[Effect of different observers on the interpretation of pulmonary perfusion scintigraphy]

The agreement in the interpretation of the scintigraphic images of pulmonary perfusion in the diagnosis of pulmonary tromboembolism (TEP) is not always the most adequate. The purpose of this study was to evaluate the degree of agreement by eight observers in the interpretation of these studies.

The study population consisted of 180 studies with clinically suspected pulmonary embolism referred to our department for scintigraphic perfusion imaging from April 98 to September 99. The patients received an intravenous injection of 111-148 MBq (3-4 mCi) of 99mTc-macroaggregated albumin whereas the images were obtained in the six routine projections. The observers consisted of five nuclear medicine physicians and three residents who independently reviewed the scintigrams (low, intermediate and high probability). In a second step, the physicians performed consensus interpretations. The Kappa statistics was used to evaluate the degree of agreement between individual observer interpretations beyond that expected by chance alone.

The mean value SEM for Kappa index was 0.58 0.02, with maximum and minimum values of 0.76 and 0.27 respectively. The interobserver variability was greater than expected. Except in two observers scans in the interpretation of images in the low and high categories showed minor discrepancies, whereas scans in intermediate probability showed the most important interobserver variations.

The study manifests the importance of a uniform criteria in lung scintigraphic interpretation.