Predicting cardiopulmonary exercise testing outcomes in congenital heart disease through multimodal data integration and geometric learning

TL;DR

This study uses machine learning to predict exercise test results for heart disease patients by combining ECG data and clinical notes.

Contribution

First successful use of geometric learning to integrate ECGs and clinical text for predicting CPET outcomes in CHD.

Findings

Combining ECGs and clinical letters improves CPET outcome prediction.

Riemannian geometry-based models outperform traditional methods in predictive accuracy.

Covariance augmentation enhances model robustness for CHD patient data.

Abstract

Cardiopulmonary exercise testing (CPET) provides a comprehensive assessment of functional capacity by measuring key physiological variables including oxygen consumption (\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$VO_2$$\end{document}), carbon dioxide production (\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$VCO_2$$\end{document}), and pulmonary ventilation (VE) during exercise. Previous research has identified peak \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$VO_2$$\end{document} and \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$VE/VCO_2$$\end{document} ratio as robust predictors of mortality risk in chronic heart failure (CHF) patients as well as in congenital heart disease (CHD). This study utilises CPET variables as surrogate mortality endpoints for patients with CHD. To our knowledge, this represents the first successful implementation of an advanced machine learning approach that predicts CPET outcomes by integrating electrocardiograms (ECGs) with information derived from clinical letters. Our methodology began with extracting unstructured patient information from clinical letters using natural language processing techniques, organising this data into a structured database. We then digitised ECGs to obtain quantifiable waveforms and established comprehensive data linkages. The core innovation of our approach lies in exploiting the Riemannian geometric properties of covariance matrices derived from both 12-lead ECGs and clinical text data to develop robust regression and classification models. Through extensive ablation studies, we demonstrated that the integration of ECG signals with clinical documentation, enhanced by covariance augmentation techniques in Riemannian space, consistently produced superior predictive performance compared to conventional approaches.

Figures