Physical characterization of a new CT iterative reconstruction method operating in sinogram space

Area of Science:

• Medical Imaging
• Radiology
• Image Reconstruction

Background:

• Siemens introduced SAFIRE, a novel iterative reconstruction algorithm.
• SAFIRE operates in the raw data domain, focusing on noise reduction with adjustable strengths.

Purpose of the Study:

• Evaluate the impact of SAFIRE on CT image quality.
• Compare SAFIRE's performance against standard filtered back projection (FBP).
• Assess the influence of SAFIRE's strength levels on image parameters.

Main Methods:

• Utilized selected phantoms for quantitative analysis.
• Evaluated image noise, noise power spectrum (NPS), contrast-to-noise ratio (CNR), spatial resolution, and CT number accuracy.
• Compared SAFIRE reconstructions with FBP.

Main Results:

• SAFIRE effectively reduces image noise, independent of reconstruction kernel, and is strength-dependent.
• NPS shifts towards lower frequencies with increasing SAFIRE strength.
• CNR consistently improves with SAFIRE, increasing at higher strengths.
• SAFIRE maintains CT number accuracy, linearity, and spatial resolution across dose levels.

Conclusions:

• SAFIRE enables significant noise reduction while preserving essential image quality parameters.
• The algorithm demonstrates potential for effective dose containment in clinical CT imaging.