The pretest probability helps determine the best imaging method

Pulmonary embolism remains an elusive diagnosis. Even though nearly all doctors are well aware of the risk factors for venous thromboembolism, pulmonary embolism still causes too many unexpected deaths. The natural history is, however, well known,1 and, once the diagnosis has been considered, appropriate treatment is usually effective and death less likely. Anticoagulants, however, are not without risk and accurate diagnosis remains essential. Pulmonary angiography is the gold standard for diagnosis, but it is invasive, perceived as dangerous, and not widely available. The burden of diagnosis therefore rests on non-invasive imaging.

Over the past two decades the next best investigation has been ventilation-perfusion (V/Q) scintigraphy, loosely known as the “lung scan.” The patient inhales a technetium-99m labelled aerosol or radioactive gas (such as krypton-81m), which demonstrates the distribution of pulmonary ventilation. This is performed in association with an intravenous injection of technetium-99m labelled microparticles of macroaggregated human serum albumin, which shows the distribution of pulmonary blood flow. When ventilation is compared with perfusion areas showing reduced perfusion not matched on …