High diagnostic performance of noninvasive fractional flow reserve computed from CT (FFR-CT) was recently reported in prospective multicenter trials. The aims of this study were to evaluate the diagnostic accuracy of FFR-CT in clinical practice and to examine the lesion characteristics showing a mismatch between FFR-CT and invasive FFR. A total of 20 patients (29 vessels) with suspected coronary artery disease were included. All patients underwent invasive coronary angiography and invasive FFR according to coronary artery CT angiography (CCTA) findings. The same raw data used for CCTA were used to evaluate FFR-CT. Results from FFR-CT were compared with invasively measured FFR. A positive ischemia was defined as FFR <0.80. Analyses from three vessels in two patients were not evaluated because of severe calcification or motion artifacts. The diagnostic accuracy, sensitivity, and specificity of FFR-CT per-vessel basis were 81, 100, and 69 %, respectively. To find the reason for mismatch in positive ischemia, lesion characteristics determined with CCTA were compared between the matched group and the mismatched group. A significant difference in bifurcation lesions with positive remodeling was observed between the matched group and the mismatched group (p < 0.01). The high sensitivity of FFR-CT may provide an additional support to the use of CCTA, although particular attention should be paid when using FFR-CT in bifurcation lesions with positive remodeling.