Recent advances in SPECT technology including cadmium zinc telluride (CZT) semi-conductor detector material may pave the way for absolute myocardial blood flow (MBF) measurements by SPECT. The aim of the present study was to compare K1 uptake rate constants as surrogates of absolute MBF and myocardial flow reserve index (MFRi) in humans as assessed with a CZT SPECT camera versus positron emission tomography (PET). METHODS Absolute MBF was assessed in 28 consecutive patients undergoing adenosine stress/rest myocardial perfusion imaging (MPI) by (99m)Tc-tetrofosmin CZT SPECT and (13)N-ammonia PET and MFR was calculated as ratio of hyperaemic over resting MBF. Results from both MPI methods were compared and correlation coefficients were calculated. Diagnostic accuracy of CZT MFRi to predict an abnormal MFR defined as PET MFR<2 was assessed using a receiver operator characteristic curve (ROC). RESULTS Median MBF at rest was comparable between CZT and PET (0.89 [0.77-1.00] ml/g/min vs. 0.92 [0.78-1.06] ml/g/min; P = ns) while it was significantly lower at stress in CZT vs PET (1.11 [1.00-1.26] ml/g/min vs 2.06 [1.48-2.56] ml/g/min; p<0.001). This resulted in median MFRi values of 1.32 [1.13-1.52] by CZT and 2.36 [1.57-2.71] by PET (p<0.001). ROC revealed a cut-off for CZT MFRi at 1.26 to predict an abnormal PET MFR yielding an accuracy of 75%. CONCLUSION Estimation of absolute MBFi values by CZT SPECT MPI with (99m)Tc-tetrofosmin is technically feasible, although hyperemic values are significantly lower than from PET with (13)N-ammonia resulting in a substantial underestimation of MFR. Nevertheless, CZT MFRi may confer diagnostic value.