Daily Cardiology Research Analysis
Three studies shift contemporary cardiology practice: an externally validated machine learning RESCUE score accurately predicts in-hospital mortality in cardiogenic shock; a composite CCTA score integrating Agatston, CAD-RADS, high-risk plaque features, and PCAT improves risk stratification in chronic coronary syndromes; and a MOMENTUM 3 analysis shows no clinical benefit of ICD/CRT-D in LVAD patients while increasing ventricular arrhythmias.
Summary
Three studies shift contemporary cardiology practice: an externally validated machine learning RESCUE score accurately predicts in-hospital mortality in cardiogenic shock; a composite CCTA score integrating Agatston, CAD-RADS, high-risk plaque features, and PCAT improves risk stratification in chronic coronary syndromes; and a MOMENTUM 3 analysis shows no clinical benefit of ICD/CRT-D in LVAD patients while increasing ventricular arrhythmias.
Research Themes
- Prognostic modeling and AI-driven risk stratification in acute cardiology
- Composite imaging biomarkers for chronic coronary syndromes
- Device therapy value reassessment in LVAD populations
Selected Articles
1. Machine learning prediction of in-hospital mortality and external validation in patients with cardiogenic shock: the RESCUE score.
Using four ML algorithms for feature selection and logistic regression for model building, the RESCUE score identified seven predictors and achieved AUC 0.86 (internal) and 0.80 (external) for in-hospital mortality in cardiogenic shock. The model generalized across etiologies and was validated in an independent 750-patient cohort.
Impact: Provides an externally validated, parsimonious risk score for a high-mortality population, enabling earlier triage and resource allocation decisions in cardiogenic shock.
Clinical Implications: RESCUE can inform MCS escalation, ICU triage, and timing of advanced therapies by quantifying mortality risk at presentation. Integration into care pathways and EHRs could standardize CS risk assessment.
Key Findings
- Seven predictors were selected: age, vasoactive inotropic score, LVEF, lactate, in-hospital cardiac arrest, need for CRRT, and mechanical ventilation.
- Model performance: AUC 0.86 (internal with 10-fold CV) and 0.80 (external validation in 750 patients).
- Applicable across all-cause cardiogenic shock, supporting generalizability.
Methodological Strengths
- Independent external validation with strong AUC.
- Transparent model (logistic regression) after ML-based feature selection and internal cross-validation.
Limitations
- Observational registry design with potential residual confounding and selection bias.
- Calibration and transportability to different health systems and care pathways were not fully explored.
Future Directions: Prospective, multi-regional impact studies assessing clinical decision support integration, calibration drift monitoring, and whether RESCUE-guided management improves outcomes.
INTRODUCTION AND OBJECTIVES: Despite advances in mechanical circulatory support, mortality rates in cardiogenic shock (CS) remain high. A reliable risk stratification system could serve as a valuable guide in the clinical management of patients with CS. This study aimed to develop and externally validate a risk prediction model for in-hospital mortality in CS patients using machine learning (ML) algorithms. METHODS: Data from 1247 patients with all-cause CS in the RESCUE registry (January 2014-December 2018) were analyzed. Key predictive variables were identified using 4 ML algorithms. A risk prediction model, the RESCUE score, was developed using logistic regression based on the selected variables. Internal validation was conducted within the RESCUE registry, and external validation was performed using an independent CS registry of 750 patients. RESULTS: The 4 ML models identified 7 predictors: age, vasoactive inotropic score, left ventricular ejection fraction, lactic acid level, in-hospital cardiac arrest at presentation, need for continuous renal replacement therapy, and mechanical ventilation. The RESCUE score demonstrated strong predictive performance, with an AUC of 0.86 (95%CI, 0.83-0.88) for in-hospital mortality. Ten-fold internal cross-validation yielded an AUC of 0.86 (95%CI, 0.77-0.95). External validation showed an AUC of 0.80 (95%CI, 0.76-0.84). CONCLUSIONS: Our ML-based risk-scoring system, the RESCUE score, demonstrated excellent predictive performance for in-hospital mortality in all patients with CS, regardless of cause. The system could be a useful and reliable tool to estimate risk stratification of CS in everyday clinical practice. CLINICAL TRIAL REGISTRATION: NCT02985008.
