Learning from Scarce Labels: Multi-View Echocardiography for Ejection Fraction Prediction

Machine Learning for Biomedical Imaging Learning from Scarce Labels: Multi-View Echocardiography for Ejection Fraction Prediction

Beyond Doppler: Scalable AI Detection of LVOT Obstruction in HCM.

BACKGROUND Accurate assessment of left ventricular outflow tract (LVOT) gradients is critical for hypertrophic cardiomyopathy management, yet Doppler-based measurements are technically demanding and require expertise. The objective of this work was to develop a multi-view deep learning model capable of classifying LVOT obstruction (>20 mm Hg) using routine 2-dimensional echocardiographic windows without reliance on Doppler imaging. METHODS We trained and externally validated a cross-attention-based video-to-video fusion framework that integrated EchoPrime-derived video representations from 3 standard transthoracic echocardiographic views to classify LVOT gradients. RESULTS Training was performed on a derivation cohort (N=1833) from a tertiary care system in the United States, with model performance evaluated on an internally held-out test set (N=275) and a Korean external validation cohort (N=46). Single-view baselines showed limited discrimination (external area under the receiver operating curves, 0.47-0.70). Conversely, the domain-specific foundational model (EchoPrime) achieved superior single-view performance (area under the receiver operating curves, 0.75-0.80 internal; 0.79-0.83 external), highlighting the importance of echo-specific pretraining and temporal modeling. The proposed multi-view fusion further enhanced predictive performance, with the late fusion model reaching an area under the receiver operating curve of 0.84 on the external cohort with significant population-shift. CONCLUSIONS These results suggest LVOT physiology is encoded in routine 2-dimensional imaging and can be leveraged for clinically relevant gradient classification without Doppler input. The proposed artificial intelligence-guided strategy demonstrates substantial cost savings compared with the screen-all approach. By integrating complementary spatial-temporal information across multiple views, our approach generalizes robustly across populations and may enable real-time decision support, extend LVOT assessment to portable or resource-limited settings, and complement Doppler-based evaluation for longitudinal hypertrophic cardiomyopathy management.

Domain Shift in Echocardiography: Interpretable Quantification and Prediction of Cross-Dataset Left Ventricular Segmentation

Cross-dataset generalisation remains a major barrier to clinical deployment of echocardiographic left ventricular segmentation, yet the sources of this shift are rarely disentangled. We examined whether transfer degradation could be estimated before deployment using handcrafted ultrasound descriptors, VAE latent features, and segmentation-derived latent features across six echocardiographic datasets. Geometry-aware preprocessing substantially improved several poor transfer cases, suggesting that much of the apparent domain shift reflects field-of-view and framing inconsistencies rather than intrinsic acoustic differences alone. Intensity z-normalisation changed dataset separability by less than 0.005, indicating that brightness and contrast are not the dominant shift axis. Absolute Dice drop on held-out source-target pairs was predicted with an R-squared value of 0.612, an MAE of 0.082, and a Spearman rho of 0.681. The variant without LV and fan-shaped features retained approximately 70% of this explanatory power, supporting mask-free transfer-risk monitoring. The most informative discrepancy measure depended on the representation, with CMD strongest in z-normalised handcrafted features, with an absolute r of approximately 0.86 and an R-squared value of approximately 0.70; log-Wasserstein strongest in VAE space, with an r of approximately -0.90 and an R-squared value of approximately 0.81; and log-MMD strongest in LV-segmentation latent features, with an r of approximately -0.92 and an R-squared value of approximately 0.84. Apparent vendor effects were largely dataset-confounded. Echocardiographic domain shift is therefore structured and measurable, and its impact on segmentation can be partly reduced through geometry-aware preprocessing and anticipated using representation-specific transfer-risk estimation.

