The evaluation of myocardial glucose utilisation with fluorine-18 fluorodeoxyglucose (FDG) and positron emission tomography is currently considered the most reliable tool for the identification of myocardial viability. However, the investigations using FDG imaging to predict improvement in left ventricular (LV) function after revascularisation have reported wide ranges for sensitivity (71%-100%) and, in particular, for specificity (33%-91%). The variable results may be related to differences in study populations but also to differences in the imaging protocols employed. Detailed analysis of the published studies has revealed differences in study populations, patient selection criteria, the methods for assessing changes in LV function post revascularisation and the timing of these assessments. Even more importantly, protocols have varied substantially with regard to imaging equipment, perfusion tracers, metabolic conditions, data analysis and interpretation of results. In addition, evaluation of patients with insulin resistance appears to represent a specific challenge. This review examines the different study protocols and methodologies used for myocardial FDG imaging in order to draw conclusions concerning optimal imaging protocols. It appears that the optimisation and standardisation of study protocols and analysis of FDG images for the assessment of myocardial viability are critical. In addition, multi-centre trials seem warranted on prediction of long-term function, congestive heart failure symptoms, survival and quality of life.