Photon-Counting CT of the Brain: In Vivo Human Results and Image-Quality Assessment

Abstract

Background and purpose: Photon-counting detectors offer the potential for improved image quality for brain CT but have not yet been evaluated in vivo. The purpose of this study was to compare photon-counting detector CT with conventional energy-integrating detector CT for human brains.

Materials and methods: Radiation dose-matched energy-integrating detector and photon-counting detector head CT scans were acquired with standardized protocols (tube voltage/current, 120 kV(peak)/370 mAs) in both an anthropomorphic head phantom and 21 human asymptomatic volunteers (mean age, 58.9 ± 8.5 years). Photon-counting detector thresholds were 22 and 52 keV (low-energy bin, 22-52 keV; high-energy bin, 52-120 keV). Image noise, gray matter, and white matter signal-to-noise ratios and GM-WM contrast and contrast-to-noise ratios were measured. Image quality was scored by 2 neuroradiologists blinded to the CT detector type. Reproducibility was assessed with the intraclass correlation coefficient. Energy-integrating detector and photon-counting detector CT images were compared using a paired t test and the Wilcoxon signed rank test.

Results: Photon-counting detector CT images received higher reader scores for GM-WM differentiation with lower image noise (all P < .001). Intrareader and interreader reproducibility was excellent (intraclass correlation coefficient, ≥0.86 and 0.79, respectively). Quantitative analysis showed 12.8%-20.6% less image noise for photon-counting detector CT. The SNR of photon-counting detector CT was 19.0%-20.0% higher than of energy-integrating detector CT for GM and WM. The contrast-to-noise ratio of photon-counting detector CT was 15.7% higher for GM-WM contrast and 33.3% higher for GM-WM contrast-to-noise ratio.

Conclusions: Photon-counting detector brain CT scans demonstrated greater gray-white matter contrast compared with conventional CT. This was due to both higher soft-tissue contrast and lower image noise for photon-counting CT.

© 2017 by American Journal of Neuroradiology.

Figures

Fig 1.

Radial noise-power spectrum (NPS) measured in an anthropomorphic head phantom for energy-integrating detector and photon-counting detector scans at 120 kVp and 370 mAs. The PCD curve was lower than the EID curve. The difference is more prominent at mid-to-high spatial frequencies.

Fig 2.

Blinded reader evaluation of image quality for energy-integrating detector and photon-counting detector head images. PCD scores are better for gray matter-versus-white matter differentiation and image noise, whereas EID scores are better for posterior fossa image quality (all P < .001, paired Wilcoxon signed rank test). Image quality scores are based on the European Guidelines for Image Quality Criteria for Computed Tomography.

Fig 3.

Example energy-integrating detector and photon-counting detector images of a 59-year-old woman (section thickness, 2 mm; increment, 2 mm; window center, 45 HU; window width, 80 HU). A, Axial EID reconstruction at the level of the basal ganglia. B, Axial PCD reconstruction at the same level as A. Lower image noise is shown for the PCD image. Zoomed-in EID (C) and PCD (D) images at the same level as A and B. C indicates caudate; I, internal capsule; L, lentiform nucleus; Th, thalamus.

Fig 4.

Sample energy-integrating detector and photon-counting detector images of a 67-year-old man (section thickness, 2 mm; increment, 2 mm; window center, 45 HU; window width, 80 HU). A, Coronal EID image at the level of the basal ganglia. B, Coronal PCD image at the same level as A shows lower image noise for the PCD image. Zoomed-in EID (C) and PCD (D) images at the same level. C indicates caudate; I, internal capsule; L, lentiform nucleus.

Fig 5.

Sample photon-counting detector images of low- and high-energy bins of a 70-year-old woman (section thickness, 2 mm; increment, 2 mm; window center, 45 HU; window width, 80 HU). A, Axial PCD image reconstructed from all detected photons (22–120 keV) at the level of the basal ganglia. B, Axial PCD image reconstructed from a low-energy bin image (22–52 keV) at the same level as A. C, Axial PCD image reconstructed from the high-energy bin image (52–120 keV) at the same level as A. The image noise for both the low- and high-energy bins is higher than that of the PCD image reconstructed from all detected photons because each bin contains only a subset of all detected photons. The low-energy bins provide good gray matter–white matter differentiation but are susceptible to beam-hardening, best seen as an artifactual increase in attenuation of the cortical GM and the subarachnoid space (arrows in B). The high-energy photons are less susceptible to beam-hardening but have poorer GM–WM differentiation. The image reconstructed from all photons is a trade-off between the good GM–WM differentiation of the low-energy image and the lower beam-hardening artifacts of the high-energy images.