Estimation of cardiac output from test-bolus during coronary computed tomography angiography: a personalized mathematical model

Abstract

Abstract Background Contrast enhancement is measured in the aorta at set intervals of time during the test-bolus series of a coronary computed tomography angiography [CCTA]. Its primary purpose is to ascertain the timing delay of the main scan after contrast administration to ensure optimal image acquisition. The change in contrast enhancement over time correlates directly with contrast medium concentration in the vessel. If the complete time density curve [TDC] is known and corrected for contrast enhancement due to recirculation it can be used to estimate the cardiac output [CO], by e.g. a modified Stewart-Hamilton equation, as shown by Mahnken et al. Purpose The current study aims to explore ways to improve the mathematical estimation of CO. In practical applications only part of the TDC is known. By integrating patient factors, contrast factors and scanning factors with a mathematical compartment model, the CO can be estimated. Estimated CO is subsequently used to simulate the distribution of contrast medium in the body, enabling prospective adjustment of contrast medium administration protocol that could be tailored to each individual, utilizing the lowest possible dose of contrast to achieve diagnostically acceptable CCTA images. Methods Test-bolus images were acquired sequentially at two second intervals over the aorta from 40 patients with a clinical indication for CCTA. According to clinical routines, premedication with beta-blockers was performed in 35 patients and sublingual nitroglycerine in 39 patients. Attenuation in the aorta was measured by a centrally placed region of interest with a width of half the aortic diameter. Each participant underwent cardiac magnetic resonance imaging [MRI] immediately after CCTA, from which actual CO was measured as a method of validation. A mathematical compartment model for distribution of contrast, outlined by Bae et al. was improved and used to estimate the CO from the TDC in the aorta during the test-bolus. The estimated CO [COEst] was compared with CO calculated from MRI [COMR], and estimated attenuation in the aorta [HUEst] was compared with measured attenuation in the aorta during test-bolus in CCTA [HUCT]. Results One patient was excluded due to unacceptable cardiac MRI. Our model showed a strong correlation between COEst and COMR for all patients (r=0.62, p<0.001) and a higher correlation for women than men when separated by gender (men r=0.58, p=0.006; women r=0.63, p=0.005). Women had an average COEst 1.92 L/min significantly lower than men (p<0.001), despite no statistical difference in COMR between men and women. Conclusion An improved mathematical model for estimation of CO can be a valuable tool to improve individualized contrast administration protocols and increase the diagnostic information extracted from a CCTA. In addition to known physical factors it is clear that gender and weight, including their derivatives, must play a more significant role in estimation. Funding Acknowledgement Type of funding source: Public hospital(s). Main funding source(s): Stavanger University Hospital