Positron emission tomography (PET) utilizing fluorine-18 fluorodeoxyglucose (FDG) has been used in the evaluation of non-small cell lung cancer (NSCLC). Recently its use in the staging of small cell lung cancer (SCLC) has been reported. However, the prognostic value of FDG-PET imaging in SCLC has not been studied. We performed a retrospective analysis to assess this, with the following hypotheses: (1) PET-positive patients would have a less favorable prognosis than PET-negative patients and (2) a high standardized uptake value (SUV) would be associated with a poor prognosis. Retrospective review of a mixed population of treated and untreated patients imaged between 1995 and 2000 was performed. Results of 62 scans in 46 patients were analyzed. There were 8 untreated and 38 treated patients. Findings were correlated with pathology, computed tomography/magnetic resonance imaging and clinical data. The sensitivity of PET scanning was 100% with pathological correlation. The prognostic value of a positive PET study was determined. Overall survival in PET-positive cases was significantly worse than that in PET-negative cases ( P=0.0108). Correlation of SUV(max) with survival showed a significant negative correlation ( P=0.0021). In the eight untreated patients, scans were strongly positive and in all cases the scan results concurred with the final clinical stage assigned on the basis of conventional methods. We conclude that FDG-PET imaging provides prognostic information in treated patients. A positive study and a high SUV(max) are significantly associated with poor survival. Additionally, FDG-PET may be helpful in staging and follow-up.