In-line spectroscopy for iodine quantification in dynamic contrast-enhanced dedicated breast CT

Abstract

Background: In 4D dynamic contrast-enhanced dedicated breast computed tomography (4D DCE-bCT), the functional properties of the breast will be characterized by monitoring the uptake and washout of iodine-based contrast agents over time. This information could be valuable in breast cancer treatment. However, prior to clinical implementation, it is crucial to validate the quantitative estimates of iodine concentrations at each time point during acquisition.

Purpose: To develop an in-line spectroscopy system capable of measuring iodine concentrations in a dynamic x-ray breast phantom in real-time.

Methods: Potassium iodide served as the contrast agent. The system was set-up at both the entrance and exit of the phantom. It comprises a fiber-coupled green LED and collimator, which together ensure that a parallel beam passes through the sample holder. Transmitted light is captured by a collimator on the opposite side and directed through a fiber optic cable to a photodetector for intensity measurement. The relationship between 13 iodine concentrations (0-6 mg I/mL) and light transmission was tested, and the system's repeatability and accuracy were determined.

Results: The system exhibited a strong correlation between iodine concentration and transmission values, achieving a root-mean-square error of 0.007. The repeated measurements had relative standard deviations of 0.04% and 0.1% for repeated water measurements at the phantom's entrance and exit, respectively. Furthermore, the accuracy measurements gave a mean error of fitting of 0.008 (± 0.07) mg I/mL.

Conclusion: The in-line spectroscopy system can effectively monitor iodine concentrations in a dynamic breast phantom, providing a reliable method for quantitative validation of 4D DCE-bCT.

Keywords: 4D DCE‐BCT; breast cancer; breast phantom; iodine quantification; spectroscopy.

FIGURE 1

In‐line spectroscopy setup. Schematic overview of the setup (top). The sample holder is assembled between collimators (bottom left [1] and [4]). Light is directed from the collimator indicated with [1] towards the sample holder [2 and 3] and captured on the other side [4]. Fluid flow is indicated with black arrows. The square tube for fluid flow is indicated with [2]. The sample holder is shown on the bottom right.

FIGURE 2

Setup to investigate the system's ability to detect a dynamic change in iodine concentration in the breast CT phantom. Schematic representation (top) and build on top of the breast CT system (bottom). The green arrows indicate fluid flow.

FIGURE 3

Absorption coefficients for different iodide concentrations at different wavelengths measured with the UV‐Vis spectrophotometer. The photometric accuracy is 0.004 A/cm. This uncertainty is smaller than the size of the symbols.

FIGURE 4

Transmission values per iodine concentration, for the setup at the phantom entrance and exit. The average relative standard deviation of the transmission values at the phantom entrance is 1.13% and at the phantom exit 0.84%. This uncertainty is smaller than the size of the symbols. Exponential curves were fitted to the data with coefficients (with 95% confidence intervals) of: a = 0.992 (± 0.0139) and b = −0.8353 mL mg/I (± 0.0200 mL mg/I) for the entrance curve and a = 0.992 (± 0.0133) and b = −0.7912 mL mg/I [± 0.0181 mL mg I⁻¹)] for the exit curve and RMSEs of 0.0072 and 0.0070.

FIGURE 5

Assessment of accuracy. The real value of the concentration is compared to the computed concentration obtained from the fit to a concentration‐transmission calibration curve for 26 data points (left). A Bland‐Altman plot is made in which the error in the fitted concentration is calculated by comparing the concentration obtained from a concentration‐transmission calibration curve to the actual concentration for 26 data points (right).

FIGURE 6

Expected iodine concentration over time and measured with the optical system at the phantom entrance and exit.