On clinical criteria alone, the diagnosis of early Parkinson's disease can be difficult and, by definition, the prospective recognition of preclinical Parkinson's disease is impossible. Positron emission tomography (PET) using [18F]dopa as tracer has been proposed as a means of identifying patients with preclinical disease. The number of subjects detected to date has been few; most have been identified by serendipity or during the course of family studies. This review examines the significance of a single abnormal scan in an apparently healthy subject in terms of the relationship between normal and abnormal values and the time course of the disease.