Quantitative Enhancement of Spectral Imaging in Photon-Counting CT via Dynamic Energy Thresholding

Area of Science:

• Medical Imaging
• Radiological Physics

Background:

• Photon-counting computed tomography (PCCT) offers superior material separation and quantitation compared to conventional spectral systems.
• Imaging performance in PCCT is critically dependent on energy thresholds, which are often fixed in current systems.
• Optimal energy thresholds can vary based on diagnostic tasks and patient anatomy, highlighting a limitation of fixed thresholds.

Purpose of the Study:

• To investigate the potential of dynamic energy threshold modulation in PCCT for optimizing spectral imaging performance.
• To determine optimal dynamic thresholds by leveraging prior anatomical information or low-dose scout scans.
• To assess improvements in dual-energy material decomposition using dynamic thresholding.

Main Methods:

• Simulations using two-bin photon-counting detectors (PCDs) under various conditions.
• Analysis of single-pixel decomposition across different material thicknesses and photon counts.
• Decomposition of a digital pelvic phantom with iodine enhancement to evaluate pixel-wise threshold modulation.

Main Results:

• Optimal energy thresholds for water-calcium decomposition shifted between 35 and 75 keV.
• Replacing fixed thresholds with optimal ones reduced Root Mean Square Error (RMSE) by over 30%, particularly in low-photon count scenarios.
• Pixel-wise threshold modulation in pelvic CT simulations resulted in 6.9% and 17.2% RMSE reductions for water and iodine, respectively, and decreased streak artifacts.

Conclusions:

• Dynamic threshold modulation in PCCT can substantially enhance spectral imaging accuracy.
• This approach offers significant benefits for low-dose protocols and imaging of complex anatomical regions.
• Integrating dynamic thresholding into PCCT systems holds promise for improved diagnostic performance.