Individual thorax geometry reduces position and size differences in reconstructed images of electrical impedance tomography

Article type: Research Article

Authors: Zhao, Zhanqi^; | Frerichs, Inéz | Pulletz, Sven | Müller-Lisse, Ullrich | Möller, Knut

Affiliations: Institute of Technical Medicine, Furtwangen University, VS-Schwenningen, Germany | Department of Anaesthesiology and Intensive Care Medicine, University Medical Center Schleswig-Holstein, Campus Kiel, Kiel, Germany | Department of Anaesthesiology and Intensive Care Medicine, Medical Center Osnabrück, Osnabrück, Germany | Department of Radiology, University of Munich, Munich, Germany

Note: [] Corresponding author: Zhanqi Zhao, Furtwangen University, Jakob-Kienzle-Straße 17, 78054 Villingen-Schwenningen, Germany. Tel.: +49 0 7720 307 4395; Fax: +49 0 7720 307 4210; E-mail: zhanqi.zhao@hs-furtwangen.de

Abstract: BACKGROUND: Due to the ill-posed problem, the electrical impedance within the thorax cannot be exactly reconstructed. OBJECTIVE: The aim of our study was to prove that reconstruction with individual thorax geometry improved the quality of EIT (electrical impedance tomography) images. METHODS: Seven mechanically ventilated patients with acute respiratory distress syndrome were examined by EIT. The thorax contours were determined from routine computed tomography (CT) images based on automatic threshold filtering. EIT raw data was reconstructed offline with (1) back-projection with circular forward model; (2) GREIT reconstruction method with circular forward model and (3) GREIT with individual thorax geometry. The resulting EIT images were compared to rescaled CT images. The distance between the lung contour and the thorax contour was calculated for each method and the differences to that in CT were denoted as position differences. Shape differences was defined as the ratio of thorax (or lungs) size in EIT and that in rescaled CT. RESULTS: Method (3) has the smallest position differences (6.6 ± 2.8, 5.3 ± 3.3, 2.3 ± 1.4 in pixel, for each reconstruction method respectively; mean ± SD). The thorax and lungs sizes in the transformed CT images were 514 ± 73 and 177 ± 39. Shape differences of thorax were 1.81 ± 0.26, 1.81 ± 0.26, 1.10 ± 0.12 and that of lungs were 1.69 ± 0.45, 1.52 ± 0.45, 1.34 ± 0.35 for each method respectively. CONCLUSION: The reconstructed images using the GREIT method with individual thorax geometry were more realistic. Improvement of EIT image quality may foster the acceptance of EIT in routine clinical use.

Keywords: Computed tomography, electrical impedance tomography, image reconstruction, individual thorax geometry

DOI: 10.3233/XST-140464

Journal: Journal of X-Ray Science and Technology, vol. 22, no. 6, pp. 797-807, 2014