Introduction: The diagnosis of pulmonary embolism (PE) is relevant in therapeutic approach and patient prognosis. Ventilation perfusion scan (VQ) is a well established method in the evaluation of PE. The interobserver agreement could be variable and it is necessary to diminish the discordances within a group.
Objective: To determine interobserver agreement in our center and to correlate the blind report with the original one (with clinical and radiological information available).
Method: 401 studies from 382 patients with a possible PE were analyzed retrospectively; a blind lecture was performed by 6 independent observers with different experience level. Interpretation results were high, low or intermediate probability for PE, based on the individual experience and revised PIOPED criteria. Weighted kappa index was applied.
Results: The original reports included 27.2% high probability, 5.5% intermediate and 67.3% low probability, almost normal or normal. The interobserver agreement ranged between 72.6 and 86% (k: 0.582-0.743) and the correlation with the report ranged between 74.3 and 81.8% (k: 0.582- 0.675). There was greater agreement between observers with longer experience.
Conclusion: In our center there is an excellent interobserver concordance with good kappa index in VQ scans required to ruled out PE. This academic task was a practical training for the residents in order to accomplish self confidence in the interpretation.
