Integrated Technology of Internal Electrode-Assisted Pulmonary Ventilation Electrical Impedance Tomography and Surgical Positioning

Inspired by clinical demands and the limitations of current monitoring and navigation systems in precision medicine, this study addresses key needs in pulmonary interventional surgery. These include real-time, high-quality lung ventilation imaging and accurate localization of distal airways and deep pulmonary lesions. To meet these needs, we derived a mathematical-physical model based on complex impedance distribution for ventilation imaging and surgical localization. The mechanisms through which the real and imaginary components of impedance function in pulmonary ventilation imaging and internal electrode localization were systematically analyzed. Building upon this foundation, a multi-task deep learning algorithm, PI-DualET, was proposed, integrating shared feature extraction layers with task-specific decoders to simultaneously process real and imaginary complex impedance measurements, jointly outputting lung ventilation images and three-dimensional internal electrode coordinates. A corresponding hardware platform was developed and validated through comparative analysis of simulations and physical phantom experiments. Results demonstrated excellent performance in both ventilation imaging and localization, with a Relative Error (RE) of 0.0261, Correlation Coefficient (CC) of 0.9856, Structural Similarity Index (SSIM) of 0.9799, and an average Center Point Error (CPE) of 1.016 mm. Subsequent in-vivo porcine experiments further confirmed the clinical feasibility of this approach. This integrated monitoring and navigation system greatly enhances information acquisition during lung interventional procedures, offering a promising technical solution for precision interventional pulmonology.