[Quantification of atelectases in artificial respiration: spiral-CT versus dynamic single-slice CT]

Abstract

Purpose: Dynamic CT (dCT) allows visualization and quantification of ventilated lung and atelectases with high temporal resolution during continuous ventilation. This study compares a quantitative image analysis in a subcarinal single slice dCT series versus a whole lung spiral-CT, in order to analyze, whether the distribution of atelectasis of a single dCT series is representative for the whole lung.

Materials and methods: dCT in sliding windows technique (slice thickness 1 mm, temporal increment 100 ms) was performed in 8 healthy pigs 3 cm caudal to the carina during continuous mechanical ventilation. Subsequently, a spiral-CT of the whole lung (slice thickness 2 mm; pitch 1.5; increment 2 mm) was acquired during inspiratory breath hold (airway pressure 20 mbar). Lung segmentation and planimetry of predefined density ranges were achieved using a dedicated software tool in both data-sets. Thus, the fractions of the following functional lung compartments were averaged over time: hyperinflated lung (- 1024 to - 910 HE), normal ventilated lung -900 to -300 HE) and atelectasis (-300 to +200 HE).

Results: Quantitative analysis of dCT-series during continuous respiration correlated with the density analysis in spiral-CT as follows: hyperinflated lung r = 0.56; normal ventilated lung r = 0.83 and atelectases r = 0.84. Analysis of spiral-CT showed the following distribution of functional lung compartments: hyperinflated lung 3.1% normal ventilated lung 77.9% and atelectasis 19.0%. In dCT, hyperinflated lung represented 6.4%, normal ventilated lung 65.2% and atelectasis 28.4% of total the lung area.

Conclusion: The results of our study demonstrate that dCT allows monitoring of atelectasis formation in response to different ventilatory strategies. However, a deviation between dCT and spiral-CT has to be taken into account. In subcarinal dCT series, hyperinflated lung areas and atelectases were overestimated due to a craniocaudal gradient of atelectases, whereas normal ventilated lung was underestimated.