Empirical correction decomposition method (ECDM): enhancing accuracy of quantitative measurement in spectral CT

Abstract Objective. Spectral CT and material decomposition methods are crucial for precise material identification and quantitative composition analysis in preclinical research and clinical diagnosis. The empirical material decomposition method is widely used for its straightforward modeling approach, independence from spectral and detector response knowledge, and operational convenience. However, this method has limited decomposition accuracy and its precision depends on the choice of calibration phantoms. Approach. To address these issues, we propose an empirical correction decomposition method (ECDM). The innovation of this method lies in its ability to conveniently estimate and correct empirical decomposition errors using a specially designed calibration phantom. First, the specially designed calibration phantom for ECDM undergoes empirical decomposition initially to establish the relationship between decomposition errors and decomposition values. Then, ECDM estimates and corrects the error of empirical decomposition values. Main results. In the phantom experiments, ECDM improves the decomposition accuracy of empirical methods, effectively reducing the different decomposition errors caused by four different sizes of calibration phantoms from a maximum of 144\% to within 25\%. In the mouse experiments, ECDM achieves accurate quantification of contrast agents in biological tissues, outperforming the other two methods. The absolute error percentages of ECDM in the decomposition results of the two standard iodine solutions are both less than 5\%. Significance. ECDM significantly improves decomposition accuracy and reduces the impact of the size of the empirical calibration phantom. Overall, our method based on spectral CT is very convenient and practical for the quantitative measurement in biomedical applications.