Neurovascular Imaging with Ultra-High-Resolution Photon-Counting CT: Preliminary Findings on Image-Quality Evaluation

BACKGROUND AND PURPOSE: The first-generation photon-counting detector (PCD) CT was recently introduced into clinical practice and represents a promising innovation in high-resolution CT imaging. The purpose of this study was to assess the image quality of ultra-high-resolution (UHR) PCD-CT compared with energy-integrating detector (EID)-CT, and to explore different reconstruction kernel sharpness levels for the evaluation of intracranial aneurysms. MATERIALS AND METHODS: Ten patients with intracranial saccular aneurysms, who had previously undergone conventional EID-CT, were prospectively enrolled. CT angiograms were acquired on a clinical dual-source PCD-CT in UHR mode, and reconstructed with four vascular kernels (Bv36, Bv40, Bv44, Bv48). Quantitative and qualitative image quality parameters of the intracranial arteries were evaluated. For the quantitative analysis (image noise, SNR, CNR), regions of interest were manually placed at standard anatomical intracranial and extracranial locations by one author. In addition, vessel border sharpness was evaluated quantitatively. For the qualitative analysis, three blinded neuroradiologists rated PCD-CT and EID-CT image quality for the evaluation of the intracranial vessels (i.e., the aneurysms and nine standard vascular branching locations) on a 5-point Likert-type scale. Additionally, readers independently selected their preferred kernel among the four kernels evaluated on PCD-CT. RESULTS: In terms of quantitative image quality, Bv48, the sharpest kernel, yielded increased image noise, and decreased SNR and CNR parameters compared to Bv36, the smoothest kernel. Compared to EID-CT, the Bv48 kernel offered better quantitative image quality for the evaluation of small intracranial vessels (p < .001). Image quality ratings of the Bv48 were superior to those of the EIDCT, and not significantly different from ratings of the B44 reconstruction kernel. When comparing side-by-side all four PCD-CT reconstruction kernels, readers selected the B48 kernel as the best to visualize the aneurysms in 80% of cases. CONCLUSIONS: UHR PCD-CT provides improved image quality for neurovascular imaging. Although the less sharp kernels provided superior SNR and CNR, the sharpest kernels delivered the best subjective image quality on PCD-CT for the evaluation of intracranial aneurysms.