Federico M. Asch

Background: Deep learning (DL) is increasingly used for the automated analysis of echocardiographic videos. These end-to-end pipelines often incorporate models targeting common preprocessing tasks to ensure the selection of videos and frames suitable for their primary analytic objective. However, publicly available models supporting these core steps remain limited. Objectives: We aimed to develop three DL models for distinct preprocessing tasks: Model-V for view classification, Model-O for orientation classification, and Model-CC for identifying end-diastolic (ED) and end-systolic (ES) frames. Methods: Model-V was trained and internally validated on 13,864 videos and externally evaluated on 5,650 videos. Model-O and Model-CC were trained and internally validated on 5,513 and 3,108 apical 4-chamber view (A4C) videos, respectively, and both were externally tested on the same set of 1,200 A4C videos. Results: Model-V discriminated A4C from non-A4C with balanced accuracies of 0.991 (0.987–0.994) and 0.895 (0.884–0.907), and Model-O distinguished Stanford from Mayo orientation with balanced accuracies of 0.995 (0.990–1.000) and 0.960 (0.948–0.971) during internal and external testing, respectively. Model-CC identified ED frames with mean absolute errors (MAEs) of 4.041 (3.878–4.210) and 3.117 (3.044–3.193) frames and ES frames with MAEs of 2.799 (2.708–2.894) and 4.827 (4.738–4.917) frames in the internal and external test sets, while failing to identify only 7.3 and 4.2% of ED frames and 1.2 and 3.6% of ES frames, respectively. Conclusions: The proposed models demonstrated good performance across key preprocessing tasks; thus, they constitute readily integrable components for end-to-end DL pipelines in automated echocardiographic analysis.

Purpose of review: The diagnostic evaluation of mitral regurgitation (MR) is complex, time-intensive, and prone to significant interobserver variability. This review examines the current evidence on artificial intelligence (AI) applications across the full MR diagnostic pathway, from pre-imaging screening to advanced multimodality imaging, and explores future directions for clinical integration. Recent findings: AI-enabled digital stethoscopes and deep learning-based electrocardiographic models provide scalable upstream strategies for early detection and population-level risk stratification, although they currently function as enrichment tools rather than standalone diagnostics. In echocardiography, AI has demonstrated strong performance for automated valve segmentation, Doppler analysis, severity grading, and phenotypic classification. In cardiac magnetic resonance, AI enables automated valve tracking, ventricular segmentation, and tissue characterization, although dedicated algorithms for direct MR quantification remain under development. Beyond automation, AI-driven approaches have identified clinically meaningful phenotypes linking valvular dysfunction and cardiac remodeling with myocardial fibrosis and increased cardiovascular risk. Summary: AI holds significant promise to improve reproducibility, consistency, and clinical integration of MR assessment. Widespread implementation, however, requires prospective validation, standardized acquisition protocols, improved model interpretability and, most importantly, proper regulations driven by a culture of safety.

BACKGROUND: Cardiac amyloidosis (CA) is an underdiagnosed yet treatable cause of heart failure in which timely diagnosis is essential to initiate life-prolonging therapies. While artificial intelligence (AI)-based tools using transthoracic echocardiography (TTE), electrocardiography, or electronic health records have demonstrated promise for CA detection, most rely on single data sources. We aimed to evaluate whether integrating clinical, laboratory, and TTE biomarkers improves the performance of an existing TTE-based AI model for CA detection. METHODS: We developed and tested a combined AI echo-clinical model (AI-ECM) incorporating demographics, laboratory biomarkers, and TTE parameters into a previously validated TTE-only AI model (Us2.Ca). Model training and internal validation were performed using the Amyloidosis Imaging International Consortium, a global multiethnic registry comprised of 727 patients with CA and 316 controls, including 202 with suspected transthyretin-CA with negative diagnostic evaluation and 114 patients with biopsy-proven extracardiac light chain amyloidosis without cardiac involvement. Ground truth CA diagnosis was adjudicated per consensus criteria. AI-ECM and Us2.Ca performance was assessed using area under the curve, accuracy, sensitivity, and specificity. RESULTS: In building the AI-ECM, feature importance analysis showed that having the Us2.Ca prediction scores, relative wall thickness, sex, and estimated glomerular filtration rate contributed most to performance. The AI-ECM demonstrated superior performance (area under the curve, 0.94; accuracy, 90%; sensitivity, 93%; specificity, 85%) compared with the Us2.Ca (area under the curve, 0.89; accuracy, 80%; sensitivity, 76%; specificity, 91%; P =0.006). While the Us2.Ca model classification was indeterminate in 9% of the cases, the AI-ECM allowed classification of all cases. The AI-ECM improved sensitivity for light chain-CA detection and maintained high accuracy across subtypes and control groups. CONCLUSIONS: A multiparametric AI model integrating basic clinical, laboratory, and TTE data with the deep learning Us2.Ca improved performance for CA detection over Us2.Ca alone. This approach represents a step toward scalable, AI-guided precision diagnostics for CA in diverse populations.

