Quantitative evaluation improves specificity of myocardial perfusion SPECT in the assessment of functionally significant intermediate coronary artery stenoses: a comparative study with fractional flow reserve measurements

Objective Myocardial perfusion SPECT (MPS) is a noninvasive method commonly used for assessment of the hemodynamic significance of intermediate coronary stenoses. Fractional flow reserve (FFR) measurement is a well-validated invasive method used for the evaluation of intermediate stenoses. We aimed to determine the association between MPS and FFR findings in intermediate degree stenoses and evaluate the added value of quantification in MPS. Methods Fifty-eight patients who underwent intracoronary pressure measurement in the catheterization laboratory to assess the physiological significance of intermediate (40–70 %) left anterior descending (LAD) artery lesions, and who also underwent stress myocardial perfusion SPECT either for the assessment of an intermediate stenosis or for suspected coronary artery disease were analyzed retrospectively in the study. Quantitative analysis was performed using the 4DMSPECT program, with visual assessment performed by two experienced nuclear medicine physicians blinded to the angiographic findings. Summed stress scores (SSS) and summed difference scores (SDS) in the LAD artery territory according to the 20 segment model were calculated. A summed stress score of ≥3 and an SDS of ≥2 were assumed as pathologic, indicating significance of the lesion; a cutoff value of 0.75 was used to define abnormal FFR. Both visual and quantitative assessment results were compared with FFR using Chi-square (χ²) test. Results The mean time interval between two studies was 13 ± 11 days. FFR was normal in 45 and abnormal in 13 patients. Considering the FFR results as the gold standard method for assessing the significance of the lesion, the sensitivity and specificity of quantitative analysis determining the abnormal flow reserve were 85 and 84 %, respectively, while visual analysis had a sensitivity of 77 % and a specificity of 51 %. There was a good agreement between the observers (κ = 0.856). Summed stress and difference scores demonstrated moderate inverse correlations with FFR values ( r = −0.542, p