PubMed1991
The quantification of myocardial perfusion abnormalities is necessary to allow comparison of repeated studies, especially in the evaluation of the success of medical, interventional or combined treatment in stable coronary artery disease or in evolving myocardial infarction. The purpose of this study was to assess inter-observer reproducibility of tomographic study processing using a semi-automatic quantitative programme. Technetium 99m hexakis-2-methoxyisobutylisonitrile (99mTc-Sestamibi) was chosen for tomographic imaging of repeated rest-stress studies in patients with stable coronary artery disease. The quantification was performed using a modification of the Cedars polar coding and comparison with the normal data base. The perfusion defects were quantified separately for each standard perfusion area [left anterior descending (LAD), right coronary (RCA) and left circumflex (LCX) arteries] and total area of hypoperfused myocardium. The inter-observer variability for 40 tomographic studies was accomplished. The defects were the largest in the LAD perfusion area (average 19.7% of the normalized LAD supply area) with an inter-observer correlation of 0.84 for this region. The greatest variability was found for the LCX region (r = 0.55) and is attributed to a small average perfusion defect (7.1%), only 18 studies having abnormal perfusion in this area. In total, an average 14.3% of the left ventricular myocardium was significantly hypoperfused, and the inter-observer correlation was 0.87. These results show good inter-observer reproducibility using semi-automatic quantitation of perfusion defects. Careful interpretation of smaller defects in the evaluation of treatment results is advised when repeated 99mTc-Sestamibi single photon emission tomography studies are processed by more than one observer.