QUANTIFICATION OF CORONARY FLOW VELOCITY VIA CONTRAST DISPERSION PATTERNS: INSIGHTS FROM COMPUTATIONAL MODELING AND COMPUTED TOMOGRAPHY EXPERIMENTS
Johns Hopkins University
Advances in multi-detector cardiac computed tomography (CT) have expanded its use beyond coronary atherosclerosis to a suite of functional myocardial imaging options that now closely parallels magnetic resonance imaging; including ventricular function, viability and perfusion. Despite these advances, there are currently no existing CT based methods to assess coronary luminal blood flow/hemodynamics. Recent studies have shown that CT derived axial transluminal contrast gradients (TCG) are greater in coronary arteries with atherosclerotic lesions when compared with normal arteries; suggesting TCG may be related to local coronary hemodynamics. Despite this provocative observation, the basic mechanisms responsible for TCG and their possible connection with coronary hemodynamics have not been explained. In the current work, we hypothesize that TCG is related to the temporal gradients of the contrast bolus and that TCG encodes coronary flow velocity. An analytical relationship between spatial (TCG) and temporal measurements of contrast dispersion is proposed and this allows for estimation of coronary flow velocity from TCG. This is a novel method (called transluminal attenuation flow encoding-TAFE) integrates: a) anatomic features of the coronary vessels, b) TCG and c) temporal gradients in contrast associated with the arterial input function (AIF) that are readily available in conventional CT to allow non-invasive CT derived coronary flow quantification. The TAFE formulation is validated in computational models as well as in CT-compatible experimental phantom studies with configurations that mimic coronary vessels. The experimental studies revealed factors that were absent in computational modeling including imaging artifacts and imaging reconstruction kernels where by imaging analysis TAFE has been modified. In addition, computational simulations of the aortic arch including a semi-patient specific aortic valve model were performed to study contrast dispersion through the arch. This study was done to assess a key assumption in TAFE, that the clinically available AIF at the descending aorta can be used as an accurate estimate of the AIF at the coronary ostium.. The work provides support for the ability of TAFE to provide quantitative estimates of coronary flow velocity but also reveals a number of issues that require further assessment for improved accuracy of TAFE.
CT angiography, transluminal attenuation gradient, Computational fluid dynamics, Imaging diagnosis, Aortic Arch, Coronary arteries, patient specific modeling