Simulating and correcting the pileup effect in deep-silicon photon-counting CT.

Area of Science:

• Medical Imaging Physics
• Computational Modeling
• Detector Technology

Background:

• Photon-counting computed tomography (CT) offers enhanced spectral and spatial resolution.
• Pulse pileup in photon-counting detectors causes count loss and spectral distortion, impacting image quality.
• Deep-silicon detectors with segmented pixels aim to mitigate pileup by optimizing count rates.

Purpose of the Study:

• To develop and validate a digital twin framework for deep-silicon photon-counting CT systems.
• To assess the framework's accuracy in simulating pulse pileup effects by comparing simulation to prototype measurements.
• To investigate the impact of pileup on image quality and evaluate a data-driven pileup correction algorithm.

Main Methods:

• Integrated a semi-nonparalyzable detector pileup model into the CatSim simulation environment.
• Validated the simulation framework against measured count data from a deep-silicon photon-counting CT prototype.
• Incorporated a typical image chain including material decomposition and pileup correction for phantom image generation.

Main Results:

• The pileup model demonstrated high accuracy, with deviations below 5% in count rate and variance compared to measurements.
• The pileup correction algorithm effectively suppressed artifacts below noise levels in monochromatic and material images.
• Iodine bias was significantly reduced from 26% to 2% without compromising contrast-to-noise ratio (CNR).

Conclusions:

• The validated digital twin framework reliably represents pulse pileup effects in deep-silicon photon-counting CT.
• The pileup correction algorithm shows strong performance, potentially reducing the need for recalibration in modulated scans.
• Future work includes optimizing simulation speed and expanding the framework to include other detector effects.