Fifty-two consecutive symptomatic patients with a low to intermediate likelihood of coronary artery disease, and who were referred for MPS, also underwent CTA. The CTA datasets were analyzed by two experienced observers who were blinded to the MPS data, and coronary artery segments were reported as < 50 % , 50 % to 69 % , 70 % to 99 % stenoses, or occluded. The MPS images were similarly analyzed for inducible perfusion abnormalities, and coronary territories were identified. At the patient level, agreement between CTA and MPS for CTA lesions at ≥ 50 % was 87 % (sensitivity, 100 % ; specificity, 84 % ; positive predictive value, 50 % ; negative predictive value, 100 % ). For CTA lesions, agreement at ≥ 70 % was 96 % (sensitivity, 86 % ; specificity, 98 % ; positive predictive value, 86 % ; negative predictive value, 98 % ).
In patients with a low to intermediate likelihood of CAD, there is good correlation between MPS and CTA for the detection of functionally significant coronary artery stenoses when CTA detects a narrowing of ≥ 70 % severity. Computed tomography coronary angiography stenoses of 70 % should be used to determine functional significance, and not 50 % , as is the usual practice at present.