K-Edge Imaging Using a Clinical Dual-Source Photon-Counting CT System

Abstract

Purpose: To evaluate the feasibility and performance of K-edge imaging of iodine (I) and gadolinium (Gd) on a clinically available photon-counting computed tomography (PCCT) system.

Methods: A dual-source clinical PCCT scanner with four energy thresholds (20, 55, 72, 90 keV) was used to scan phantoms containing pure and mixed solutions of I and Gd across multiple concentrations (1-10 mg/mL) and radiation doses (1-8 mGy). Multi-material decomposition was performed using a calibration-based, image-domain algorithm to generate material-specific maps. Quantitative accuracy was assessed using Bland-Altman analysis and contrast-to-noise ratio (CNR), while noise and bias trends were statistically analyzed using non-parametric tests.

Results: K-edge imaging was successfully achieved on a clinical PCCT system with accurate decomposition of I and Gd across varying concentrations, solution types (pure/mixed), and dose levels. Quantitative bias was significantly influenced by radiation dose, concentration, and solution type (p < 0.0004). Increased radiation dose and contrast concentration improved quantification accuracy, with maximum bias reductions of 0.9 (I) and 0.3 mg/mL (Gd). CNR correlated linearly with concentration (R² > 0.99) and moderately with dose (R² = 0.85-0.94), achieving peak values of 13 (I) and 16 (Gd) at 8 mGy. Mixed solutions showed reduced performance compared to pure solutions, i.e., CNR of 5 mg/mL Gd solutions increased by 0.6 per mGy in pure solutions while by 0.5 per mGy in mixtures. Noise was dependent on dose but not on concentration or solution type.

Conclusion: This study establishes the feasibility of K-edge imaging using a clinical PCCT system and demonstrates accurate, simultaneous decomposition of I and Gd in pure and mixed solutions. These findings support the clinical translation of K-edge imaging and highlight PCCT's potential for advanced dual-contrast and molecular imaging applications.

Keywords: Dual-contrast imaging; K-edge imaging; Material decomposition; Photon-counting CT.

Figure 1.. Calibration and experimental setup.

The calibration phantom (A), containing a syringe filled with the calibration material (water, I, or Gd) and water-mimicking inserts, was scanned with a dual-source photon-counting CT. Additional 10 cm extension rings (B) were added after completion of each set of calibration scans. To evaluate the feasibility and accuracy of K-edge imaging, syringes filled with contrast agent solutions were placed within a thoracic phantom and scanned (C).

Figure 2.. Material-specific images of gadolinium and iodine using K-edge imaging.

Inserts containing varying concentrations of I and Gd were placed within a thoracic phantom and scanned. Conventional images (A) display a standard grayscale CT attenuation image (WL/WW: 200/600 HU), illustrating the inability to separate the two contrast agents. Iodine maps (B) (WL/WW: 2.5/5 mg/mL) show successful localization of iodine within inserts, with gadolinium maps (C) (WL/WW: 2.5/5 mg/mL) also achieving successful separation of gadolinium and iodine within the same volume. An overlay image (D) highlights the spatial distribution of both contrast agents, enabling enhanced visualization of co-localized and distinct regions of each contrast agent.

Figure 3.. Bland-Altman plot of iodine and gadolinium material maps at each radiation dose.

Material decomposition of I (A) and Gd (B) demonstrated stable quantification across different concentrations, types of solutions, and radiation doses for all parameters. Solid lines represent the mean difference between measured and expected concentration, whereas dashed lines represent the 95^th percentile confidence interval.

Figure 4.. Contrast-to-noise ratio of iodine and gadolinium in pure solutions at different concentrations and radiation doses.

Higher concentration of material and radiation dose improved contrast in I (A) and Gd (B) material maps. CNR presented an approximately ten-fold increase between the lowest and highest concentration, and a two-fold increase between the lowest and highest radiation dose.

Figure 5.. Effect of radiation dose and contrast agent concentration on contrast of material decomposition maps.

I (A) and Gd (B) CNR were primarily influenced by contrast agent concentration, while radiation dose rendered a smaller effect on CNR. CNR was higher for mixed solutions where the concentration of one CA corresponding to the material map was greater than the other CA, i.e., CNR at 8 mGy of 5:2.5 mg/mL I:Gd is 7.9 while the 5:5 solution is lower at 7.1.