT combined with exercise rehabilitation. Following the intervention, patients in the study group demonstrated increases in peak oxygen uptake (peak VO2), high-density lipoprotein cholesterol (HDL-C) and other parameters compared to pre-treatment levels. The combination of CSWT and exercise rehabilitation was found to improve myocardial markers, related anti-inflammatory and anti-atherosclerotic indicators in patients with CHD. Moreover, it alleviated myocardial ischemia while enhancing parameters such as oxygen uptake in cardiopulmonary exercise testing (CPET). These findings may offer novel insights into anti-ischemic treatment and cardiac rehabilitation for patients with CHD in future clinical practice.

ObjectiveCongenital heart disease with pulmonary arterial hypertension (CHD-PAH), caused by CHD, is associated with high clinical mortality. Hence, timely diagnosis is imperative for treatment.ApproachTwo non-invasive diagnosis algorithms of CHD-PAH were put forward in this review, which were direct three-divided and two-stage classification models. Pre-processing in both algorithms focuses on segmentation of heart sounds into discrete cardiac cycles. Both the dual-threshold and Bi-LSTM (Bi-directional Long Short-Term Memory) methods demonstrate efficacy. In the feature extraction phase, the direct three-divided model integrate time-, frequency-, and energy-domain features with deep learning features. While the two-stage classification model sequentially extracts sub-band envelopes and short-time energy of cardiac cycle. In the classification phase, considering the lack of CHD-PAH data, ensemble learning was widely used.Main resultsAn accuracy of 88.61% was achieved with direct three-divided model and 90.9% with two-stage classification model.SignificanceBy analyzing and discussing these algorithms, future research directions of CHD-PAH assisted diagnosis were discussed. It is hoped that it will provide insight into prediction of CHD-PAH. Thus saving people from death due to untimely assistance.

Heart sound analysis is a non-invasive and economical technique that can aid in diagnosing cardiovascular disease. A novel End-to-End heart sound classification method was proposed in this paper, in which a combination of multi-scale dense network and multi-head recurrent neural network technology was used. It can be used to diagnose congenital heart disease (CHD) without using the manual extraction of features. An Fβ score of 94.33% and an accuracy of 94.41% were achieved by the method on dataset A, which consisted of 1,000 individuals and 5,000 signals. Similarly, the widely used dataset B (Physio Net/CinC 2016 dataset), comprising 764 individuals and 3,240 signals, resulted in an Fβ score of 93.75% and an accuracy of 92.97%. The results show the proposed method had a significant potential to assist in diagnosing CHD. The SHAP algorithm which is a kind of Interpretable method was applied in this study to interpret the prediction results of model. It was shown that the model’s prediction process is similar to a doctor’s diagnosing mode.

The multi-window time-frequency reassignment helps to improve the time-frequency resolution of bark-frequency spectral coefficient (BFSC) analysis of heart sounds. For this purpose, a new heart sound classification algorithm combining feature extraction based on multi-window time-frequency reassignment BFSC with deep learning was proposed in this paper. Firstly, the randomly intercepted heart sound segments are preprocessed with amplitude normalization, the heart sounds were framed and time-frequency rearrangement based on short-time Fourier transforms were computed using multiple orthogonal windows. A smooth spectrum estimate is calculated by arithmetic averaging each of the obtained independent spectra. Finally, the BFSC of reassignment spectrum is extracted as a feature by the Bark filter bank. In this paper, convolutional network and recurrent neural network are used as classifiers for model comparison and performance evaluation of the extracted features. Eventually, the multi-window time-frequency rearrangement improved BFSC method extracts more discriminative features, with a binary classification accuracy of 0.936, a sensitivity of 0.946, and a specificity of 0.922. These results present that the algorithm proposed in this paper does not need to segment the heart sounds and randomly intercepts the heart sound segments, which greatly simplifies the computational process and is expected to be used for screening of congenital heart disease.

