Uniform background assumption produces misleading lung EIT images.

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Physiological Monitoring

Background:

• Electrical impedance tomography (EIT) is explored for monitoring lung ventilation.
• Linear EIT reconstruction often assumes a homogeneous background conductivity.
• This assumption can lead to inaccuracies in clinical applications.

Purpose of the Study:

• To investigate the impact of the homogeneous background assumption on EIT image accuracy.
• To quantify errors in lung EIT imaging caused by anatomical and physiological variations.
• To evaluate the benefit of using subject-specific forward models.

Main Methods:

• Simulated EIT measurements using a 3D finite element model of thorax conductivity.
• Inclusion of variations in heart and lung properties, gravitational collapse, and ventilation.
• Analysis of three common linear EIT reconstruction algorithms.
• Comparison of results with and without subject-specific forward models.

Main Results:

• The asymmetric heart position can bias EIT ventilation images by up to 60% towards the left lung.
• Gravitational lung collapse can cause overestimation of conductivity changes in non-dependent lung by up to 100%.
• Using realistic conductivity distributions in the forward model significantly reduces these errors.

Conclusions:

• Subject-specific, anatomically accurate forward models are crucial for accurate lung EIT.
• Careful analysis of EIT images is needed for patients with heterogeneous or changing lung conductivity.
• Improved EIT reconstruction methods are essential for reliable lung ventilation monitoring.