The diagnosis of pulmonary embolism is challenging, and autoptic series have demonstrated that a high percentage of cases are not recognized ante-mortem. A number of predisposing factors, symptoms and signs associated with pulmonary embolism have been recognized, and should be used to raise the suspicion of the disease. These include immobilization, recent surgery, active cancer, previous thromboembolism, syncope, dyspnoea, chest pain, haemoptysis, signs of deep vein thrombosis, hypocarbic hypoxemia. Once pulmonary embolism is suspected, the clinical probability of the disease should be assessed; to this end, three clinical rules have been proposed and validated (the revised Geneva score, the Wells score and the PISA-PED score) while others await clinical validation. In case of low clinical probability, a negative a D-dimer test is sufficient to rule out the diagnosis, while if the clinical probability is high, or the D-dimer test is positive, further tests are necessary. Computer tomography angiography or perfusion lung scan are the imaging tests of choice, depending on local availability and experience. If the clinical probability and the results of the imaging test are concordant, a definitive diagnosis can be obtained; if the results are discordant, further testing is necessary. In particular, in the specific case of a small clot (i.e. segmental or subsegmental) incidentally recognized at a computer tomography obtained for other reasons in a patient without a clinical suspicion of pulmonary embolism, an occurrence whose frequency is rapidly increasing in clinical practice, a final diagnosis cannot be made without further confirmatory testing.