With the aim of evaluating the efficiency of our diagnostic approach in patients with suspected acute pulmonary embolism (PE), we prospectively studied 143 patients investigated by means of a ventilation/perfusion (V/Q) lung scan. A pre-test clinical probability of PE (Pclin) was assigned to all patients by the clinicians and scans were interpreted blinded to clinical assessment. A 2-year follow-up of our patients was systematically performed and possible in 134 cases. Distribution of clinical probabilities was high Pclin in 22.5%, intermediate Pclin in 24% and low Pclin in 53.5%, whereas the distribution of scan categories was high Pscan in 14%, intermediate Pscan in 18%, low Pscan in 57% and normal Pscan in 11%. The final prevalence of PE was 24.5%. High Pscan and normal Pscan were always conclusive (19 and 15 cases respectively). Low Pscan associated with low Pclin could exclude PE in 43/45 cases (96%). None of the patients in whom the diagnosis of PE was discarded had a major event related to PE during the 2-year follow-up. Overall, the combined assessment of clinical and scintigraphic probabilities allowed confirmation or exclusion of PE in 80% of subjects (107/134) and proved to be a valuable tool for selecting patients who needed pulmonary angiography, which was required in 20% of our patients (27/134).