BACKGROUND:International studies have shown that use of a subcutaneous implantable cardioverter defibrillator (S-ICD) could reduce lead-related complications while maintaining adequate defibrillation performance; however, data from the Chinese population or other Asian groups are limited. MATERIAL AND METHODS:SCOPE is a prospective, multicenter, observational cohort study. Two hundred patients with primary prevention indication for sudden cardiac death (SCD), who are candidates for S-ICD, will be enrolled. From the same population, another 200 patients who are candidates for transvenous implantable cardioverter defibrillator (TV-ICD) will be enrolled after being matched for age, sex, SCD high-risk etiology (ischemic cardiomyopathy, and non-ischemic cardiomyopathy, ion channel disease, and other) and atrial fibrillation in a 1: 1 ratio with enrolled S-ICD patients. All the patients will be followed for 18 months under standard of care. RESULTS:The primary endpoint is proportion of patients free from inappropriate shock (IAS) at 18 months in the S-ICD group. The lower 95% confidence bound of the proportion will be compared with a performance goal of 90.3%, which was derived from the previous meta-analysis. The comparisons between S-ICD and TV-ICD on IAS, appropriate shock, and complications will be used as secondary endpoints without formal assumptions. CONCLUSIONS:This is the first prospective multicenter study focusing on the long-term performance of S-ICD in a Chinese population. By comparing with the data derived from international historical studies and a matched TV-ICD group, data from SCOPE will allow for the assessment of S-ICD in the Chinese population in a contemporary real-world implantation level and programming techniques, which will help us to further modify the device implantation and programming protocol in this specific population in the future.

Aiming at the problems of obscure clinical auscultation features of pulmonary hypertension associated with congenital heart disease and the complexity of existing machine-aided diagnostic algorithms, an algorithm based on the statistical characteristics of the high-frequency components of the second heart sound signal is proposed. Firstly, an endpoint detection adaptive segmentation method is employed to extract the second heart sounds. Subsequently, the high-frequency component of the heart sound is decomposed using the discrete wavelet transform. Statistical features including the Hurst exponent, Lempel-Ziv information and sample entropy are extracted from this component. Finally, the extracted features are utilized to train an extreme gradient boosting algorithm (XGBoost) classifier, which achieves an accuracy of 80.45% in triple classification. Notably, this method eliminates the need for a noise reduction algorithm, allows for swift feature extraction, and achieves effective multi-classification using only three features. It is promising for early screening of pulmonary hypertension associated with congenital heart disease.

We present the case of a patient referred for conduction system pacemaker implantation after a complete block associated with transcatheter aortic valve replacement and reduced left ventricular function. In this case, we achieved bilateral septal pacing through the anodal capture of the right ventricular septum during bipolar pacing. This approach generated superior ventricular mechanical synchrony compared to left ventricular septum pacing (LVSP).

Objective Formononetin (FOR), a traditional Chinese medicine, has been widely used for nerve protection and nerve function rehabilitation after cerebral stroke. However, the role of FOR in autophagic lysosome function in cerebral ischemia reperfusion damage has not been investigated. This study aimed to explore whether the therapeutic benefits of FOR were influenced by the regulation of autophagy flux. Methods Male Sprague-Dawley rats were separated into sham, model, and MCAO+FOR (30 mg/kg) groups after undergoing middle cerebral artery occlusion (MCAO) and ischemia-reperfusion (I/R). Then, the brain tissues in the ischemic penumbra were obtained to detect the proteins in autophagic/lysosomal pathway with antibodies of Beclin-1, LC3, SQSTM1/P62, Ubiquitin, LAMP-2, Cathepsin B (CTSB) and Cathepsin D (CTSD) by Western blot and immunofluorescence, respectively. Meanwhile, the therapeutic effectiveness was evaluated by measuring infarct volume, neurological impairments, and neuronal necrosis. Results The findings of this study demonstrate that FOR treatment exhibits a dual effect by enhancing the autophagic activities of Beclin-1 and LC3 in neurons, while simultaneously improving the autophagic clearance function, as evidenced by reinforced lysosomal activities of LAMP-2, CTSB, and CTSD, as well as reduced autophagic accumulation of Ubiquitin and P62 in the MCAO+FOR group compared to the MCAO group. Additionally, 7 d of FOR treatment dramatically reduced neurological deficits, infarct volume, and neuronal death caused by cerebral ischemia. Conclusion These findings suggest that the neuroprotective mechanism of FOR therapy in accelerating recovery from ischemic stroke may involve the increase of autophagy flux in the penumbra.

Heart sound signals are vital for the machine-assisted detection of congenital heart disease. However, the performance of diagnostic results is limited by noise during heart sound acquisition. A limitation of existing noise reduction schemes is that the pathological components of the signal are weak, which have the potential to be filtered out with the noise. In this research, a novel approach for classifying heart sounds based on median ensemble empirical mode decomposition (MEEMD), Hurst analysis, improved threshold denoising, and neural networks are presented. In decomposing the heart sound signal into several intrinsic mode functions (IMFs), mode mixing and mode splitting can be effectively suppressed by MEEMD. Hurst analysis is adopted for identifying the noisy content of IMFs. Then, the noise-dominated IMFs are denoised by an improved threshold function. Finally, the noise reduction signal is generated by reconstructing the processed components and the other components. A database of 5000 heart sounds from congenital heart disease and normal volunteers was constructed. The Mel spectral coefficients of the denoised signals were used as input vectors to the convolutional neural network for classification to verify the effectiveness of the preprocessing algorithm. An accuracy of 93.8%, a specificity of 93.1%, and a sensitivity of 94.6% were achieved for classifying the normal cases from abnormal one.

