Congress:
EuroSafe Imaging 2019
Keywords:
Action 3 - Image quality assessment based on clinical indications, Liver, CT, Diagnostic procedure, Equipment, Physics, Image verification
Authors:
I. Fitton, A. Luciani, E. Mousseaux, C. A. Cuenod, A. Rahmouni, O. Clement
DOI:
10.26044/esi2019/ESI-0055
Background/introduction
Computed Tomography (CT) values characterize the linear attenuation coefficient of the tissue in each volume element relative to the µ-value of water.
The Hounsfield Units (HU) or CT numbers are defined as a normalized index of X-ray attenuation based on a scale of –1000 HU (air) to +1000 HU (bone),
including water equal to 0 HU at standard temperature and pressure [1].
According to most investigators,
healthy liver tissue CT numbers must be between 50 HU and 70 HU [2,
3].
Some clinical recommendations advocate the use of HU values measurement for disease assessment,
considering CT image matrix as reliable and accurate.
For instance liver steatosis can be detected and characterized by the liver/spleen ratio measurement in HU [4,
5],
while portal venous phase CT valued of liver metastases can be a marker of response to novel therapy regimens [6].
The follow-up of HU values has also already been used in patients with non-alcoholic fatty liver disease [7],
and even in patients with idiopathic hemochromatosis treated by phlebotomy by providing an index of hepatic iron stores without requiring a liver biopsy [8].
Mean CT attenuation liver values were also used by Altenbernd et al.
[9] for the detection of hepatic metastases in patients with uveal melanoma.
All these medical studies do not provide specific information regarding the HU calibration criteria,
which should be necessary to ensure the reproducibility and the accuracy of their conclusions.
However,
the CT numbers of tissues are not constant data; they may be influenced by the convolution kernel,
the reconstruction algorithms [10] and the patient morphology.
The HU variability may also be related to the difference in the energy spectra of the X-rays; this was already reported more than thirty five years ago [11,
12].
Recently,
the variability of CT attenuation values had been confirmed [13,
14]: Birnbaum et al.
showed in a phantom study that CT attenuation values were different for all tissues between manufacturers and among different generations of CT equipments.
Thus by considering HU variability between MDCT scanners,
Lamba et al.
[14] concluded that established absolute HU thresholds can alter the diagnosis performance of CT examinations.
These studies provided valuable information in HU variability understanding between MDCT manufacturers,
however the reliability of CT attenuation values for a specific quantitative clinical CT application during lifetime of the same CT equipment,
has never been considered.
The aim of our study was therefore to evaluate the fluctuation of liver HU absolute values and contrast in a phantom over several years,
with acquisitions being performed on two identical CT equipments and to propose some recommendations for CT scanner tissue calibrating.