To develop and validate machine learning (ML) models integrating two-dimensional speckle tracking echocardiography (2D-STE) parameters with clinical variables for robust identification of severe coronary artery disease (sCAD).
In this retrospective cohort study, five distinct ML models (Random Forest [RF], Support Vector Machine [SVM], K-Nearest Neighbors [KNN], Multi-Layer Perceptron [MLP], and Extremely Randomized Trees [Extra Trees]) were constructed to identify sCAD on a cohort of 204 patients (80% training set, 20% independent test set). Within the independent test set, two junior sonographers' diagnostic performance for sCAD was compared first without and then with ML assistance over a 2-week interval. SHapley Additive exPlanations (SHAP) analysis was applied to visualize and interpret the models, identifying key features driving sCAD prediction accuracy, with results visualized through dependence diagrams and force plot. Furthermore, a clinical nomogram integrating key predictors identified by ML models was developed to enable individualized quantification of sCAD risk.
Utilizing five features, the MLP demonstrated the best performance with an area under the curve (AUC) of 0.870 and a sensitivity of 0.944. The SHAP visualization analysis for this model indicated that “LV AP4 Endo Peak L. Time SD” significantly influenced its predictions. The MLP model (AUC = 0.870) outperformed both junior sonographers (AUC = 0.687) and a nomogram constructed from ML-selected features (AUC = 0.712). Additionally, the results revealed that junior sonographers achieved significantly improved performance when assisted by the ML models.
The developed ML models could differentiate patients with angiography-confirmed sCAD from those without. Importantly, these models significantly improved the diagnostic performance of junior sonographers when used as an assistive tool.
Objective: Studies have confirmed that a RFOF (redundant foramen ovale flap) can mimic CoA (coarctation of the aorta) in terms of traditional indicators; however, there is currently no more comprehensive and feasible indicator to differentiate between the two. This study aims to investigate the effect of RFOF on cardiac output, and explore the ability of multiple left cardiac parameters to differentiate RFOF from CoA.
Methods: This prospective diagnostic cohort study consecutively enrolled fetuses with suspected CoA and normal controls. Comprehensive echocardiographic measurements, including cardiac output and novel morphological parameters-including RCO/LCO (right cardiac output/left cardiac output), the FOFd/LAd (foramen ovale flap/left atrial diameter ratio), aortic arch angle, and LIOA (left ventricular inflow-outflow tract angle)-were performed by operators blinded to the postnatal outcome. ROC (Receiver operating characteristic) curve analysis was performed to evaluate the predictive ability of all indicators for pregnancy outcomes.
Results: Analysis revealed that in the RFOF Group, FOFd/LAd had the highest AUC (area under the curve), followed by RCO/LCO. In the CoA Group, the aortic isthmus/ductus arteriosus diameter (Aoi/DA) and aortic arch angle showed the highest AUC values.
Conclusion: RFOF can mimic the ultrasonic manifestations of CoA in terms of traditional indicators. FOFd/LAd > 0.65 (AUC = 1.000) is identified as a specific marker for RFOF, while RCO/LCO > 1.60 (AUC = 0.966) provides strong supportive evidence. Conversely, an aortic arch angle > 75.84° receiver operating characteristic (AUC = 0.995) combined with Aoi/DA < 0.75 (AUC = 0.999), suggests a higher probability of CoA. FOFd/LAd, aortic arch angle, and LIOA can help distinguish between RFOF and CoA.
Summary: What's already known about this topic? The established consensus indicates that RFOF can mimic the sonographic features of CoA prenatally by causing imbalances in traditional parameters of cardiac chamber dimensions and great vessel diameter ratios. However, reliable tools for prenatally distinguishing between these two conditions based on hemodynamic and left heart morphological characteristics remain unavailable. What does this study add? This study adds a new, multiparameter approach that combines hemynamic (RCO/LCO) and novel morphological (FOFd/LAd, Aortic Arch Angle, LIOA) assessments. This approach moves beyond the limitations of traditional ratios and provides clinicians with practical tools to reduce the false positive diagnosis of CoA in fetuses with RFOF.
To explore medical students’ perceptions of integrating handheld focused cardiac ultrasound (FoCUS) into the internal medicine curriculum and identify factors influencing learning outcomes.
A cross-sectional pilot survey was conducted among 145 fourth-year medical students at “Grigore T. Popa” University of Medicine and Pharmacy, Iași, Romania, during the 2024–2025 academic year. During their internal medicine rotation, students performed FoCUS echocardiography at the bedside, using a handheld ultrasound device. A structured online questionnaire assessed perceptions of training quality, usefulness in curriculum, impact on understanding cardiac anatomy and physiology, improvement in echocardiographic skills, and interest in integrating ultrasound more broadly into the curriculum.