Introduction: Congenital heart disease (CHD) is a cardiovascular disorder caused by structural defects in the heart. Early screening holds significant importance for the effective treatment of this condition. Heart sound analysis is commonly employed to assist in the diagnosis of CHD. However, there is currently a lack of an efficient automated model for heart sound classification, which could potentially replace the manual process of auscultation.Methods: This study introduces an innovative and efficient screening and classification model, combining a locally concatenated fusion approach with a convolutional neural network based on coordinate attention (LCACNN). In this model, Mel-frequency spectral coefficients (MFSC) and envelope features are locally fused and employed as input to the LCACNN network. This model automatically analyzes feature map energy information, eliminating the need for denoising processes.Discussion: The proposed classification model in this study demonstrates a robust capability for identifying congenital heart disease, potentially substituting manual auscultation to facilitate the detection of patients in remote areas.Results: This study introduces an innovative and efficient screening and classification model, combining a locally concatenated fusion approach with a convolutional neural network based on coordinate attention (LCACNN). In this model, Mel-frequency spectral coefficients (MFSC) and envelope features are locally fused and employed as input to the LCACNN network. This model automatically analyzes feature map energy information, eliminating the need for denoising processes. To assess the performance of the classification model, comparative ablation experiments were conducted, achieving classification accuracies of 91.78% and 94.79% on the PhysioNet and HS databases, respectively. These results significantly outperformed alternative classification models.

Feature extraction methods and classifier selection are two critical steps in heart sound classification. To capture the pathological features of heart sound signals, this paper introduces a feature extraction method that combines mel-frequency cepstral coefficients (MFCC) and power spectral density (PSD). Unlike conventional classifiers, the adaptive neuro-fuzzy inference system (ANFIS) was chosen as the classifier for this study. In terms of experimental design, we compared different PSDs across various time intervals and frequency ranges, selecting the characteristics with the most effective classification outcomes. We compared four statistical properties, including mean PSD, standard deviation PSD, variance PSD, and median PSD. Through experimental comparisons, we found that combining the features of median PSD and MFCC with heart sound systolic period of 100-300 Hz yielded the best results. The accuracy, precision, sensitivity, specificity, and F1 score were determined to be 96.50%, 99.27%, 93.35%, 99.60%, and 96.35%, respectively. These results demonstrate the algorithm's significant potential for aiding in the diagnosis of congenital heart disease.

BACKGROUND: Dual-phase cardiac computed tomography (CCT) has been applied to detect left atrial appendage (LAA) thrombosis, which is characterized as the presence of left atrial appendage filling defects (LAADF) in both early- and delayed-phase scanning. However, the clinical implication of LAAFD in exclusive early-phase scanning (LAAFD-EEpS) of CCT in patients with atrial fibrillation (AF) is unclear. METHODS: The baseline clinical data and dual-phase CCT findings in 1183 AF patients (62.1 ± 11.6 years, 59.9% male) was collected and analyzed. A further analysis of CCT and transesophageal echocardiography (TEE) data (within 5 days) in a subgroup of 687 patients was performed. LAAFD-EEpS was defined as LAAFD present in early-phase and absent in delayed-phase scanning of dual-phase CCT. RESULTS: A total of 133 (11.2%) patients were detected with LAAFD-EEpS. Patients with LAAFD-EEpS had a higher prevalence of ischemic stroke or transient ischemic attack (TIA) (p < 0.001) and a higher predefined thromboembolic risk (p < 0.001). In multivariate analysis, a history of ischemic stroke or TIA was independently associated with LAAFD-EEpS (odds ratio [OR] 11.412, 95% confidence interval [CI] 6.561–19.851, p < 0.001). When spontaneous echo contrast in TEE was used as the reference standard, the sensitivity, specificity, positive predictive value, and negative predictive value of LAAFD-EEpS was 77.0% (95% CI 66.5–87.6%), 89.0% (95% CI 86.5–91.4%), 40.5% (95% CI 31.6–49.5%), 97.5% (96.3–98.8%), respectively. CONCLUSIONS: In AF patients, LAAFD-EEpS is not an uncommon finding in dual-phase CCT scanning, and is associated with elevated thromboembolic risk.