Premium
SU‐E‐I‐18: Evaluation of a Lung Density CT Standard Reference Material
Author(s) -
ChenMayer H,
Avila R,
John L,
Levine Z,
Yankelevitz D
Publication year - 2015
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4924015
Subject(s) - standard deviation , nuclear medicine , imaging phantom , scanner , centroid , histogram , hounsfield scale , medicine , mathematics , materials science , physics , computed tomography , radiology , statistics , computer science , optics , image (mathematics) , geometry , artificial intelligence
Purpose: To evaluate quantitative lung density CT using a Standard Reference Material (SRM) with SI traceable physical densities in a clinical setting free from physiological noise. Methods: The 5‐density (0.06 to 0.23 g/cc) suite of polyurethane foams developed as a CT lung density reference (SRM‐2088) was measured inside an anthropomorphic chest phantom in a 64‐row scanner, to assess the mean HU value and its distribution. Two tube current settings (200 mA and 50 mA) were used to represent high‐ and low‐dose CT clinically, with back‐to‐back repeat scans, and for each scan the image reconstruction was performed at slice thicknesses of 0.625 mm and 1.25 mm. A volume histogram was generated for the entire suite, and was fitted to a function with 5 Gaussian peaks, each reporting a centroid (mean HU) and standard deviation (SD in HU) as a measure of the accuracy and precision of the measurement. Statistical analysis was performed using the two levels of mA and recon slice thickness as factors to study the effect on the centroid and the SD. Results: Neither dose nor thickness has significant effect on the centroid, but both have significant effect on the SD. At 200 mA, the observed average SD is about half of that at 50 mA. With the thickness doubled, the SD decreased by a factor of 1.17, providing a parameter relating spatial resolution and noise. The SD is more than 3 times higher when compared to the SRM in‐air measurement (without the attenuation and scattering from the chest wall) performed earlier, whereas for the centroid excellent agreement was observed. Conclusion: The study provides the underlying uncertainty assessment for these standard references in the lung density range in a set of scanning conditions, forming the basis for estimating the sensitivity to changes in lung density in a clinical setting. Dr. David Yankelevitz receives royalty payments from General Electric Company.