Computer Methods and Programs in Biomedicine Update
Noninvasive hemodynamic estimation in patients with heart failure using electrocardiograms, pulse waves, and heart sounds
By Yuichiro Iida, Takayuki Uchida, Toru Fujii and 3 others.
Abstract
Background
Right heart catheterization, the gold standard for hemodynamic measurement in patients with heart failure, is invasive and performed only when its benefits outweigh the risks. While there have been attempts to noninvasively estimate hemodynamic parameters, a simple and accurate method for estimating multiple parameters has not yet been established.
Objectives
This study aimed to accurately estimate multiple hemodynamic parameters using noninvasive measurements.
Methods
Fifty patients with heart failure, scheduled for right heart catheterization, underwent simultaneous recording of ECG, pulse wave, and heart sound signals. Features were extracted from the three noninvasive waveforms and used in machine learning models to estimate multiple hemodynamic parameters.
