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

Area of Science:

• Medical Imaging
• Radiology
• Computed Tomography

Background:

• Photon-counting detector CT (PCD-CT) represents a significant advancement in computed tomography (CT) technology.
• The latest generation of PCD-CT systems offers potential improvements in various clinical applications, notably in chest imaging.

Purpose of the Study:

• To evaluate the image quality of ultra-high-resolution (UHR) photon-counting detector CT (PCD-CT) for lung imaging.
• To compare the effectiveness of four sharp reconstruction kernels in optimizing UHR PCD-CT lung image quality.

Main Methods:

• A retrospective analysis of 25 unenhanced chest CT scans acquired using a clinical dual-source PCD-CT scanner in UHR mode.
• Images were reconstructed using four sharp kernels: Bl64, Br76, Br84, and Br96.
• Quantitative analysis of image noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR), alongside subjective assessment of image sharpness, noise, overall quality, and airway details by two radiologists.

Main Results:

• The Br76 kernel demonstrated the lowest image noise and the highest SNR and CNR among all tested kernels.
• Subjective evaluations by radiologists indicated that the Br76 kernel provided the best image sharpness, lowest perceived noise, highest overall image quality, and superior airway detail.
• All quantitative and qualitative improvements with the Br76 kernel were statistically significant (P < 0.001 to P < 0.01).

Conclusions:

• The Br76 reconstruction kernel is optimal for achieving superior quantitative and qualitative image quality in ultra-high-resolution photon-counting detector CT of the lungs.
• Utilizing the Br76 kernel enhances diagnostic confidence and visualization of pulmonary structures in UHR PCD-CT examinations.