The influence of image reconstruction algorithms on linear thorax EIT image analysis of ventilation

作者：Zhao Zhanqi^*; Frerichs Inez; Pulletz Sven; Mueller Lisse Ullrich; Moeller Knut

来源：Physiological Measurement, 2014, 35(6): 1083-1093.

DOI：10.1088/0967-3334/35/6/1083

• electrical impedance tomography
• image analysis
• mechanical ventilation
• ventilation distribution
• computed tomography

Analysis methods of electrical impedance tomography (EIT) images based on different reconstruction algorithms were examined. EIT measurements were performed on eight mechanically ventilated patients with acute respiratory distress syndrome. A maneuver with step increase of airway pressure was performed. EIT raw data were reconstructed offline with (1) filtered back-projection (BP); (2) the Drager algorithm based on linearized Newton-Raphson (DR); (3) the GREIT (Graz consensus reconstruction algorithm for EIT) reconstruction algorithm with a circular forward model (GRC) and (4) GREIT with individual thorax geometry (GR(T)). Individual thorax contours were automatically determined from the routine computed tomography images. Five indices were calculated on the resulting EIT images respectively: (a) the ratio between tidal and deep inflation impedance changes; (b) tidal impedance changes in the right and left lungs; (c) center of gravity; (d) the global inhomogeneity index and (e) ventilation delay at mid-dorsal regions. No significant differences were found in all examined indices among the four reconstruction algorithms (p > 0.2, Kruskal-Wallis test). The examined algorithms used for EIT image reconstruction do not influence the selected indices derived from the EIT image analysis. Indices that validated for images with one reconstruction algorithm are also valid for other reconstruction algorithms.

• 出版日期2014-6