Philips Azurion - Mean CT Number (XperCT); Uniformity; Modulation Transfer Function

Figure 143 XperC
T reconstrucon of the Catphan 500 phantom, CTDI
w
50 mGy, slice thickness 5.2 mm
Mean CT Number (XperCT)
XperCT is calibrated to produce values that are accurate for the Hounseld scale, using several
calibraon methods:
• Detector gain
• Water beam hardening
• Paent scaer
• Intra-detector scaer
Truncaon is corrected through parabolic extrapolaon of proles.
Accuracy of mean CT numbers is limited mainly due to variaons in an-scaer grid (relevant for neuro
and abdomen), paent truncaon (relevant for abdomen) and incompleteness of circular orbit
(resulng in cone-beam arfacts). From clinical experience, from 2006 onwards, Philips Medical
Systems has observed that the mean CT number is accurate to within approximately 20H (neuro) and
50H (abdomen, depending on paent size, due to truncaon). For the intended use of this product, this
level of accuracy is sucient.
Uniformity
Uniformity within an axial single slice (near the plane of rotaon) is dependent on the type of object
scanned. Philips Medical Systems has measured uniformity using water phantoms where uniformity is
beer than 10H. This level of uniformity is also observed with XperCT Head. Uniformity with XperCT
Abdomen is approximately 50H, mainly limited by scaer and truncaon.
Uniformity across axial slices is less than within axial slices (near plane of rotaon), due to
incompleteness of semi-circular orbit of approximately 200 degrees, resulng in cone-beam arfacts
which are clearly recognizable by the user (streaks).
Modulaon Transfer Funcon
XperCT uses a linear reconstrucon algorithm with a single reconstrucon kernel. The graph below
shows the measured modulaon transfer funcon for a typical XperCT reconstrucon.
Technical In
formaon Protecon Against Stray Radiaon
Azurion Release 1.2 Ins
trucons for Use 320 Philips Healthcare 4522 203 52421