BACKGROUND The use of computed tomography pulmonary angiography (CTPA) has been increased during the last decade. OBJECTIVE We studied the adherence to current diagnostic recommendations for evaluation of pulmonary embolism in a teaching hospital of Tehran University of Medical Sciences. METHODS The registered medical records (Wells scores and serum D-dimer level) of all patients whose CTPA was performed with suspicion of pulmonary thromboembolism (PTE) were studied retrospectively. Modified Wells score of each patient was determined without being aware of the CTPA results. The patients were categorized to those with a high (likely) clinical probability (score > 4) and low (unlikely) clinical probability (score≤ 4) of PTE. RESULTS During a 6-month period, 82 patients who underwent CTPA were included. The prevalence of PTE was 62.2% in the group of subjects with a likely clinical risk. In 45 (54.8%) of those patients whose CTPA was requested, the PTE was unlikely based on modified Wells criteria. In the clinically unlikely group, serum D-dimer assay was done in 15 out of 45 (33.3%), while it was inappropriately checked in 10 out of 37 (27.0%) with a clinically likely risk. General adherence rate to diagnostic algorithm of PTE was 43.9%. CONCLUSIONS There is still excessive unjustified concern of PTE in less trained physicians leading to excessive diagnostic work-up. Loyalty to the existing guideline for management of suspected PTE in educational hospitals and supervision of attending physicians could prevent overuse of CTPA.