Development and Clinical Applications of a Virtual Imaging Framework for Optimizing Photon-counting CT.

Area of Science:

• Medical Imaging Physics
• Computational Imaging
• Radiology

Background:

• Photon-counting CT (PCCT) offers advanced imaging capabilities.
• Accurate simulation tools are crucial for developing and optimizing new CT technologies.
• Existing simulation platforms may not fully capture PCCT-specific physics.

Purpose of the Study:

• To develop a virtual imaging framework simulating a novel PCCT system (NAEOTOM Alpha, Siemens).
• To validate the simulator against experimental data using various imaging parameters and dose levels.
• To assess the framework's utility in clinical applications like COPD quantification and lung nodule radiomics.

Main Methods:

• Adapted the DukeSim platform to model the PCCT prototype's geometry and photon-counting detector physics using Monte Carlo methods.
• Validated the simulator by comparing reconstructed images (at varying doses and kernels) with experimental measurements.
• Quantitatively assessed image quality using metrics like HU values, noise magnitude, NPS, and MTF.
• Applied the framework to simulate COPD quantifications and lung nodule radiomics on computational phantoms.

Main Results:

• The PCCT simulator accurately replicated experimental data with minimal discrepancies in image quality metrics.
• Lung lesion radiomics accuracy improved with reduced pixel size and slice thickness.
• COPD quantification accuracy was enhanced by higher doses, thinner slices, and softer reconstruction kernels.
• The framework demonstrated successful implementation of PCCT acquisition and physics attributes.

Conclusions:

• The developed virtual imaging platform accurately simulates PCCT systems.
• This framework facilitates systematic comparison of new PCCT technologies.
• It enables optimization of imaging parameters for specific clinical tasks, improving diagnostic accuracy and efficiency.