Pulmonary emphysema: size distribution of emphysematous spaces on multidetector CT images--comparison with...

To test the hypothesis that the frequency-size distribution of low-attenuation areas could be a parameter to quantify pulmonary emphysema.

Ethics committee approval and written informed consent were obtained. Multidetector computed tomographic (CT) scans were performed with simultaneous acquisition of four 1-mm sections of the whole chest in 80 patients (57 men, 23 women; age range, 38-79 years) who were referred for surgical resection of lung cancer. From the raw data, 1.25-mm-thick sections were reconstructed at 10-mm intervals. The relative area of lung with attenuation coefficients lower than -960 HU (RA(960)) and the 1st percentile of the distribution of attenuation coefficients were calculated. The cumulative frequency-size distributions of the RA(960) and the 1st percentile, when represented on a log-log plot, followed linear relationships. The slopes of these lines (D(960) and D(p1)) were compared with areas found macroscopically to have emphysema and with two different microscopic measurements assessed on resected specimens. Spearman correlation coefficients of each CT index with macroscopic and microscopic measurements were calculated.

The RA(960) and the 1st percentile showed statistically significant correlations with macroscopic and microscopic indexes (P < .001), whereas D(960) and D(p1) did not (P > or = .165).

The RA(960) and the 1st percentile reflect the extent of emphysema as compared to macroscopic and microscopic measurements, while D(960) and D(p1) do not.