We have demonstrated the ability to identify coronary calcium, a reliable biomarker of coronary artery disease, using nongated, 2-shot, dual energy (DE) chest x-ray imaging. Here we will use digital simulations, backed up by measurements, to characterize DE calcium signals and the role of potential confounds such as beam hardening, x-ray scatter, cardiac motion, and pulmonary artery pulsation. For the DE calcium signal, we will consider quantification, as compared to CT calcium score, and visualization.Methods:
We created stylized and anatomical digital 3D phantoms including heart, lung, coronary calcium, spine, ribs, pulmonary artery, and adipose. We simulated high and low kVp x-ray acquisitions with x-ray spectra, energy dependent attenuation, scatter, ideal detector, and automatic exposure control (AEC). Phantoms allowed us to vary adipose thickness, cardiac motion, etc. We used specialized dual energy coronary calcium (DECC) processing that includes corrections for scatter and beam hardening.Results:
Beam hardening over a wide range of adipose thickness (0–30 cm) reduced the change in intensity of a coronary artery calcification (ΔICAC) by < 3% in DECC images. Scatter correction errors of ±50% affected the calcium signal (ΔICAC) in DECC images ±9%. If a simulated pulmonary artery fills with blood between exposures, it can give rise to a residual signal in DECC images, explaining pulmonary artery visibility in some clinical images. Residual misregistration can be mostly compensated by integrating signals in an enlarged region encompassing registration artifacts. DECC calcium score compared favorably to CT mass and volume scores over a number of phantom perturbations.Conclusion:
Simulations indicate that proper DECC processing can faithfully recover coronary calcium signals. Beam hardening, errors in scatter estimation, cardiac motion, calcium residual misregistration etc., are all manageable. Simulations are valuable as we continue to optimize DE coronary calcium image processing and quantitative analysis.