Voxelwise characterization of noise for a clinical photon-counting CT scanner with a model-based iterative...

Area of Science:

• Medical Imaging
• Radiology
• Detector Technology

Background:

• Photon-counting detector (PCD) technology offers potential for reduced noise in computed tomography (CT).
• Voxelwise noise characterization is crucial for evaluating new CT technologies.

Purpose of the Study:

• To perform a voxelwise noise characterization of a clinical PCD-CT scanner.
• To evaluate the impact of a model-based iterative reconstruction algorithm (QIR) on image noise.

Main Methods:

• Repeated axial acquisitions of water and Catphan phantoms using PCD-CT and conventional energy-integrating detector (EID) CT.
• Noise maps, non-uniformity index (NUI), noise histograms, and noise power spectrum (NPS) were computed for filtered back projection (FBP) and iterative reconstructions.

Main Results:

• PCD-CT demonstrated significantly lower mean noise and NUI compared to EID-CT.
• Noise reduction increased with iterative power, reaching up to 72.5% for PCD-CT.
• QIR algorithm reduced noise without altering histogram shape and with limited NPS shift.

Conclusions:

• PCD-CT substantially reduces image noise and improves spatial uniformity.
• The QIR algorithm effectively decreases noise in PCD-CT while preserving noise texture.
• This study supports PCD technology's potential for reduced radiation exposure in CT.