Contrast and quantum noise in single-exposure dual-energy thoracic imaging with photon-counting x-ray detectors

Jeffrey Dhari,Jesse Tanguay

DOI: https://doi.org/10.1088/1361-6560/ad75df

IF: 3.5

2024-09-03

Physics in Medicine and Biology

Abstract:Objective: Photon-counting x-ray detectors (PCDs) can produce dual-energy (DE) x-ray images of lung cancer in a single x-ray exposure. This study quantifies the dependence of contrast-to-noise ratio (CNR) on tube voltage, energy threshold and patient thickness in single exposure, DE, bone-suppressed thoracic imaging with PCDs, and elucidates how the processes inherent in x-ray detection by PCDs contribute to CNR degradation. Approach: We modeled the DE CNR for five theoretical PCDs, ranging from an ideal PCD that detects every photon in the correct energy bin and rejects scatter, to a non-ideal PCD that suffers from charge-sharing and electronic noise, and detects scatter. Model predictions were compared with experimental data from images acquired using a CdTe PCD. The imaging phantom simulated attenuation, scatter and contrast in lung nodule imaging. We quantified CNR improvements achievable with anti-correlated noise reduction (ACNR) and measured the range of exposure rates where pulse pile-up is negligible. Main Results: At the optimal energy thresholds, the modeled CNR with and without ACNR was within 10% of the experimental CNR. CNR improvements with ACNR were approximately five-fold. CNR increased <20% when increasing the tube voltage from 90 kV to 130 kV. Optimal energy thresholds ranged from 50 keV to 70 keV across all tube voltages and patient thicknesses with and without ACNR. Charge sharing and scatter had the largest effect on CNR, degrading it by ∼30% and ∼15% respectively. Dead-time losses were less than 5%. Significance: We (1) employed analytical and computational models to assess the impact of different factors on CNR in single-exposure DE imaging with PCDs, (2) evaluated the accuracy of these models in predicting experimental trends, and (3) quantified improvements in CNR achievable through ACNR. To the best of our knowledge, this study represents the first systematic investigation of single-exposure DE imaging of lung nodules with PCDs.

engineering, biomedical,radiology, nuclear medicine & medical imaging