A Machine Learning Approach to Non-Exercise Prediction of Peak VO 2 in Advanced Heart Failure

Announcing a new article publication for Cardiovascular Innovations and Applications journal. Peak oxygen consumption (VO₂), assessed via cardiopulmonary exercise testing (CPET), is central to heart transplant evaluation in advanced heart failure. However, many patients have limitations that prevent them from completing CPET. This study was aimed at developing machine learning 524 heart transplant candidates (ML) models to estimate peak VO₂ by using non-exercise clinical data who underwent comprehensive clinical, laboratory, echocardiographic, and hemodynamic assessment were retrospectively analyzed.

The Boruta algorithm selected key predictors among 55 variables. Multiple ML algorithms were trained to estimate peak VO₂, and performance was assessed according to root mean square error (RMSE), mean absolute error (MAE), and R². Feature importance was evaluated with SHAP values.

Models were developed with two sets of variables: a full model (17 features) and a simplified model (seven features). Extreme Gradient Boosting performed best overall (RMSE: 2.44, MAE: 1.89, R²: 0.689). In the simplified model, gradient boosted decision trees were optimal (RMSE: 3.20, MAE: 2.54, R²: 0.481). Key predictors included TAPSE/PASP, mixed venous saturation, proBNP, age, hemoglobin, pulmonary vascular resistance, and BMI.

ML-based models with non-exercise data can estimate peak VO₂ with reasonable accuracy, thus offering a practical tool for evaluating transplant candidates unable to undergo CPET.

Read More: https://www.scienceopen.com/hosted-document?doi=10.15212/CVIA.2025.0024

CVIA is available on the ScienceOpen platform and at Cardiovascular Innovations and Applications. Submissions may be made using ScholarOne Manuscripts. There are no author submission or article processing fees. Cardiovascular Innovations and Applications is indexed in the EMBASE, EBSCO, ESCI, OCLC, Primo Central (Ex Libris), Sherpa Romeo, NISC (National Information Services Corporation), DOAJ, Index Copernicus, Research4Life and Ulrich’s web Databases. Follow CVIA on Twitter @CVIA_Journal; or Facebook.

Barkin Kultursay, Murat Karacam and Seda Tanyeri Uzel et al. A Machine Learning Approach to Non-Exercise Prediction of Peak VO2 in Advanced Heart Failure. CVIA. 2025. Vol. 10(1). DOI: 10.15212/CVIA.2025.0024