The effects of radiation dose and CT manufacturer on measurements of lung densitometry

Background: To evaluate the effect of radiation dose and scanner manufacturer on quantitative CT scan measurements of lung morphology in smokers. Methods: Low-dose and high-dose, inspiratory, multislice CT scans were obtained in 50 subjects at intervals of approximately 6 months (mean [+/- SD] interval, 0.5 +/- 0.2 years). In another 30 subjects, multislice CT scans were acquired first using a GE LightSpeed Ultra (General Electric Healthcare; Milwaukee, WI), followed a mean time of 1.2 +/- 0.4 years later by using a Siemens Sensation 16 scanner (Siemens Medical Solutions; Erlangen, Germany). Custom software was used to measure lung volume, mass, mean density, and the extent of emphysema using threshold cutoffs of -950, -910, and -856 Hounsfield units (HU) and the lowest 15th and 5th percentile points. Results: The change in radiograph dose significantly affected measurements of emphysema assessed using mean lung density, threshold, or percentile methods. There were also interactions between dose and total lung volume for all of the measurements except the -950-HU threshold and the lowest fifth percentile point. These two emphysema measurements suggest that there was more emphysema found in the CT scans obtained using a lower radiograph dose. Only the mean lung density and -856-HU threshold showed significant effects between CT scanner manufacturers and interactions between total lung volume and scanner. All other measures of lung structure were not different between the two CT scanners. Conclusion: CT scan measurements of very low density lung structures are significantly affected by radiation dose but are less sensitive to the lung volume. Image acquisition parameters including radiation dose, scanner type, and the subject's breath size should be standardized to estimate emphysema severity in longitudinal studies.