Abstract Background Two-dimensional electrical impedance tomography (2D-EIT) is increasingly used at the bedside to monitor regional ventilation (V) and perfusion (Q). However, single-slice reconstruction may miss pathology localized outside the electrode plane. We hypothesised that three-dimensional EIT (3D-EIT) employing a dual-belt electrode configuration would improve the topographic accuracy of V/Q mismatch detection compared with conventional 2D-EIT. Methods In a prospective single-centre study (Zhongshan Hospital, Fudan University, February-June 2024) we enrolled 78 consecutive adults with known or suspected focal lung disease. Simultaneous recordings were obtained with: A custom 3D-EIT system (Infivision 1900, Huarui-Bio, China; 64 electrodes, two belts 4 cm apart) reconstructing a 12 × 12 × 8 cm³ lung volume;The belt supplying the second intercostal space exported as 2D-EIT data;A commercial 2D-EIT monitor (Dräger PulmoVista® 500, Germany) at the same level;Non-contrast thoracic CT acquired within 30 min. EIT images were segmented into six regions-of-interest (ROI, 3D) or four quadrants (2D). V and Q defects were defined as pixel values 20 % of the global maximum. Concordance with CT (reference) was quantified by Cohen’s κ and voxel-wise sensitivity/specificity. Results Seventy-eight patients (age 63 ± 11 y, BMI 24 ± 3 kg m−²) were analysed; CT diagnoses included malignancy with atelectasis (n = 34), pleural mesothelioma (n = 12), pneumonia/consolidation (n = 20) and mixed pathology (n = 12). Case 1 - Pleural mesothelioma, left lower lobe (LLL): 2D-EIT depicted complete V and Q loss in the entire left lung (false extrapolation); 3D-EIT precisely localised preserved Q in LLL despite absent V (V/Q 0), matching CT. Case 2 - Right lung tumour with middle-lobe collapse: 2D-EIT reported global right-sided V/Q absence; 3D-EIT revealed intact lateral-zone Q, enabling differentiation between atelectatic (zero V/Q) and merely compressed (low V, preserved Q) parenchyma. Case 3 - Right upper-lobe (RUL) mass: 2D-EIT showed right-sided V/Q loss; 3D-EIT confined the defect to RUL, consistent with CT. Pooled analysis Sensitivity for focal V-defect detection: 3D-EIT 0.92 (95 % CI 0.86-0.96) vs 2D-EIT 0.73 (0.64-0.81), P 0.001.Sensitivity for Q-defect: 0.88 vs 0.70, P = 0.002.Specificity remained 0.95 for both modalities. κ agreement with CT anatomical segment: 3D-EIT 0.81 (substantial) vs 2D-EIT 0.57 (moderate). Conclusions Compared with conventional 2D-EIT, 3D-EIT significantly improves the topographic accuracy of bedside V/Q mapping by capturing cranio-caudal heterogeneity missed by single-plane recordings. The enhanced spatial resolution may facilitate earlier recognition of region-specific pathology and guide targeted interventions in acute and chronic pulmonary diseases. This abstract is funded by: none
Wang et al. (2026) studied this question.