Abstract:

Objective  To explore the correlation between CT indexes of emphysema and lung function of patients with chronic obstructive pulmonary disease (COPD) and effects of section thickness and reconstruction algorithms. Methods  Thirty-nine COPD patients at stable phase were selected and underwent chest CT (deep inspiratory phase) and lung function test. CT images were reconstructed by standard algorithm with section thicknesses of 0.625, 1.25, 5, 7.5, and 10 mm and by bone algorithm with section thicknesses of 0.625 and 1.25 mm. Low attenuation area percent (LAA%), the lowest 15th percentile of the histogram of attenuation values (Perc 15), mean lung density, and lung volume were measured using the best threshold and default threshold. The effects of different section thicknesses and reconstruction algorithms on indexes and the correlation between indexes and lung function were analyzed. Results  LAA% decreased with the increase of section thickness and bone algorithm was higher than standard algorithm. Perc 15 and mean lung density increased with the increase of section thickness and bone algorithm was lower than standard algorithm, while section thickness and algorithm had little effect on the measurement of lung volume. For each section thickness and reconstruction algorithm, FEV1, FEV1%pred, FEV1/FVC, MMEF%pred, DLCO, DLCO%pred, DLCO/VA%pred negatively correlated with LAA% (P<0.05) and positively correlated with Perc 15 (P<0.05). FEV1%pred and FEV1/FVC positively correlated with mean lung density (P<0.05). RV, TLC, and RV/TLC positively correlated with lung volume on CT (P<0.05). For pulmonary ventilation indexes and diffusing capacity, their correlations with LAA% and Perc 15 were higher than those with mean lung density and lung volume. For lung volume indexes, the correlations with lung volume on CT were higher than those with other parameters. The section thickness did not have obvious effect on the correlation between CT assessment of emphysema and lung function. The correlations between LAA% or Perc 15 and some pulmonary ventilation index (such as FEV1) of bone reconstruction were higher than those of standard reconstruction. Conclusion  LAA% and Perc 15 measured from chest CT images can ideally reflect the ventilation indexes and diffusing capacity of lung function. The lung volume on CT can ideally reflect the volume indexes of lung function. Attention should be paid to the consistency of section thickness and reconstruction algorithm when performing CT emphysema evaluation and comparison. Modifications are needed if inconsistency exists.

Key words: chronic obstructive pulmonary disease, pulmonary emphysema, CT, section thickness, restruction algorithm