Ejection fraction on a budget: mapping the accuracy-compute trade space for video-based ejection fraction estimation

Deep video networks estimate left ventricular ejection fraction (EF) from echocardiograms with expert-level accuracy, but the compute cost of running them is rarely reported. This leaves anyone building a handheld or bedside tool without clear guidance on what to deploy. We measured the accuracy-versus-compute tradeoff for EF estimation on EchoNet-Dynamic by training 22 configurations that varied clip length (8 to 64 frames), frame sampling period (1 to 4), and backbone: R(2+1)D-18, R3D-18, MC3-18, X3D-S, X3D-M, and a 2D ResNet-18 with temporal pooling. All models used one fixed training recipe. Every configuration was evaluated for accuracy using mean absolute error, R-squared, and Bland-Altman agreement; clinical utility using sensitivity and specificity at the clinically relevant EF cutoffs of 40% and 50%, plus error stratified by EF band; and cost using floating-point operations, parameter count, GPU and CPU latency, and peak memory under a single frozen measurement protocol. We stress-tested the main findings with replicate training seeds. Sparse temporal sampling outperformed dense sampling at matched frame budgets. A sampling period of 4 outperformed a period of 1 at every tested frame count while also reducing per-video cost. In the seed-replicated 8-frame comparison, the advantage averaged one full point lower mean absolute error across all nine cross-seed pairings. A standard R3D-18 achieved the best accuracy in the study, with a mean absolute error of 3.99, while requiring 19% less CPU latency than the reference configuration. A 16-frame, period-4 R(2+1)D-18 cut reference cost in half with no statistically confirmed loss in accuracy. Removing temporal modeling entirely substantially reduced accuracy, with a mean absolute error of 5.65, setting a practical floor for how inexpensive this task can be. We release the code, cost-measurement protocol, and per-configuration results.

Point-of-care echocardiography screening for hypertrophic cardiomyopathy using automated deep-learning analysis

Abstract Aims Hypertrophic cardiomyopathy (HCM) remains underdiagnosed due to limited access to expert imaging. We developed and validated a deep-learning (DL)-based echocardiographic model adaptable to point-of-care ultrasound (POCUS) for scalable HCM screening. Methods and results We retrospectively analysed 134 956 expert transthoracic echocardiograms (TTE) from 73 598 patients at Sheba Medical Center (2007–2022). A TTE-trained DL model integrating structural features and temporal motion patterns from parasternal long-axis and apical four-chamber views estimated HCM probability. Performance was evaluated in an independent test cohort and clinical subgroups. External validation used bedside POCUS studies from non-cardiologists with handheld devices. The test cohort included 12 096 patients with 119 confirmed HCM cases (prevalence 0.98%; median age 75 years, 57% male). HCM-positive patients showed increased expert TTE-measured septal (1.67 [1.5, 2.0] vs. 1.01 [0.9, 1.19] cm) and posterior wall thickness (1.1 [1.0, 1.3] vs. 0.9 [0.8, 1.0] cm) (P < 0.001). The model achieved excellent discrimination with an area under the curve of 0.982 (95% CI 0.966–0.993), sensitivity 88.2%, and specificity 97.3%, robust across subgroups. The POCUS cohort (n = 1047, median age 73 years, 55% male) represented multimorbid inpatients with 65 (6.2%) classified as screen-positive by the algorithm. These showed higher expert TTE-measured septal thickness (1.26 [1.07, 1.46] vs. 1.06 [0.9, 1.2] cm; 22% vs. 4% with IVS ≥1.5 cm; P ≤ 0.01). Among 49 (75%) POCUS-flagged positive patients with formal TTE and clinical data, 8 (16%) were confirmed by expert adjudication to have HCM. Specificity is limited by occasional confounding amyloidosis detection (4% of POCUS-flagged patients). Conclusion This DL-based model identifies HCM and demonstrates feasibility for POCUS screening, supporting earlier detection and broader diagnostic access.

ROI-Guided Echocardiographic Image Analysis and Case Aggregation for Differentiating Atrial and Ventricular Septal Defects

This study implemented strict subject-disjoint partitioning to eliminate data leakage, and simultaneously introduced cross-frame case aggregation to emulate the multi-frame visual synthesis process of expert echocardiographers, suggesting that the proposed workflow has the potential to serve as an adjunctive tool for septal defect screening.