Density-based MDCT quantification of lobar lung volumes: a study of inter- and intraobserver reproducibility.

PURPOSE: The authors assessed the reproducibility of multidetector-row computed tomography (MDCT) volumetry of the total and emphysematous parenchyma of pulmonary lobes.
MATERIALS AND METHODS: Two observers analyzed 23 MDCT examinations of patients with emphysema during two sessions held 3 months apart. Both lungs and all lobes were delimited by a combination of semiautomated and manual segmentation. Emphysematous parenchyma was obtained by applying density thresholds of -1,024/-950 HU. To assess the reproducibility of total volume (V), volume of emphysema (VE) and emphysema index (EI), intra- and interobserver differences of those measurements were assessed.
RESULTS: Total volumetry of the lungs was highly reproducible (intra- and interobserver variability of +/-3.4%). Variability between measurements was slightly greater or emphysema volume and index (EI). Lobar analyses showed large ranges of intra- and interobserver variability (intraobserver V=+/-3.7%-10.6%; VE=+/-17.3%-32.9%; EI=+/-17.8%-34%; interobserver V=+/-13.3%-98.3%; VE=+/-11%-137.6%; EI=+/-28.9%-96.4%).
CONCLUSIONS: MDCT quantification of total and emphysematous lung volume and emphysema index is overall reproducible. Quantitative assessment of those parameters performed on single lobes is affected by variability. An improvement of the reproducibility of q-MDCT is expected from the use of advanced methods for lobar segmentation.