Related Experiment Video
Updated: Jun 16, 2026
Identifying Coronary Artery Calcification on Non-gated Computed Tomography Scans
Published on: August 28, 2018
AI-CVD-HF: A heart failure risk prediction model based on coronary artery calcium scans compared with PREVENT-HF
Seyed Reza Mirjalili1, Kyle Atlas1, Anthony P Reeves2
1HeartLung.AI, Houston, TX, 77021, USA.
Insights
A new AI-CVD-HF model uses artificial intelligence (AI) features from coronary artery calcium (CAC) scans to predict heart failure (HF) risk. This AI-CVD-HF model shows improved accuracy and calibration compared to the PREVENT-HF score.
Area of Science:
- Cardiology
- Artificial Intelligence in Medicine
- Medical Imaging Analysis
Background:
- The AI-CVD initiative aims to extract comprehensive information from coronary artery calcium (CAC) scans beyond traditional CAC scores.
- Current methods for heart failure (HF) risk prediction have limitations, necessitating novel approaches.
Purpose of the Study:
- To develop and validate a novel heart failure (HF) prediction model, termed AI-CVD-HF.
- To leverage artificial intelligence (AI) derived features from non-contrast CAC scans for HF risk stratification.
- To compare the performance of the AI-CVD-HF model against the established PREVENT-HF prediction tool.
Main Methods:
- AI-derived features were extracted from CAC scans of 6743 asymptomatic participants from the MESA and FHS-O studies.
- Random forest and random survival forest models were employed for feature selection and HF risk prediction.
- Model performance was evaluated using 5-fold cross-validation, comparing area under the receiver-operating-characteristic curve (AUC), area under the precision-recall curve (AUPRC), net-benefit, and calibration metrics against PREVENT-HF.
- External validation was performed to assess the generalizability of the AI-CVD-HF model.
Main Results:
- The AI-CVD-HF model incorporated AI-CAC score, ventricular volume ratio, atrial volume, ventricular mass, visceral fat volume, skeletal muscle density, thoracic aortic calcification, age, and sex.
- AI-CVD-HF achieved a 10-year AUC of 0.83, outperforming PREVENT-HF (0.79, p=0.01).
- The model demonstrated a 32% higher AUPRC (0.25 vs 0.19, p=0.049) and superior calibration (Brier score, slope) compared to PREVENT-HF.
- Performance remained consistent across diverse demographic groups and in external validation.
Conclusions:
- AI-CVD-HF, utilizing AI features from CAC scans, offers superior discrimination and calibration for heart failure prediction compared to PREVENT-HF.
- This study highlights the expanded utility of CAC scans for cardiovascular risk assessment beyond coronary artery disease.
- The AI-CVD-HF model represents a significant advancement in leveraging medical imaging for proactive HF risk management.
Background:
The AI-CVD initiative seeks to extract actionable information from coronary artery calcium (CAC) scans beyond the CAC score. We aimed to develop a heart failure (HF) prediction model, AI-CVD-HF, based on AI-derived features from non-contrast CAC scans, and compare it with PREVENT-HF.
Method:
AI features from CAC scans of 6743 asymptomatic participants in the Multi-Ethnic Study of Atherosclerosis (MESA) and Framingham Heart Study-Offspring (FHS-O) (mean age 62.3 ± 10.1; 47.2% male; median follow-up 17.1 years; 429 HF events) were analyzed. Features were selected using random forest and modeled with random survival forest. Performance was assessed against the base PREVENT-HF using 5-fold cross-validation for area under the receiver-operating-characteristic curve (AUC), area under the precision-recall curve (AUPRC), net-benefit, and calibration. External validation experiment was conducted to assess generalizability.
Results:
Selected features for AI-CVD-HF model included AI-CAC score, left-to-right ventricular volume ratio, left atrial volume, left ventricular mass, visceral fat volume, skeletal muscle mean density, thoracic aortic calcification, age and sex. The 10-year AUC of AI-CVD-HF was 0.83 (95% CI:0.81-0.85), compared with PREVENT-HF (0.79, 95% CI:0.75-0.83; p = 0.01) and also demonstrated 32% higher AUPRC (0.25[95% CI:0.11-0.39] vs 0.19[95% CI:0.07-0.31]; p = 0.049), and consistent performance across age, sex, and race/ethnicity. Calibration metrics favored AI-CVD-HF over PREVENT-HF (Brier score [0.0356 vs 0.0383], slope [1.02 vs 1.20]). External validation showed performance and calibration consistent with internal validation.
Conclusions:
AI-CVD-HF, using AI-derived features from CAC scans, demonstrated more favorable discrimination and calibration than PREVENT-HF for HF prediction, extending the utility of CAC scans beyond coronary artery disease risk assessment.
Related Concept Videos
01:25Imaging Studies for Cardiovascular System VI: Calcium -Scoring CT
01:28Imaging Studies for Cardiovascular System V: CT
01:26Coronary Artery Disease IV: Preventive Measures
01:20Imaging Studies for Cardiovascular System III: X-Ray
Definition and Purpose
An X-ray, or radiograph, is a non-invasive method that uses ionizing radiation to take images of internal structures. It is mainly used in cardiac imaging to examine the heart, lungs, and major blood vessels, aiming to identify abnormalities in the heart's size, shape, and position, such as heart failure, congenital defects, and vascular...
01:30Heart Failure IV: Classification and Diagnostic Evaluation
01:27Heart Failure I: Introduction