Lung volume estimation feasible with automatic segmentation of chest CT

Automatic lung segmentation of routine chest CT scans allows for a technically stable estimation of lung volume, though work remains to reduce variation, according to a study published in the May issue of Radiology.

“CT of the chest is widely applied because it combines fast image acquisition with detailed anatomic depiction in a wide range of clinical applications,” wrote lead author Mathias Haas, MD, of the Charité Universitätsmedizin Berlin in Germany, and colleagues. “Without further diagnostic expenditure, the acquired data can be used for CT volumetry. There is a range of clinical scenarios where CT lung volumetry is of interest.”

Although lung volumes can be easily calculated from chest CT scans, challenges remain in using this information as an indicator for pulmonary disease when it’s obtained from routine CT.

The researchers assessed the reliability of volumetric measurements in routine CT lung scans by examining 302 chest CT scans from 40 adult cancer patients whose lungs were unaffected by disease. Lung volume was then calculated using the appropriate software, which yielded an average of 7.2 successfully evaluable CT scans per patient.

The study’s results highlighted the following findings: lung volume was approximately normally distributed, as demonstrated with a mean of 5283 cm3. Between different scans in one and the same patient, the median intra-individual standard deviation in lung volume was 853 cm3, which was 16 percent of the mean lung volume.

No significant correlation was found between calculated lung volume and the CT scanner used.

“With high intra-subject inter-test variation, the standardization of automated CT lung volumetry remains a challenge,” the authors concluded.