Optimal flow adjustment of veno‐venoarterial extracorporeal membrane oxygenation with an adaptive prediction model: cannula sizes screening and pump speeds estimation

Abstract

Veno‐venoarterial extracorporeal membrane oxygenation is an upgraded configuration that provides well‐oxygenated blood to the systemic and pulmonary circulations. Sufficient sized cannulas, gas exchange capability, and other equipment should be selected to fulfil the desired extracorporeal supporting flow for temporary life support. Therefore, an accurate prediction model is used to screen the optimal combinations of cannula sizes and determine adequate extracorporeal flow for gas exchange. Prediction models use a substitution‐rate matrix to choose optimal combinations under hemodynamic stability, and a generalised regression neural network to estimate the desired pump speed required to maintain sufficient flow. Experimental studies are used to validate the proposed model and suggest further modifications in the base/reference for the intensive care unit.