Background:Although stress single-photon emission computed tomography (SPECT) using a cadmium-zinc-telluride (CZT) camera facilitates radiation dose reduction, only a few studies have evaluated its diagnostic accuracy in Japanese patients by applying fractional flow reserve (FFR) measurements.Methods and Results:We prospectively evaluated 102 consecutive patients with suspected or known coronary artery disease with a low-dose stress/rest protocol (^99mTc radiotracer 185/370 MBq) using CZT SPECT. Within 3 months, coronary angiography was performed and a significant stenosis was defined as ≥90% diameter narrowing on visual estimation, or as a lesion of <90% and ≥ 50% stenosis with FFR ≤0.80. To detect individual coronary stenosis, the respective sensitivity, specificity, and accuracy were 86%, 75%, and 82% for left anterior descending artery stenosis, 76%, 81%, and 79% for left circumflex artery stenosis, and 87%, 92%, and 90% for right coronary artery stenosis. When limited to 92 intermediate stenotic lesions in which FFR was measured, stress SPECT showed 77% sensitivity, 91% specificity, and 84% accuracy, whereas the diagnostic value decreased to 52% sensitivity, 68% specificity, and 58% accuracy based only on visual estimation of ≥75% diameter narrowing.Conclusions:CZT SPECT demonstrated a good diagnostic yield in detecting hemodynamically significant coronary stenoses as assessed by FFR, even when using a low-dose ^99mTc protocol with an effective dose ≤5 mSv. (Circ J 2016; 80: 1217–1224)