2. Implantable Cardioverter-Defibrillators and Cardiovascular Resynchronization Therapy with Left Ventricular Assist DevicesA MOMENTUM 3 Trial Analysis.
In HeartMate 3 LVAD recipients, having an ICD or CRT-D did not improve survival, rehospitalization, quality of life, or 6-minute walk distance at 2 years, but was associated with higher ventricular arrhythmia incidence. Findings were consistent in propensity-matched analyses.
Impact: Challenges routine ICD/CRT-D use in contemporary LVAD care by demonstrating lack of clinical benefit and potential harms, informing guideline updates and device programming.
Clinical Implications: Consider individualized ICD/CRT-D strategies in LVAD patients; emphasize careful programming, arrhythmia monitoring, and avoid default implantation when not indicated by pre-LVAD history.
Key Findings
- ICD/CRT-D showed no differences in survival, rehospitalization, quality of life, or functional capacity versus no device over 2 years.
- ICD/CRT-D presence was associated with increased ventricular arrhythmias (HR 2.4; 95% CI 1.3–4.3).
- CRT-D vs ICD alone: no survival advantage; higher ventricular arrhythmia rates (HR 1.3; 95% CI 1.0–1.7).
Methodological Strengths
- Large trial-derived cohort with standardized follow-up and propensity-matched sensitivity analyses.
- Multiple clinically relevant endpoints assessed over 2 years.
Limitations
- Post-hoc, non-randomized comparison of device status; residual confounding and device programming heterogeneity.
- Findings specific to HeartMate 3 era; generalizability to other LVAD platforms requires caution.
Future Directions: Prospective randomized or pragmatic trials testing ICD programming strategies or de-implementation pathways in LVAD recipients; subgroup analyses based on pre-LVAD arrhythmic history.
BACKGROUND: The benefit of implantable cardioverter-defibrillators (ICDs) and cardiovascular resynchronization therapy defibrillators (CRT-Ds) in patients supported with a HeartMate 3 left ventricular assist device (LVAD) remains uncertain. METHODS: An analysis was done of the Multicenter Study of MAGLEV Technology in Patients Undergoing Mechanical Circulatory Support Therapy with HeartMate 3 (MOMENTUM 3) randomized clinical trial and the first 1000 patients in the Continued Access Protocol (CAP) trial. Patients were divided into 3 groups based on the presence of an ICD and/or CRT-D: No device (n = 153, 11%), ICD only (n = 699, 50.4%), and CRT-D (n = 535, 38.6%). We assessed the association of ICDs or CRT-Ds with overall mortality, ventricular arrhythmias (VAs), rehospitalization rates, quality of life, and the 6-minute walk test distance at 2 years' follow-up. RESULTS: Patients with an ICD or CRT-D had similar survival to those without (hazard ratio [HR], 1.3; 95% CI 0.8-2.1, P = .36) with no differences in rehospitalizations, quality of life or 6-minute walk test distance. VA occurred more frequently in patients with an ICD or CRT-D (HR, 2.4; 95% CI 1.3-4.3, P = .006). Compared with an ICD alone, patients with a CRT-D demonstrated similar survival (HR, 1.1; 95% CI 0.9-1.5, P = .36). However, they had increased rates of VA (HR, 1.3; 95% CI 1.0-1.7, P = .03). There were no differences in rate of rehospitalization between those with an ICD or CRT-D and those without (P = .19) or between those with an ICD and those with a CRT-D (P = .32). A propensity-matched sensitivity analysis confirmed these findings. CONCLUSIONS: In this post-hoc analysis of the MOMENTUM 3 trial, the presence of an ICD or CRT-D at the time of HM3 LVAD implantation was associated with an increased incidence of VA but was not associated with survival, quality of life, or functional capacity. TRIAL REGISTRATION: Momentum 3 portfolio, NCT02224755 (Pivotal) and NCT02892955 (CAP).
