Abstract
Abstract
Photon-counting detector computed tomography (PCD-CT) has
demonstrated improvements in conventional image quality compared to
energy integrating detector (EID) CT. PCD-CT has the advantage of
being able to operate in conventional and spectral mode
simultaneously by sorting photons according to selected energy
thresholds. However, to reconstruct spectral images PCD-CT requires
extensive calibration and specifically fine-tuning a spectral
response. This response is then used to perform material
decomposition (MD). We propose a step-wedge phantom made of water
and iodine to calibrate a prototype PCD-CT system. Four methods were
tested and compared based on calibration accuracy and CT image
quality. The exhaustive PCD response was not well calibrated, but a
reduced model was defined that was able to perform accurate
water/iodine MD and to reduce the ring artifact intensity. The
impact of the number of energy bins (from two to seven) was also
studied. The number of bins did not affect the spectral
accuracy. However, compared to the two energy bin configuration, the
seven bin configuration decreased the noise by 10% and 15% in the
water and iodine maps, respectively. The model was tested on ex-vivo tissue samples injected with iodine to demonstrate
the results of the water/iodine MD on biological materials.