Sean David Mobberley BS, Presenter: Nothing to Disclose

Jered Sieren, Abstract Co-Author: Research Consultant, VIDA Diagnostics, Iowa City, IA

Matthew K. Fuld BS, Abstract Co-Author: Nothing to Disclose

Keith Gunderson, Abstract Co-Author: Nothing to Disclose

Eric A. Hoffman PhD, Abstract Co-Author: Shareholder, VIDA Diagnostics Medical Advisory Board, Siemens AG

Reliable, repeatable, and accurate quantification of volumetric CT data for assessment of lung density, particularly for longitudinal and multi-center studies remains elusive. With the advent of new dual source scanners that adjust x-ray energy spectra and modify reconstruction kernels, it is of interest to understand how the quantitative measures of lung density compare to single source scan modes.

Three supine anesthetized sheep were studied using a Dual Source MDCT scanner (Siemens Definition Flash). Two dual-energy spiral scans at 80/140Sn kV and 100/140Sn kV, were acquired using a B35f kernel (specific to single energy mode) and D30f kernel (specific to dual energy mode) as well as three single energy scans at 80, 100, and 140 kV respectively (B35f kernel). Density histograms were evaluated for both the lung and tracheal segments and the median HU values were measured along with the standard deviations. In addition, a customized phantom was imaged using the same technique to evaluate the consistency of the in vivo measurements. The phantom air hole was surrounded by water-like material to simulate chest characteristics. Scans were obtained at 100 kV in single source mode (B35f kernel) and 100/140Sn kV in dual source mode (B35f and D30f kernel).

When using raw projection data gathered in the dual source scanning mode, the median air value in the tracheal regions of both sheep and phantom show consistently more negative (closer to -1000) values when the raw data is reconstructed with either the kernel designed for use in the dual energy (D30f) mode or the kernel traditionally used in single energy mode (B35f). When image data is gathered in single source mode, the traditional kernel (B35f) demonstrated a positive shift of up to 35 HU.

It appears from our data that the x-ray chain rather than the reconstruction kernel may play a greater role in generating tracheal air which is closer to the true air value of -1000HU. This may relate to efforts made in the dual energy mode to separate the energy spectra for improved dual energy-based material decomposition. More investigation is needed to further understand these dual energy modes and their effects on CT density.

By bringing air values in the trachea closer to the correct -1000HU value, characterization and distribution of emphysema may be closer to the truth and thus may be more repeatable over time.

Mobberley, S, Sieren, J, Fuld, M, Gunderson, K, Hoffman, E, Improved Quantitation of Tracheal Air Density with a Dual Energy CT Scanning Mode.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9012375.html