Virtual noncontrast (VNC) images reconstructed from contrast-enhanced PCCT provided accurate abdominal attenuation values compared with true noncontrast (TNC) imaging, supporting the potential to omit a separate TNC acquisition.
Across all measurements, VNC attenuation errors were <5 HU in 76% and <10 HU in 95%; there was no meaningful overall difference between VNC reconstructed from arterial versus portal venous acquisitions.
VNC had inferior image quality and higher/blotchier noise than TNC, but remained diagnostically usable in 99–100% of patients.
Access publication
Figure 2: CT attenuation measurements in a 58-year-old man referred for follow-up examination after an acute type B aortic dissection. Axial CT images acquired with photon-counting detector CT depict exemplary region-of-interest (ROI) measurements on venous virtual noncontrast reconstructions. avg = average attenuation within region of interest; d = diameter of region of interest.
Protocol
Detail | Value |
|---|
Scanner | NAEOTOM Alpha |
Detector | Dual-source, two cadmium telluride PCDs |
Scan area | Thoracoabdominal |
Scan phases | TNC + arterial + portal venous |
Acquisition mode | Single-source multienergy, QuantumPlus |
Tube voltage [kV] | 120 |
Tube current | Automated modulation, CARE Dose4D / CARE kV |
IQ level | 68 |
Rotation time [s] | 0.25 |
Pitch | 1.2 |
Detector slice collimation | 144 × 0.4 mm |
Contrast | 70 mL iopromide, 370 mg I/mL |
Injection rate | 4 mL/s |
Arterial timing | Bolus tracking; 140 HU threshold at 90 kV + 12-s delay |
Portal venous timing | 70 s after contrast injection |
Reconstruction kernel | Br36 |
Iterative reconstruction | QIR 4 |
Slice thickness [mm] | 2.0 |
Reconstruction increment [mm] | 1.6 |
Reconstruction plane | Axial |
VNC source phases | Arterial and portal venous |
Reference standard | True noncontrast (TNC) images |
Key Results
Parameter | Main finding |
|---|
Study population | 100 patients; mean age 72 ± 10 years; 81 men |
Overall VNC error <5 HU | 76% of measurements |
Overall VNC error <10 HU | 95% of measurements |
Arterial vs venous VNC error | 3.3 vs 3.5 HU; P = .16 |
Left liver attenuation error | 1.4 ± 1.3 HU arterial VNC; 1.5 ± 1.2 HU venous VNC |
Liver error <5 HU | Left lobe 99%; right lobe 98% for both arterial and venous VNC |
Diagnostic subjective quality | 99–100% |
VNC noise | 33% higher than TNC in phantom |
Noise texture | VNC more blotchy; average spatial frequency 0.14 vs 0.20 mm⁻¹ |
Spatial resolution | Similar overall; TTF50 0.29 mm⁻¹ for both |
Detectability index | VNC approximately 15% lower, but always >20 |
The liver showed particularly small VNC attenuation errors, with left-lobe mean errors around 1.5 HU. Phantom testing confirmed higher and blotchier VNC noise but similar middle-frequency spatial resolution and a detectability index consistently above 20.
Take-home: PCCT VNC can provide quantitatively accurate abdominal attenuation without a dedicated noncontrast acquisition. The principal trade-off is higher, blotchier image noise and somewhat poorer subjective image quality, although diagnostic quality remained very high.
Source: Mergen V, Racine D, Jungblut L, et al. Virtual Noncontrast Abdominal Imaging with Photon-counting Detector CT. Radiology. 2022;305(1):107–115. doi:10.1148/radiol.213260.
