Aim. To evaluate the informative value of a 3D physicomathematical model of local hemodynamics based on coronary computed tomography angiography (CCTA) data for noninvasive determination of the functional significance of coronary artery (CA) stenosis. Material and methods . The study included 15 patients (mean age 57,6±8,9 years) with stable coronary artery disease (CAD). All patients underwent CCTA, and 9 patients underwent myocardial perfusion scintigraphy (MPS). Based on the CCTA images, 3D coronary artery models were constructed, in which fractional flow reserve (CT-FFR) was calculated at rest and during simulated hyperemia using COMSOL Multiphysics®. The modeling results were compared with the stenosis degree and ischemia signs based on the MPS data. Results . A total of 76 coronary vascular segments were analyzed. CT-FFR in simulated hyperemia values were significantly lower (p<0,001) in segments with stenosis ≥50% (0,53 [0,37-0,65]) and ≥70% (0,47 [0,34-0,55]) compared to non-obstructive lesions <50% (0,88 [0,83-0,96]) and <70% (0,87 [0,81-0,95]). A negative correlation was found between the stenosis degree and CT-FFR in simulated hyperemia (ρ=-0,485, p<0,05). ROC analysis demonstrated high diagnostic accuracy of CT-FFR in simulated hyperemia for stenoses ≥50% (AUC=0,9; sensitivity 93,7%; specificity 86,7%) and ≥70% (AUC=0,9; sensitivity 100%; specificity 83,6%). In vascular regions with perfusion defects according to MPS data, CT-FFR in simulated hyperemia were significantly lower (0,49±0,21 vs 0,82±0,18, p<0,001). Logistic analysis showed that a decrease in CTFFR in simulated hyperemia ≤0,68 was associated with an increased odds of regional ischemia (odds ratio 0,497; 95% confidence interval 0,33-0,73; p<0,001). Conclusion. The pilot study results confirm that the three-dimensional CT-FFR model, based on CCTA, is highly informative for assessing the anatomical and functional significance of coronary artery stenoses and can be used for non-invasive diagnosis of myocardial ischemia.