Background: Among cardiac measures, diastolic parameters demonstrate the earliest and most consistent age-related changes. This can be leveraged to develop a continuous left ventricular (LV) Diastolic Age from routine echocardiographic parameters. Analogous to how epigenetic clocks weight molecular markers against mortality risk, we calibrated Diastolic Age by weighting echocardiographic features against the validated PREVENT–Heart Failure (HF) risk score. Methods: We analyzed 1,952 participants from the Project Baseline Health Study (median age 50 [36–64] years, 54% female). The measure was derived using partial least-squares regression anchored on PREVENT-HF and calibrated within a healthy reference subgroup. External validation was performed in the WASE (n=1,708) and Stanford Cardiovascular Aging (n=313) cohorts. Associations with ASE-defined LV diastolic dysfunction (LVDD), epigenetic clocks, and major adverse cardiovascular events (MACE) were examined. Results: Diastolic Age correlated strongly with chronological age (r=0.78) with robust external validation (WASE r=0.76; Stanford r=0.82; calibration slopes ≈1.0). It increased progressively across grades of diastolic dysfunction and discriminated LVDD with an AUC of 0.89 (95% CI 0.87–0.92), and was independently associated with hypertension, diabetes, and elevated C-reactive protein. While correlated with the Levine (r=0.76) and Horvath (r=0.41) epigenetic clocks, residual analyses indicated that Diastolic Age captures a distinct cardiac-specific dimension of biological aging. Over median follow-up of 4.2 years, it independently predicted MACE (HR 2.30, 95% CI 1.70–3.18), with accelerated diastolic aging across all age groups among those with events. Discrimination was comparable to ASE-defined LVDD (C-index 0.83 vs. 0.82). Conclusions: Diastolic Age provides a continuous, echocardiography-derived measure of cardiac biological aging that complements categorical diastolic grading and epigenetic aging clocks, and independently predicts cardiovascular outcomes.

Background: Advanced age may influence the benefits derived from transcatheter edge‐to‐edge repair for primary mitral regurgitation (MR). We explored the impact of age on clinical, echocardiographic, functional, and patient‐reported outcomes after contemporary mitral transcatheter edge‐to‐edge repair. Methods: Patients with primary MR were identified from a pooled, patient‐level analysis of the EXPAND and EXPAND G4 studies. Key outcomes included 30‐day and 1‐year major adverse events, MR severity, New York Heart Association class, Kansas City Cardiomyopathy Questionnaire – Overall Summary score, and mortality. Results: A total of 854 patients were included. The oldest quartile ranged in age from 86 to 99 years, while the younger quartiles ranged from 21 to 85 years. The 30‐day major adverse events rate was not different in the oldest compared with the younger quartiles (5.7% versus 4.3%, P =0.46). At 1 year, MR ≤1+ was present in 81% of the oldest versus 88% of the younger patients ( P =0.08). LV end‐diastolic volumes improved to similar extent (−12.0±28.8 mL, P =0.003 and −14.1±31.5 mL, P P =0.34 across groups). Functional status and Kansas City Cardiomyopathy Questionnaire – Overall Score improved significantly across age groups, without differences according to age. Adjusted 1‐year mortality was higher in the oldest compared with the younger quartiles (adjusted hazard ratio 1.74 [95% CI, 1.02–2.98], P =0.04). Conclusions: Although 1‐year mortality was higher among the very elderly, mitral transcatheter edge‐to‐edge repair using next‐generation devices was associated with low complication rates and substantial improvements in ventricular remodeling, functional status, and quality of life across age groups. These findings support the use of mitral transcatheter edge‐to‐edge repair in appropriately selected elderly patients and suggest that age alone should not be a barrier to treatment.