Daily Cardiology Research Analysis
Analyzed 183 papers and selected 3 impactful papers.
Summary
Top findings span clinical AI, disease mechanisms, and population health. A publicly released 12‑lead ECG deep‑learning model (ECG2HF) robustly predicts 10‑year incident heart failure across three health systems. Mechanistically, a reduced TBX5‑dependent atrial gene regulatory network links atrial fibrillation and heart failure, while a nationwide study shows frailty is more predictive of 1‑year mortality after myocardial infarction in men than women, despite higher frailty prevalence in women.
Research Themes
- AI-enabled ECG risk prediction for incident heart failure
- Gene regulatory networks bridging atrial fibrillation and heart failure
- Sex-specific frailty effects on post-myocardial infarction outcomes
Selected Articles
1. Artificial Intelligence-Enabled ECG Analysis to Predict Incident Heart Failure.
A 12‑lead ECG deep-learning model (ECG2HF) trained in 94,636 patients and externally validated in 93,868 across three hospitals achieved AUCs of 0.84–0.86 for predicting 10‑year incident heart failure. It outperformed a 15‑variable clinical score by improving discrimination and net reclassification, and the model is publicly available.
Impact: Provides a generalizable, openly available AI tool for long-horizon HF risk that consistently outperforms a validated clinical score across independent health systems.
Clinical Implications: Supports population‑level and point‑of‑care HF prevention by identifying high‑risk individuals from routine ECGs, enabling earlier lifestyle, biomarker, and therapy interventions.
Key Findings
- ECG2HF achieved AUC 0.86 (MGH), 0.85 (BWH), and 0.84 (BIDMC) for 10-year HF prediction.
- Improved discrimination and net reclassification versus the Pooled Cohorts Equations to Prevent HF (AUC gain up to 0.061; NRI 0.16–0.23).
- Model is publicly available and trained on raw 12-lead ECG waveforms, enabling scalable deployment.
Methodological Strengths
- Large development cohort with multi-institution external validation and consistent performance.
- Validated EHR-based NLP outcome adjudication and rigorous risk reclassification analyses.
Limitations
- Observational design with potential site and demographic biases from three Northeastern US systems.
- No interventional testing of whether ECG2HF-guided care reduces HF events.
Future Directions: Prospective impact and implementation trials, fairness and domain‑shift assessments, and integration into EHR workflows with decision support to test clinical utility.
BACKGROUND: ECG-based artificial intelligence may enable efficient prediction of incident heart failure (HF) risk to facilitate preventive efforts. Prior models are proprietary, with modest or inconsistent accuracy. We sought to develop and validate a generalizable and publicly available convolutional neural network to predict incident HF using the 12-lead ECG waveform (ECG-to-HF [ECG2HF]). METHODS: We developed ECG2HF in 94 636 patients receiving longitudinal ambulatory care at Massachusetts General Hospital (MGH), and validated it in 3 test sets: MGH, Brigham and Women's Hospital (BWH), and Beth Israel Deaconess Medical Center (BIDMC), among 93 868 individuals aged 30 to 79 years without HF. HF events at 10 years were identified using a validated electronic health record-based natural language processing model. Discrimination was quantified using the area under the receiver operating characteristic curve. We then compared discrimination and net reclassification (at <10%, 10% to 20%, ≥20% 10-year risk categories) using ECG2HF versus the 15-component Pooled Cohorts Equations to Prevent HF score. RESULTS: The test sets comprised MGH (13 954 individuals, 441 events, age 57±13 years, 48% women), BWH (54 396 individuals, 1809 events, age 57±13 years, 55% women), and BIDMC (25 457 individuals, 901 events, age 57±13 years, 53% women). Over 10 years, the cumulative risk of HF was 4.6% (95% CI, 4.1-5.0) in MGH, 5.0% (4.8-5.2) in BWH, and 4.4% (4.1-4.7) in BIDMC. ECG2HF discriminated 10-year incident HF in each test set (area under the receiver operating characteristic curve: MGH 0.86 [0.84-0.87]; BWH 0.85 [0.84-0.86]; BIDMC 0.84 [0.83-0.86]). Compared with the Pooled Cohorts Equations to Prevent HF, ECG2HF provided favorable discrimination (improvement in area under the receiver operating characteristic curve MGH/BWH 0.061 [0.025-0.097]; BIDMC 0.038 [-0.0096 to 0.086]) and net reclassification (NRI MGH/BWH 0.16 [0.077-0.24]; BIDMC 0.23 [0.10-0.35]) of 10-year HF risk. CONCLUSIONS: ECG2HF is a publicly available 12-lead ECG-based artificial intelligence model that discriminates the risk of future HF with favorable and consistent performance across 3 large health care samples from the northeastern United States. ECG2HF may enable efficient prioritization of high-risk individuals for HF-related preventive measures.
2. A reduced TBX5-dependent gene regulatory network links atrial fibrillation and heart failure.
Comparative atrial multi-omic profiling across AF and HF models shows coordinated disruption of a TBX5‑driven gene regulatory network, including loss of Klf15, and emergence of a Sox9‑centered fibroblast network. Atrial TBX5 is downregulated in mouse and human HF, suggesting a shared genomic injury response linking AF and HF.
Impact: Defines a unifying atrial gene regulatory mechanism connecting AF and HF, highlighting TBX5 as a nodal regulator and revealing fibroblast Sox9 networks as potential targets.