Among 145 participants (mean age 23.4 ± 2.1 years, 67.6% female), 130 (89.6%) reported using the handheld devices, with a mean frequency of 1.66 ± 0.73 times per semester. Students rated FoCUS as highly useful for bedside learning (3.92 ± 1.25/4), for understanding cardiac anatomy and physiology (4.02 ± 1.29/5), and for enhancing echocardiographic skills (3.75 ± 1.35/5). Higher usage frequency corresponded with increased perceived learning efficacy. Training quality (β = 0.461, 95% CI: 0.216–0.706) and practical demonstrations (β = 0.607, 95% CI: 0.040–1.175) were significant predictors of educational outcomes. Strong support was expressed for formal ultrasound curriculum inclusion (97.6%) and interdisciplinary expansion (99.2%).
Integrating handheld FoCUS into the internal medicine curriculum is associated with substantial perceived educational benefits. Training quality and hands-on demonstrations are key determinants of successful learning outcomes, supporting the systematic inclusion of FoCUS as a core component of medical education.
Exaggerated blood pressure response (ExBPR) to exercise, often defined as peak systolic blood pressure (SBP) ≥ 210 mm Hg in men and ≥190 mm Hg in women, has limited clinical utility due to inconsistent prognostic data. Workload-indexed SBP, expressed as SBP/metabolic equivalent of task (MET) slope, has emerged as a potentially superior marker of cardiovascular risk. This study aims to evaluate the association of exercise SBP parameters with left ventricular (LV) remodeling and function, and to unveil which exercise SBP marker more accurately reflects adverse cardiac remodeling and function.
We retrospectively studied 455 individuals who underwent echocardiography and treadmill testing within 1 day at a single center in Korea. Echocardiographic parameters included left ventricular (LV) end-diastolic dimension (EDD), relative wall thickness (RWT) and e′ velocity. SBP/MET slope was calculated as (peak—resting SBP)/(maximal METs−1).
Participants with SBP/Met slope >6.2 mm Hg/MET were older, had higher body mass index, and exhibited smaller LV EDD index, higher RWT, and lower e′ velocity (all p < 0.05). They also showed higher resting and peak SBP and lower exercise capacity. ExBPR was associated with similar structural and functional abnormalities but not with exercise capacity. In individuals without ExBPR, higher SBP/MET slope identified smaller LV EDD index, higher RWT, and lower e′ velocity (all p < 0.05). Multivariable analyses confirmed independent associations of SBP/MET slope with concentric LV structure represented by high RWT and lower e′ velocity, beyond resting SBP. Adding SBP/MET slope to ExBPR improved prediction of reduced e′ velocity (p = 0.021).
SBP/MET slope was independently associated with concentric LV structure and diastolic dysfunction, providing incremental clinical value over ExBPR in detecting subclinical cardiac abnormalities.
Patients with aortic stenosis(AS)exhibit elevated covert brain injury. The association of AS-related cerebral impairment with cardiac remodeling and its prognostic impact remains unclear, prompting exploration of integrated cardio-cerebral assessment preliminarily.
Patients with moderate-to-severe aortic stenosis (AS) undergoing surgery at The First Affiliated Hospital of Guangxi Medical University (June 2021–December 2024) were enrolled (n = 384). Patients were stratified into a cerebral hypoperfusion/embolism group (n = 146) and an event-free group (n = 238) based on the presence or absence of preoperative cerebral impairment due to severe AS-related hemodynamic compromise or embolic events. Preoperative clinical/echocardiographic and surgical data were collected. The composite primary endpoint was major adverse cardiac events (MACE) (malignant arrhythmia, heart failure rehospitalization, or all-cause death). Multivariate regression identified predictors of AS-related cerebral impairment and MACE risk factors. The C-index assessed the cardio-cerebral integrated model's prognostic value.
The cerebral hypoperfusion/embolism group showed more pronounced left atrial (LA) and left ventricular (LV) remodeling. Multivariate logistic regression linked enlarged ascending aorta, left heart remodeling (increased LA size, elevated E/e' ratio, decreased LV global longitudinal strain and LA reservoir strain), reduced cardiac output (CO), and pulmonary artery systolic pressure to AS-related cerebral impairment (all p < 0.05). Over an 18-month median follow-up, 79 MACE occurred. Adjusted multivariate Cox analysis confirmed AS-related cerebral impairment as an independent MACE predictor (hazard ratio, 2.24 [95% CI: 2.13, 3.44]; p < 0.001). The cardio-cerebral integrated model (clinical + echo + ischemia) demonstrated superior MACE prediction (C-index, 0.862; likelihood ratio, 62.12) vs. clinical-only (C-index, 0.631; likelihood ratio, 19.32; both p < 0.001) or clinical + echo models (C-index, 0.757; likelihood ratio, 46.34; both p < 0.001).
AS-related cerebral impairment independently correlates with cardiac remodeling and adverse outcomes. Integrated cardio-cerebral assessment enhances risk stratification.
This virtual navigation of a fetal heart with hypoplastic left heart syndrome demonstrated the potential of this technology to provide realistic views of cardiac anatomy and to enable exploration of the ventricles, valves, and outflow tracts.
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：