Reconstruction Kernel Optimization for Ultra-High-Resolution Photon-Counting Detector Computed Tomography of the Lung

Background The latest generation of computed tomography (CT) systems based on photon-counting detector promises significant improvements in several clinical applications, including chest imaging. Purpose The aim of the study is to evaluate the image quality of ultra-high-resolution (UHR) photon-counting detector CT (PCD-CT) of the lung using four sharp reconstruction kernels. Material and Methods This retrospective study included 25 patients (11 women and 14 men; median age, 71 years) who underwent unenhanced chest CT from April to May 2023. Images were acquired in UHR mode on a clinical dual-source PCD-CT scanner and reconstructed with four sharp kernels (Bl64, Br76, Br84, Br96). Quantitative image analysis included the measurement of image noise, and the calculation of signal-to-noise ratio, and contrast-to-noise ratio. Two radiologists independently rated the images on a 5-point Likert scale for image sharpness, image noise, overall image quality, and airway details. The 4 image sets were compared pairwise in the statistical analysis. Results Image noise was lowest for Br76 (74.16 ± 22.05, P < 0.001). Signal-to-noise ratio was significantly higher in the Br76 images (13.34 ± 3.47), than in the other 3 image sets (all P < 0.001). The Br76 images demonstrated the highest contrast-to-noise ratio among all reconstructions (1.54 ± 0.86, all P < 0.001). Subjective image sharpness, image noise, overall image quality, and airway detail were best in the Br76 images (all P < 0.001 to P < 0.01, for both readers). Conclusions The use of the Br76 reconstruction kernel provided the best quantitative and qualitative image quality for UHR PCD-CT of the lungs.