Of all the tracer techniques currently available for the detection of myocardial viability, it is the classic pattern of fluorine-18 deoxyglucose-perfusion mismatch that is clearest from the conceptual point of view and consistently gives good predictive values. Measurements of absolute rates of glucose uptake depend on the much criticized lumped constant, never validated for myocardial ischaemia, and may provide little additional information or may even be confusing because of the bi-directional changes in glucose uptake in response to increasing ischaemia. Labelled nitroimidazole compounds are currently of interest because they are "ischaemia-avid" and because they can be imaged by a gamma camera. Nevertheless, much more work is required to show whether retention of nitroimidazole in ischaemic tissue may reflect viability.