Issac Leichter PhD, Presenter: Employee, Siemens AG

Eli Ben-Shimol BSc, Abstract Co-Author: Nothing to Disclose

Itai Kadosh BSc, Abstract Co-Author: Nothing to Disclose

Richard Lederman MD, Abstract Co-Author: Research Consultant, Siemens AG

Naama Rachel Bogot MD, Abstract Co-Author: Research Consultant, Siemens AG

Jacob Sosna MD, Abstract Co-Author: Consultant, ActiViews Ltd Research grant, Koninklijke Philips Electronics NV

Lung nodules can be missed on CT, particularly central nodules and nodules located next to vessels. Our purpose was to improve visualization of lung nodules by a novel method using spectral data analysis of the high and low energy images obtained with Dual Energy CT.

The pilot study included 14 datasets, from 6 patients with suspected lung nodules, obtained by Dual Energy CT (DECT) based on a double layer detector (Philips). The DECT, operated at 120 kVp, provided two series of images obtained by photons with effective energy levels of 56 keV and 82 keV. Nine datasets (3 patients) were obtained with intravenous iodinated contrast media. Since the variation in the HU values with the photon energy is not linear, subtracting the low energy (LE) image from the high energy (HE) image yielded an image with a new contrast. The separation of nodules from lung vessels and surrounding structures on the subtracted images was assessed for applying to a novel color display.

The HU values of tissues denser than water were higher in the LE image, but those of adipose tissues were higher in the HE image. The gray levels (GL) in the subtracted images were translated to yield positive values. The noise in the subtracted images was significantly higher (p<0.001) than in the original images and required further filtration. The difference between the mean GL values of lung nodules and blood vessels was not significant either in the HE images (7.4%, p=0.06) or in the LE images (4.3%, p=0.25). Following image subtraction, the mean contrast of lung nodules to blood vessels significantly increased to 41.3% (p<0.001). In examinations without contrast media, the contrast increased from 11.6% in the HE image to 22% in the subtracted image (p<0.05), while with contrast media it increased from 5.2% to 55.8% (p<0.001). Appropriate windowing of the subtracted images with the significantly higher contrast allowed for a color display of lung nodules.

A novel method of processing the high and low energy images obtained with Dual Energy CT images can significantly increase the contrast between lung nodules and blood vessels, even without the use of contrast media.

Because lung nodules located next to vessels can be missed on CT, novel processing of the Dual Energy CT images can be used to improve nodule visualization, even without the use of contrast media.

Leichter, I, Ben-Shimol, E, Kadosh, I, Lederman, R, Bogot, N, Sosna, J, A Novel Method of Processing Dual-Energy CT Images for Improved Visualization of Lung Nodules.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11004850.html