Multi-contrast K-edge imaging on a bench-top photon-counting CT system: Acquisition parameter study

TL;DR

This study explores how varying experimental parameters on a bench-top photon-counting CT system can optimize the separation and quantification of multiple contrast agents using K-edge imaging, providing practical guidelines for system configuration.

Contribution

It demonstrates the impact of filter type, projection time, and energy bin width on contrast agent separation and quantification in a PCCT system, offering tailored recommendations for different agents.

Findings

Successful quantification of contrast agents via K-edge subtraction.

Decreasing projection time reduces K-edge CNR, but not proportionally.

Filter type and bin width effects depend on the specific contrast agent.

Abstract

Purpose: Photon-counting computed tomography (PCCT) shows promise for medical imaging in regards to material separation and imaging of multiple contrast agents. However, many PCCT setups are under development and are not optimized for specific contrast agents or use cases. Here, we demonstrate how experimental system parameters may be varied in order to enhance performance and we propose a set of recommendations to achieve this based on contrast agent. Approach: A table-top PCCT system with a cadmium zinc telluride (CZT) detector capable of separating six energy bins was used to image multiple contrast agents in a small phantom. The contrast agents were separated and the concentration was quantified using K-edge subtraction. To increase system performance, we investigated three parameters: beam filter type and thickness, projection acquisition time, and energy bin width. The results from the parameters were compared based on PCCT signal and contrast to noise ratio (CNR) or noise in K-edge images. The concentrations of the contrast agents were quantified in K-edge images and compared to known concentrations. Results: The bench-top PCCT system was able to successfully quantify the contrast agents through K-edge subtraction. Decreasing projection acquisition time showed a decrease in K-edge CNR. However, it did not scale as the square root of time. Filter type and bin width demonstrated a dependence on the specific contrast agent. Conclusions: The presented bench-top system demonstrated the ability to separate contrast agents using K-edge subtraction and accurately determine contrast concentration in K-edge images. Specific parameters for future use will be chosen based on contrast agent.