Approximate Cramér-Rao lower bound analyses for semi-optimal energy thresholds selection in photon counting CT

Area of Science:

• Medical Imaging
• Photon Counting CT
• Image Reconstruction

Background:

• Energy threshold selection in Photon Counting CT (PCCT) is critical for minimizing noise in material decomposition.
• Theoretical optimization using Cramér-Rao lower bound (CRLB) is often impractical due to the need for complete system models.

Purpose of the Study:

• To develop practical methods for selecting semi-optimal energy thresholds in PCCT that do not require system models.
• To enable effective noise reduction in material decomposition without complex theoretical analysis.

Main Methods:

• Proposed approximate linear and nonlinear CRLB methods using look-up tables generated from threshold scans of reference phantoms.
• Introduced a 'total noise' metric to guide energy threshold selection for minimizing noise in basis line integrals.
• Validated methods via simulations (2- and 3-material decomposition) and experiments (2-material decomposition) using PcTK toolbox and a bench-top PCCT system.

Main Results:

• Simulations showed proposed methods yielded lower or similar noise compared to the equally distributed counts (EDC) condition for three-material decomposition.
• The nonlinear approach achieved optimal noise levels in simulations for four energy bins.
• Experiments demonstrated near-optimal performance for the linear method, outperforming EDC by 15-18% for specific materials.

Conclusions:

• The proposed energy threshold selection methods are practical, computationally efficient, and effective for PCCT.
• These methods outperform the EDC condition in both simulations and experiments, offering near-optimal noise reduction.
• The approaches require only a simple threshold scan and are broadly applicable across various decomposition scenarios.