The use of computerised tomography in the diagnosis of pulmonary embolism has been the subject of clinical research while, at the same time, technical progress has provided the current multidetector-row spiral equipment. Computerised tomography has been assessed both with respect to reference strategies as well as in extensive pragmatic trials. The preliminary evaluation of the clinical probability and the assay of d-dimers has progressively become imperative. The value of the venous doppler ultrasound of the legs, in particular in the elderly, is limited by the variable accessibility according to the centre. In rare cases, uncertainty persists, for example with a good quality negative multidetector-row spiral computerised tomography associated with a high clinical probability, leaving room for complementary explorations. The confrontation between clinicians and radiologists is then all the more pertinent.