Neural Networks for Prognostication of Patients With Heart Failure
- 1 August 2018
- journal article
- research article
- Published by Ovid Technologies (Wolters Kluwer Health) in Circulation: Heart Failure
- Vol. 11 (8), e005193
- https://doi.org/10.1161/circheartfailure.118.005193
Abstract
Background: Prognostication of heart failure patients from cardiopulmonary exercise test (CPET) currently involves simplification of complex time-series data into summary indices. We hypothesized that prognostication could be improved by considering the totality of the data generated during a CPET, instead of using summary indices alone. Methods and Results: Complete data from 1156 CPETs were used to predict clinical deterioration (characterized by initiation of mechanical circulatory support, listing for heart transplantation or mortality) 1 year post-CPET. We compared the prognostic value (area under the receiver operating characteristic curve) of (1) the most predictive summary indices, (2) staged data collected at discrete intervals using multivariable regression models, and (3) breath-by-breath data using a feedforward neural network. The top-performing models were compared with the commonly used CPET risk score, using absolute net reclassification index. All models were trained and assessed using a 100-iteration Monte Carlo cross-validation. A total of 190 (16.4%) patients experienced clinical deterioration. The summary indices demonstrated subpar discriminative value (area under the receiver operating characteristic curve ≤0.800). Each multivariable model outperformed the summary indices, with the neural network incorporating breath-by-breath data achieving the best performance (area under the receiver operating characteristic curve =0.842). When compared with the CPET risk score (area under the receiver operating characteristic curve =0.759), the top-performing model obtained a net reclassification index of 4.9%. Conclusions: The current practice of considering summary indices in isolation fails to realize the full value of CPET data. This may lead to less accurate prognostication of patients and in consequence, inaccurate selection of patients for advanced therapy. Clinical practices, like CPET prognostication, must be continuously reevaluated to ensure optimal usage of valuable (and oft-underutilized) data sources.This publication has 29 references indexed in Scilit:
- The 2011 Canadian Cardiovascular Society Heart Failure Management Guidelines Update: Focus on Sleep Apnea, Renal Dysfunction, Mechanical Circulatory Support, and Palliative CareCanadian Journal of Cardiology, 2011
- Six-Minute Walk Test and Cardiopulmonary Exercise Testing in Patients With Chronic Heart FailureCirculation: Heart Failure, 2009
- Prognostic value of the VE/VCO2 slope calculated from different time intervals in patients with suspected heart failureInternational Journal of Cardiology, 2007
- Prognostic value of end-tidal carbon dioxide during exercise testing in heart failureInternational Journal of Cardiology, 2007
- Estimates of Life Expectancy by Adolescents and Young Adults With Congenital Heart DiseaseJournal of the American College of Cardiology, 2006
- Enhanced prognostic value from cardiopulmonary exercise testing in chronic heart failure by non-linear analysis: oxygen uptake efficiency slopeEuropean Heart Journal, 2005
- Peak VO2 and VE/VCO2 slope in patients with heart failure: a prognostic comparisonAmerican Heart Journal, 2004
- Heart-Rate Recovery Immediately after Exercise as a Predictor of MortalityThe New England Journal of Medicine, 1999
- Oxygen uptake efficiency slope: A new index of cardiorespiratory functional reserve derived from the relation between oxygen uptake and minute ventilation during incremental exerciseJournal of the American College of Cardiology, 1996
- Simplifying Neural Networks by Soft Weight-SharingNeural Computation, 1992