Clinical Implications: Although preclinical, findings nominate TBX5‑KLF15 and fibroblast SOX9 circuits as candidate therapeutic axes and may inform biomarker development for atrial remodeling in AF–HF overlap.
Key Findings
- Atrial TBX5 expression is downregulated in mouse and human heart failure.
- Over 100 transcription factors are coordinately dysregulated in atria of Tbx5 cKO (AF) and TAC (HF) models.
- Wild-type TBX5-driven atrial GRN (including Klf15) is disrupted, while a Sox9-centered fibroblast network emerges.
Methodological Strengths
- Cross-model, cross-species integration (mouse AF and HF models with human data).
- Network-level analysis of atrial transcriptional and genomic regulation identifying nodal factors.
Limitations
- Preclinical models without interventional rescue demonstrating reversibility in vivo.
- Sample sizes and specific cellular contributions may vary across models; limited immediate clinical translation.
Future Directions: Test TBX5/KLF15 restoration and SOX9 inhibition in vivo, map cell‑state transitions longitudinally, and assess translatability via patient‑derived tissues and biomarkers.
Atrial fibrillation (AF) and heart failure (HF) frequently coexist and worsen one another's outcomes. To investigate shared molecular mechanisms, we compared atrial gene regulatory networks (GRNs) in the mouse Tbx5 conditional knockout (Tbx5 cKO) AF model and the transverse aortic constriction (TAC) HF model. Here we show highly correlated changes in atrial transcriptional and genomic profiles, including downregulated atrial Tbx5 expression in both mouse and human HF. More than 100 transcription factor genes were coordinately dysregulated in the atria of the Tbx5 cKO and TAC models. The wild-type atrial TBX5-driven GRN, including Klf15, a repressor of cardiomyocyte hypertrophy, was disrupted in Tbx5 cKO and TAC models. Conversely, a disease-specific network featuring Sox9 emerged in activated fibroblasts of Tbx5 cKO and TAC models. Our results identify coordinated disruption of TBX5-dependent atrial gene regulation in AF and HF, suggesting that a shared genomic injury response may underlie the reciprocal risk between these conditions.
3. Sex-specific associations between frailty and long-term outcomes in patients with acute myocardial infarction: a national population-based study.
In 931,133 AMI patients, women had more frailty, yet severe frailty conferred a 26% greater adjusted 1‑year mortality impact in men than women. Men received more intensive therapies across frailty strata, indicating that sex‑specific biology and care patterns both likely contribute to outcome disparities.
Impact: Largest national analysis to date quantifying sex differences in frailty’s prognostic effect after AMI, challenging current risk models and informing sex‑specific pathways.
Clinical Implications: Integrate frailty screening into AMI pathways for both sexes, but recognize that frailty portends higher adjusted mortality in men—guiding escalation, rehab, and follow‑up intensity.
Key Findings
- Among 931,133 AMI patients, frailty (including severe) was more prevalent in women.
- After multivariable adjustment, severe frailty’s association with 1‑year mortality was 26% greater in men (rHR 1.26; P‑interaction <0.001).
- Men received more intensive therapies at every frailty level, indicating care pattern differences alongside biological factors.
Methodological Strengths
- Nationwide linked registries (MINAP with admissions and mortality) with very large sample size.
- Robust multivariable modeling with interaction testing and absolute risk differences.
Limitations
- Observational design with potential residual confounding and frailty misclassification using administrative indices.
- Therapy intensity differences may reflect unmeasured clinical nuance (e.g., contraindications, preferences).
Future Directions: Prospective trials of sex‑informed frailty pathways in AMI, evaluation of targeted rehab and social support, and biomarker studies to disentangle biology vs. care effects.
BACKGROUND: Frailty and female sex are both recognised independent predictors of adverse outcomes after acute myocardial infarction (AMI). While females presenting with AMI are known to have a higher burden of frailty than males, it is unknown whether this fully explains sex-based disparities in outcomes, or if the prognostic impact of frailty itself differs between the sexes. METHODS: We conducted a retrospective national cohort study using data from the Myocardial Ischaemia National Audit Project (MINAP), linked to hospital admission and mortality registries in England and Wales between 2005 and 2019. Frailty was assessed using the Secondary Care Administrative Records Frailty (SCARF) index and categorised as fit, mild, moderate, or severe. Multivariable Cox proportional hazards models were used with a primary outcome of all-cause mortality at 1-year. FINDINGS: Of 931,133 patients with AMI, 317,967 (34.1%) were female. Frailty was more prevalent in females than in males (severe frailty: 53,065 [16.7%] vs. 64,018 [10.4%]). Males received more intensive therapeutic care across all frailty levels. After multivariable adjustment, the relationship between severe frailty and 1-year all-cause mortality was 26% greater in males than in females (relative hazard ratio [rHR]: 1.26, 95% CI 1.19-1.32, P-interaction <0.001). This corresponded to an adjusted absolute risk difference of 1.19% (95% CI 0.58%-1.79%). INTERPRETATION: In this national AMI cohort, while frailty was more prevalent in females, its association with 1-year mortality was significantly greater in males. This sex-specific effect of frailty challenges current risk-assessment paradigms and underscores the need for sex-informed care pathways. FUNDING: National Institute for Health and Care Research and British Heart Foundation Centre of Research Excellence, Leicester.