Accuracy of electrical impedance tomography to detect perfusion defects in pulmonary embolism

Timely detection of pulmonary embolism (PE) is crucial, particularly in critical settings where rapid confirmation or exclusion is required and computerized tomography pulmonary angiography (CTPA) poses safety challenges. We assessed the experimental and clinical accuracy of electrical impedance tomography (EIT)-ventilation-perfusion (V̇/Q̇) maps and a novel wasted-ventilation index for detecting PE.

Ten piglets underwent EIT-V̇/Q̇ mapping before and after proximal or distal pulmonary artery occlusions. EIT-perfusion maps were validated against dynamic contrast-enhanced CT (DCE-CT) and quantitative clot-burden analysis from whole-lung CTPA. To assess specificity, models of non-occlusive perfusion impairment were added (6 piglets). The wasted-ventilation index was refined across 114 piglet conditions (66 PE) and subsequently validated in 66 patients with acute respiratory failure (257 exams) and 10 patients with chronic thromboembolic disease (31 exams), totaling 288 EIT-exams.

Strong positive correlations between estimates of regional perfusion obtained by EIT vs. DCE-CT or CTPA were found. Agreement showed mean bias of - 3.04 ± 3.02% between EIT and DCE-CT, and + 3.45 ± 2.81% between EIT and CTPA. The wasted-ventilation index performed with AUC = 0.989 in animals (P < 0.0001; sensitivity 96%, specificity 94%) and AUC = 0.923 in patients (P < 0.001; sensitivity 81%, specificity 94%). The index consistently decreased after thrombolysis.

The novel wasted-ventilation index showed high accuracy in detecting pulmonary arterial occlusions of varying severity, distinguishing them from other perfusion problems. Its agreement with CTPA, DCE-CT, and reference methods supports EIT-V̇/Q̇ mapping as a reliable diagnostic aid when conventional imaging is unfeasible or risky.