GE Revolution CT, Руководство пользователя

Страница: 91 / 702

16.4 Voxel Dimensions
CAUTION
The image set resolution, that is, the dimensions of the voxels (volume elements)
that constitute the image set, is determined by the size of the field-of-view, the
matrix size, and the inter-slice distance. Ideally, voxels should be isotropic (with
the same dimensions along all three axes). That is, the inter-slice distance
should be the same as the voxel dimension in the acquisition plane. In practice,
however, considerations such as patient irradiation dose levels will usually lead
to the choice of a larger inter-slice distance. Be aware that details with
dimensions in the order of or less than the inter-slice distance cannot be
identified or measured with any degree of reliability.
16.5 ROI
CAUTION
Area measurement accuracy using a region of interest graphic (rectangle,
smooth curve, ellipse or free draw) is equal to the displayed area ± the perimeter
of the region multiplied by (largest voxel value) 2 /2. Mean and standard deviation
values for the intensity of the pixels in the region are also affected by this
accuracy. If the ROI is rotated, the area measurement can vary up to 5 %.
Region of interest statistics are based on the pixels INSIDE the graphic defining
the region.
CAUTION
CT Numbers are NOT absolute; misdiagnosis is possible. System and patient
variables may affect CT Number accuracy. If you rely solely upon CT numbers
without taking variables into consideration, you could misdiagnose an image.
16.6 CT Number
(Reference 1020.33 (j)(1), 1020.33 (j)(2))
Besides anatomic location and area, each CT pixel also represents a CT number, which in turn
indicates tissue density.
An image pixel represents a three dimensional volume, or voxel. It represents anatomy with a
location, an area, and a pixel (density) value. The system flattens the 0.625, 1.25, 2.5, 3.75, and
5 mm scan thickness into a two dimensional screen image. If a pixel represents a variety of
tissues, the system averages the contents to produce an averaged, rather than accurate, pixel
value. Uniform tissues (within the voxel) produce fairly accurate pixel values.
CT pixel shading shows relative density. Denser materials weaken X-ray and produce whiter
pixels. (Assumes Inverse Video OFF)
Reformat displays non axial planes created from contiguous pixels extracted from multiple
images. 3D locates similar pixel values within contiguous images, and generates a
mathematical model to produce images that appear three dimensional. BMD samples pixel
values to estimate bone or tissue density.
Reconstruction assigns one value to every image pixel. CT uses pixel values of -32767 to
+32767. MR uses pixel values of +16,000. The screen pixel translates the assigned value into
Revolution CT User Manual
Direction 5480385-1EN, Revision 1
Chapter 3 Safety 91