Grading quality score of contrast electrical impedance tomography for lung ventilation-perfusion: development and preliminary evaluation of the impedance-time curve.
Source: PubMed, NCBI / U.S. National Library of Medicine
The electrical impedance tomography (EIT) with saline-bolus method has been commonly used to assess regional ventilation-perfusion (V/Q) match at the bedside. Inherent technical challenges could result in low-quality impedance curves when applying this method. This study aimed to develop and assess a quality assurance method to rate the impedance-time curve before analysis. Five dimensions in three aspects for curve scoring were determined: signal quality of EIT data collection, impedance curve during tidal breathing (variability of tidal impedance), and impedance curve related to bolus injection (platform limb, descending limb, and ascending limb). The Quality Score of EIT data was created by assigning each dimension as optimal (0 point), sub-optimal (1 point), or poor (10 points). Curves with a cumulative score of 10 or above are considered failed and are not recommended for further analysis. We retrospectively included all EIT saline contrast data from a single center in 2023. Two raters independently scored each curve. 118 data were scored with an overall failing rate of 22.9%. 8/118 failed due to signal collection quality, 4/118 failed due to the tidal impedance curve and 19/118 failed due to the contrast curve. The interrater agreement is acceptable with kappa Coefficient: tidal impedance (κ = 0.82), platform limb (κ = 0.73), descending limb (κ = 0.89), ascending limb (κ = 0.90), and overall
Abstract
The electrical impedance tomography (EIT) with saline-bolus method has been commonly used to assess regional ventilation-perfusion (V/Q) match at the bedside. Inherent technical challenges could result in low-quality impedance curves when applying this method. This study aimed to develop and assess a quality assurance method to rate the impedance-time curve before analysis. Five dimensions in three aspects for curve scoring were determined: signal quality of EIT data collection, impedance curve during tidal breathing (variability of tidal impedance), and impedance curve related to bolus injection (platform limb, descending limb, and ascending limb). The Quality Score of EIT data was created by assigning each dimension as optimal (0 point), sub-optimal (1 point), or poor (10 points). Curves with a cumulative score of 10 or above are considered failed and are not recommended for further analysis. We retrospectively included all EIT saline contrast data from a single center in 2023. Two raters independently scored each curve. 118 data were scored with an overall failing rate of 22.9%. 8/118 failed due to signal collection quality, 4/118 failed due to the tidal impedance curve and 19/118 failed due to the contrast curve. The interrater agreement is acceptable with kappa Coefficient: tidal impedance (κ = 0.82), platform limb (κ = 0.73), descending limb (κ = 0.89), ascending limb (κ = 0.90), and overall pass/fail (κ = 0.95). This scoring system reveals common quality problems that could cause potential inaccurate assessment of lung perfusion and regional V/Q. The quality score is initially repeatable across raters, but further validating investigation is needed.