3. Composite cardiac computed tomography angiography score for improved risk assessment in chronic coronary syndromes.
In 759 CCS patients, a composite CCTA score integrating Agatston, CAD-RADS, high-risk plaque count, and PCAT predicted a composite endpoint and outperformed individual CT metrics. The number of high-risk plaques was the strongest single predictor (HR 2.74).
Impact: Demonstrates additive value of combining widely available CCTA metrics to refine risk stratification, supporting practical imaging-based precision cardiology.
Clinical Implications: Composite CCTA scoring could identify high-risk CCS patients for intensified preventive therapy and closer follow-up, potentially guiding lipid-lowering targets and revascularization strategies.
Key Findings
- Composite CCTA score combining Agatston, CAD-RADS, high-risk plaque count, and PCAT predicted all-cause death/MI/revascularization better than individual metrics.
- High-risk plaque count per patient was the strongest single predictor (HR 2.74; 95% CI 1.56–4.80; p<0.001).
- Predictive value was independent of age and conventional risk factors in multivariable Cox models.
Methodological Strengths
- Systematic integration of routine CCTA metrics with prespecified composite endpoint.
- Appropriate multivariable Cox regression demonstrating independence from conventional risk factors.
Limitations
- Observational design with modest event rate (5.1%) and median follow-up of ~1.6 years.
- Potential center-specific imaging protocols and PCAT measurement variability.
Future Directions: Prospective multi-center validation, calibration across vendors, and testing whether composite CCTA-guided management improves outcomes.
Agatston score, the degree of lumen narrowing categorized by CAD-RADS, high-risk atherosclerotic plaque features and pericoronary adipose tissue attenuation (PCAT) are parameters, which can be assessed non-invasively by coronary computed tomography angiography (CCTA) and aid risk stratification in patients with chronic coronary syndromes (CCS). However, few studies have so far compared the prognostic value of all those parameters together. To develop and test the prognostic value of a composite CCTA score, derived from Agatston score, CAD-RADS, high-risk plaques and PCAT in patients undergoing CCTA due to CCS. Consecutive patients with clinical indication for CCTA and available clinical follow-up of ≥ 6 months after the CCTA examination were included. (i) Agatston score, (ii) CAD-RADS, (iii) the number of plaques with at least one high-risk feature and (iv) PCAT in the proximal 4 cm of the right coronary artery (RCA) were measured, and a composite CCTA score was generated considering all four parameters. The primary endpoint encompassed all-cause mortality, myocardial infarction, and coronary revascularization (> 60 days after the CCTA scan) during follow-up. In total, 759 patients (median age 68.0 (IQR 59.0-76.0) years, 352 (46.4%) female) were included. During a median follow-up of 591.5 (IQR 505.5-686.8) days, 39 (5.1%) patients reached the primary endpoint. Cox-proportional regression demonstrated that the Agatston score, the number of high-risk plaques and CAD-RADS predicted the primary endpoint, independent of age and conventional cardiovascular risk factors. The number of high-risk plaques per patient provided the most robust prediction of the primary endpoint (HR = 2.74, 95%CI = 1.56-4.80, p < 0.001), whereas the composite CCTA score outperformed all other parameters (HR = 1.54, 95%CI = 1.19-1.98, p < 0.001). The Agatston score, CAD-RADS and high-risk plaque features may provide complementary prognostic information in patients with CCS. A composite CCTA score, derived by these imaging markers may identify high-risk individuals, who may benefit from more intensified treatment and clinical follow-up in future studies.