Windowing (computed tomography)

This article describes windowing in computed tomography, for windowing in signal processing see window function .

CT of the lungs, shown in the lung window with C = -700 HU and W = 1500 HU; there is shown a atelectasis in the right middle lobe. As can be clearly seen, there is no deep black in the entire image, as the window width extends beyond -1000 HU. The structures of the lungs are clearly visible. It is also noticeable that the image cannot distinguish between bones and soft tissue. Since the upward windowing only extends up to the CT number of 50 HU (-700 + 1500/2), bones and soft tissue are displayed in white.

CT of the lungs, shown in the soft tissue window with C = 40 HU and W = 350 HU; a funnel breast is shown . By choosing the center with a value of 40 and a window width of 350, soft tissue can be assessed very well. However, tissue structures in the lungs cannot be seen at all. Just as little as denser parts of the bones, whose HU values ​​sometimes far exceed the selected upper limit of 215 HU (40 + 350/2)

In computed tomography , the windowing determines which section of the scale of the measured density values ​​(= Hounsfield scale) is assigned to gray values ​​from black to white in the image. This optimizes the image contrast of the tissue to be examined.

Windowing is also used in magnetic resonance imaging for the display of signal intensities and in digital projection X-rays.

background

The Hounsfield scale is determined by the absorption coefficients of air and water as fixed points and expanded by inter- or extrapolation. It is basically unlimited, but often only the range from -1024 HU to 3071 HU is displayed on clinical systems. These attenuation values ​​occurring in tissue can be differentiated by measurement; however, the human eye can only distinguish between 60 and 80 shades of gray. In order to make the entire image information accessible to the viewer, only the part of the Hounsfield scale that represents the image content to be assessed is shown in the image through various windows. As when looking outside through a window of a house, the viewer only sees part of the overall picture. Depending on the windowing, different anatomical details can be seen in the same image.

definition

The visible section of the Hounsfield scale is given with two values, the center and the width, and is calculated as follows:

${\displaystyle Zentrum-{\frac {Breite}{2}}\qquad bis\qquad Zentrum+{\frac {Breite}{2}}}$

On clinical evaluation workstations and CT images, the term “center” is usually abbreviated to “c” (English center ) and the term “width” is usually abbreviated to “w” (English width ). Instead of “c”, “l” (English level ) is sometimes also written.

Typical windowing

The following table shows example windowings with common values, shown in Hounsfield units for center and latitude.

Window name	center	width
Lung window	-600	1600
Bone window	300	2000
Soft tissue window	60	360
Brain window	40	80
CT angiography window	100	900

Example lung window

For the assessment of the lungs, a section from the Hounsfield scale is selected, which includes the density values ​​of the air in the lungs and the lung tissue. The lowest density value of this section is assigned to the 'gray value' black, the highest to the value white. All density values ​​in between are linearly assigned the gray values ​​between black and white. A lung window with a center at -600 U and a width of 1600 U thus ranges from -1400 U to 200 U. Density areas smaller than -1400 are shown consistently black, those above 200 consistently white.

The lung window, however, takes on a certain special position: Since the smallest possible attenuation value on the Hounsfield scale is -1000 (the attenuation of air), there are no CT numbers in the image that are smaller than -1000. Strictly speaking, it is a windowing that ranges from -1000 to 200 and cannot contain the value black for the reasons mentioned. The darkest value in the picture is a shade of gray.

Individual evidence

1. ↑ Handels H .: Medical image processing: image analysis, pattern recognition and visualization for computer-aided medical diagnostics and therapy , Teubner Verlag, 2009, p. 285, ISBN 3835100777 , here online
2. ^ Calendar W. A .: Computed tomography. Basics, device technology, image quality, applications with multi-slice spiral CT. Publicis MCD advertising agency Munich 2000; ISBN 3-89578-082-0 .