Parametric EIT for monitoring cardiac stroke volume

S. Zlochiver*, D. Freimark, M. Arad, A. Adunsky, S. Abboud

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

47 Scopus citations

Abstract

The bio-impedance technique appears appropriate for non-invasive cardiac stroke volume (SV) measurement, as the thoracic conductivity distribution is altered during the cardiac cycle due to the heart contraction and blood perfusion. In the present work, the feasibility of a parametric electrical impedance tomography (EIT) for assessing the cardiac SV was studied. An impedance model of the thorax was constructed from segmented axial MRI images along 19 phases of the cardiac cycle. The heart was simulated as an ellipsoid, with its axes' lengths set as the reconstruction parameters, while all other tissues' geometry and conductivity values were kept fixed. A Newton-Raphson parametric optimization scheme was utilized, yielding a correlation between the reconstructed and anatomical left ventricular volumes of 0.97 (p = 2 × 10-11). An analysis of noise sensitivity showed that the proposed algorithm requires an SNR greater than 65 dB. The simulation results were compared to physical data, collected with a portable EIT system (PulmoTrace™, CardioInspect). The validation study was employed for a group of N = 28 healthy patients, and a comparison with impedance cardiography measurements (BioZ®, Cardiodynamics) was made, showing a correlation of r = 0.86 (p = 4 ×10-9). The preliminary results demonstrate that parametric EIT has the potential to measure SV, and may be applicable for both clinical and home environment usage.

Original languageEnglish
Article numberS12
Pages (from-to)S139-S146
JournalPhysiological Measurement
Volume27
Issue number5
DOIs
StatePublished - 1 May 2006

Keywords

  • Clinical trial
  • Parametric reconstruction
  • SV measurement
  • Simulation study

Fingerprint

Dive into the research topics of 'Parametric EIT for monitoring cardiac stroke volume'. Together they form a unique fingerprint.

Cite this