Predicting Long-Term Mortality in Patients With Acute Heart Failure by Using Machine Learning.

Area of Science:

• Cardiology
• Medical Informatics
• Machine Learning

Background:

• High mortality rates in acute heart failure (AHF) necessitate improved risk stratification.
• Current risk-assessment tools for long-term mortality in AHF patients are limited.

Purpose of the Study:

• To develop and validate a machine learning (ML)-based model for predicting long-term all-cause mortality in patients hospitalized for AHF.

Main Methods:

• A CoxBoost algorithm-based ML model was developed using clinical and echocardiographic data from 2704 AHF patients.
• The model was validated in an independent cohort of 1608 patients and compared against BIOSTAT-CHF and AHEAD scores.

Main Results:

• The ML model achieved an AUROC of 0.761 (training) and 0.760 (test) for 3-year mortality.
• It significantly outperformed existing risk scores (BIOSTAT-CHF AUROC 0.714, AHEAD AUROC 0.681).
• The model identified high-risk patients irrespective of heart failure phenotypes.

Conclusions:

• The developed ML model provides accurate long-term mortality prediction for AHF patients.
• This tool enables optimal risk stratification, potentially improving clinical decision-making and patient outcomes.