Regional lung ventilation and perfusion by electrical impedance tomography compared to single-photon emission computed tomography

Objective: Electrical impedance tomography (EIT) is a noninvasive imaging modality that allows real-time monitoring of regional lung ventilation (<(V)over dot>) in intensive care patients at bedside. However, for improved guidance of ventilation therapy it would be beneficial to obtain regional ventilation-to-perfusion ratio (<(V)over dot> / <(Q)over dot>) by EIT. Approach: In order to further explore the feasibility, we first evaluate a model-based approach, based on semi-negative matrix factorization and a gamma-variate model, to extract regional lung perfusion (<(Q)over dot>) from EIT measurements. Subsequently, a combined validation of both <(V)over dot> and <(Q)over dot> measured by EIT against single-photon emission computed tomography (SPECT) is performed on data acquired as part of a porcine animal trial. Four pigs were ventilated at two different levels of positive end-expiratory pressure (PEEP 0 and 15 cm H2O, respectively) in randomized order. Repeated injections of an EIT contrast agent (NaCl 10%) and simultaneous SPECT measurements of <(V)over dot> (81(m)Kr gas) and <(Q)over dot> (99(m)Tc-labeled albumin) were performed. Main results: Both <(V)over dot> and <(Q)over dot> from EIT and SPECT were compared by correlation analysis. Very strong (r(2) = 0.94 to 0.95) correlations were found for <(V)over dot> and <(Q)over dot> in the dorsal-ventral direction at both PEEP levels. Moderate (r(2) = 0.36 to 0.46) and moderate to strong (r(2) = 0.61 to 0.82) correlations resulted for <(V)over dot> and <(Q)over dot> in the right-left direction, respectively. Significance: The results of combined validation indicate that monitoring of <(V)over dot> and <(Q)over dot> by EIT is possible. However, care should be taken when trying to quantify <(V)over dot> / <(Q)over dot> by EIT, as imaging artefacts and model bias may void necessary spatial matching.