Photon Counting Computed Tomography With Dedicated Sharp Convolution Kernels: Tapping the Potential of a New...

Area of Science:

• Medical Imaging
• Radiology
• Computed Tomography

Background:

• Coronary stent imaging is crucial for diagnosing and managing cardiovascular diseases.
• Photon counting detector (PCD) CT offers potential advantages over conventional CT.
• Optimizing image reconstruction for PCD CT is essential for enhancing diagnostic accuracy.

Purpose of the Study:

• To evaluate the effectiveness of a dedicated sharp convolution kernel for PCD CT in coronary stent imaging.
• To assess the impact of iterative reconstruction on image noise when using the sharp kernel.
• To compare the performance of the sharp kernel with conventional kernels.

Main Methods:

• An in vitro phantom simulating coronary artery stenting was used.
• Scans were performed using PCD CT with both conventional and a dedicated sharp convolution kernel.
• Images were reconstructed using filtered back-projection and iterative reconstruction (SAFIRE).
• Subjective and quantitative image quality metrics were assessed by blinded readers.

Main Results:

• The dedicated sharp kernel significantly improved image quality, stent diameter difference, and image sharpness compared to conventional kernels.
• While noise was higher with the sharp kernel, iterative reconstruction reduced it by up to 59%.
• Interreader and intrareader reliability were excellent.

Conclusions:

• A dedicated sharp convolution kernel enhances PCD CT coronary stent imaging.
• Iterative reconstruction effectively mitigates the increased noise associated with the sharp kernel.
• This approach offers superior qualitative and quantitative image characteristics for coronary stent visualization.