Objective      To observe the sites with high incidences of beam-hardening (BH) artifacts in dual-energy CT myocardial perfusion images

and compare the image quality and effect for BH correcting in different monochromatic energy images via dual-source dual-energy CT myocardial perfusion imaging.       Methods      Eighty-four patients requiring cardiac enhancement-CT examination with suspected or definite coronary artery disease underwent dual-energy (100 kVp/Sn 140 kVp) myocardial perfusion imaging with dual-source CT in our department

and were enrolled in this study. Iodine maps and monoenergetic images under 65

75

85

95 and 105 keV were obtained after the process of specialized post-processing software. M image (weighted polychromatic 120 kV) was generated by the system automatically and served as control. The sites and numbers of beam-hardening artifacts were observed and counted. The rate of artifacts

objective measurements of myocardial iodine content

CT value

signal-to-noise ratio (SNR)

contrast to noise ratio (CNR) and subjective image quality grades were used to comprehensively assess the value of monoenergetic images.       Results      The number of artifacts was 274 on 65 keV images

and was reduced gradually with the increase of keV (P0.05). The sites with high incidence of artifacts were the left ventricular basal segment 36.5% (100/274)

followed by ventricular septal segment 25.91% (71/274)

superior-vena-cava-besides segment 24.82% (68/274) and posterior segment 12.77% (35/274)(P0.05). The iodine content of myocardium was -2.37±0.84 mg/g under artifacts

significantly lower than non-under artifacts [3.5±0.92 mg/g

t=-19.36

P0.01]. The non-artifact myocardial CT number of 85 keV was 93.52±18.52 HU for non-artifact

which was the closest to polychromatic 120 keV (96.06±16.32 HU)

and significant difference was seen between them (P0.01). SD of 85 keV was the lowest (12.98±3.16). SNR of non-under artifacts in the myocardium was higher in 75 keV (6.89±1.79) and 85 keV (6.5±1.7)

with no statistical difference between the 2 groups (t=1.5

P=0.14). CNR was the highest in 75 keV (-18.79±6.76

P0.01)

and there was no statistical difference between 85 keV (-15.4±5.67) and 65 keV (-15.53±4.56

t=0.27

P=0.79). The 85 keV images got the highest image quality scores. The evaluation for image quality of the 2 radiologists was highly consistent (Kappa=0.79

P0.05).       Conclusion      The affected regions of beam-hardening artifacts are common in the left ventricular basal segment

ventricular septal segment and superior-vena-cava-besides segment in turn. Myocardial density of 85 keV is closest to polychromatic 120 kV image

but 85 keV can most effectively correct beam-hardening artifacts with better image quality.