Objective: Electrical impedance tomography (EIT) shows potential for radiation-free and noninvasive hemodynamic monitoring. However, many factors degrade the accuracy and repeatability of these measurements. Our goal is to estimate the impact of this variability on the EIT-based monitoring of two important central hemodynamic parameters: stroke volume (SV) and pulmonary artery pressure (PAP). Approach: We performed simulations on a 4D () bioimpedance model of a human volunteer to study the influence of four potential confounding factors (electrode belt displacement, electrode detachment, changes in hematocrit and lung air volume) on the performance of EIT-based SV and PAP estimation. Results were used to estimate how these factors affect the EIT measures of either absolute values or relative changes (i.e. trending). Main results: Our findings reveal that the absolute measurement of SV via EIT is very sensitive to electrode belt displacements and lung conductivity changes. Nonetheless, the trending ability of SV EIT might be a promising alternative. The timing-based measurement of PAP is more robust to lung conductivity changes but sensitive to longitudinal belt displacements at severe hypertensive levels and to rotational displacements (independent of the PAP level). Significance: We identify and quantify the challenges of EIT-based SV and PAP monitoring. Absolute SV via EIT is challenging, but trending is feasible, while both the absolute and trending of PAP via EIT are mostly impaired by belt displacements.

Additional Metadata
Keywords bioimpedance model, cardiac output (CO), electrical impedance tomography (EIT), pulmonary artery pressure (PAP), simulations, stroke volume (SV)
Persistent URL dx.doi.org/10.1088/1361-6579/aa9828
Journal Physiological Measurement
Braun, F. (Fabian), Proença, M. (Martin), Lemay, M. (Mathieu), Bertschi, M. (Mattia), Adler, A, Thiran, J.-P. (Jean-Philippe), & Solà, J. (Josep). (2018). Limitations and challenges of EIT-based monitoring of stroke volume and pulmonary artery pressure. Physiological Measurement, 39(1). doi:10.1088/1361-6579/aa9828