Different approaches for quantifying ventilation distribution and lung tissue properties by functional EIT.

Area of Science:

• Medical Imaging
• Physiological Monitoring
• Electrical Impedance Tomography (EIT)

Background:

• Functional tomograms provide quantitative insights into lung physiology.
• Assessing the impact of data errors on tomogram accuracy is crucial for clinical application.
• Existing methods for functional lung imaging require robust error quantification.

Purpose of the Study:

• To quantitatively evaluate five methods for constructing functional lung tomograms.
• To assess the sensitivity of these methods to acquired data errors.
• To compare the performance of different data preprocessing techniques.

Main Methods:

• Generation of artificial, error-free lung data using a 'living thorax model'.
• Inclusion of physiological time courses and simulated EIT system errors (up to 50x initial level).
• Application of low-pass filtering and Principal Component Analysis (PCA) for data preprocessing.

Main Results:

• Methods fitting local to global time courses outperformed standard deviation or max/min detection methods.
• VT methods provided the best quantification of ventilation distribution.
• Lung filling capacity was least affected by increasing error levels.
• Filtering introduced significant errors in ventilation distribution analysis, while PCA showed promise for erroneous real data.

Conclusions:

• Functional tomogram accuracy is significantly influenced by data errors.
• VT methods offer superior ventilation distribution quantification.
• PCA preprocessing is a promising approach for improving the analysis of real-world, potentially erroneous EIT data.