Artificial Intelligence Predicts Hospitalization for Acute Heart Failure Exacerbation in Patients Undergoing...

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Imaging

Background:

• Heart failure (HF) is a major cause of morbidity, mortality, and economic burden globally.
• Accurate prediction of HF exacerbations is crucial for timely intervention and improved patient management.
• Current risk prediction models have limitations in accurately forecasting HF hospitalizations.

Purpose of the Study:

• To develop and validate an artificial intelligence (AI) model for predicting acute heart failure (HF) exacerbation hospitalizations.
• To assess the added value of incorporating SPECT/CT myocardial perfusion imaging parameters into AI risk prediction.
• To compare the predictive performance of the AI model against traditional clinical and stress test parameters.

Main Methods:

• Developed an AI risk prediction model using data from 4,766 patients undergoing SPECT/CT (internal cohort).
• The model integrated clinical factors, stress variables, SPECT imaging, and deep learning-derived calcium scores.
• Validated the model internally using 10-fold cross-validation and externally on a separate cohort of 2,912 patients.

Main Results:

• The AI model demonstrated superior prediction of HF admissions with an AUC of 0.87 ± 0.03 in the internal cohort.
• This outperformed stress left ventricular ejection fraction (AUC 0.73 ± 0.05) and a clinical-only model (AUC 0.81 ± 0.04).
• External validation confirmed the AI model's robust performance (AUC 0.80 ± 0.04).

Conclusions:

• Integrating SPECT myocardial perfusion imaging with AI significantly enhances the prediction of HF hospitalizations.
• The proposed AI-based risk assessment enables early interventions to prevent hospital admissions.
• This approach holds potential for improving patient care and outcomes in heart failure management.