Introduction and objectives: The correct interpretation of myocardial perfusion single photon emission computed tomography (SPECT) requires knowledge of the technique reproducibility. The objective was analyze the interobserver correlation of different experience in the interpretation of myocardial perfusion SPECT in patients following acute myocardial infarction (AMI) in order to improve the quality of our site.
Methods: Sixty cases (56 +/- 11 years, 87 % men) with transmural AMI who had recently undergone successful thrombolysis were included. Resting perfusion with (99m)Tc-sestamibi was performed at one week post-AMI.
Analysis: Semiquantitative interpretation using 17 segment-model by 2 independent specialists and 5 observers, was performed blindly. Left ventricular ejection fraction (LVEF) was measured with isotopic ventriculography one month after AMI, with a mean of 38 %.
Results: Using independent and then agreed on perfusion analysis, average involved segments/patient was 9.3 +/- 4 and the sum of severity 25 +/- 13. Readings of other observers ranged from 7 +/- 3.7 to 9.4 +/- 3.9 and 16.7 +/- 9.7 to 24.6 +/- 13, respectively, consistent with the reading of the specialists of between 0.779-0.871 (kappa: 0.565-0.741). There was no significant difference when the number of segments were analyzed in 40 % of the cases and for intensity in 60 % of them in more experienced observers. Correlation with consensus reading for the number of segments ranged from 0.84 to 0.94 and for severity from 0.79 to 0.89. Identification of culprit arteries was acceptable, with r values between 0.612 and 0.683 and kappas between 0.629 and 0.656. Correlation of the number of involved segments and severity with LVEF performed one month after AMI was 0.73 and 0.74, respectively.
Conclusions: There was good correlation in the interpretation of myocardial perfusion SPECT, with a significantly better fit in more experienced observers. This academic exercise was also helpful in improving our residents' skills in cardiology.