OBJECTIVE To evaluate the accuracy of dual-source CT coronary angiography (DSCTCA) for the depiction of functionally relevant coronary artery lesion(FRCAL), by using myocardial perfusion imaging (MPI) with single photon emission computed tomography (SPECT). METHODS DSCTCA, (99)Tc(m)-MIBI SPECT myocardial perfusion imaging (MPI) and conventional coronary angiography (CCA) were performed in 59 patients with clinical suspected CAD. Coronary artery diameter narrowing of 50% or greater at DSCTCA was defined as stenosis and was compared with MPI findings. CCA was served as a reference standard for DSCTCA. RESULTS (1) Agreement between DSCTCA and CCA was good (kappa = 0.93 for patient-based analysis, Kappa = 0.88 for vessel-based analysis). (2) DSCTCA revealed stenoses in 86 segments corresponding to 60 arteries in 34 patients. (3) MPI revealed 19 reversible, 21 partially reversible, and 5 fixed defects in 25 patients. (4) About 65.0% (39/60) of all the narrowed coronary arteries were determined to be FRCAL. Sensitivity, specificity, accuracy, positive predictive values and negative predictive values, respectively, of DSCTCA in the detection of all MPI defects were 92.0%, 67.6%, 78.0%, 67.6% and 92.0% on a per-patient basis and 86.7%, 89.0%, 88.6%, 65.0% and 96.6% on a per-artery basis. (5) ROC analysis showed that predictive value of DSCTCA in FRCAL was similar with those of CCA (AUCs = 0.80, 0.82). CONCLUSIONS DSCTCA can evaluate FRCAL indirectly. When DSCTCA results are negative, it can help ruled out patients with FRCAL. The positive DSCTCA results should combine MPI in predictor of myocardial ischemia.