Positron-emission tomography (PET) employing fluorodeoxyglucose (FDG) has proven to be a highly sensitive and specific diagnostic method in the staging and restaging of various neoplasms, including melanoma, complementing morphologic imaging. FDG uptake has been correlated with proliferation rate, and thus, the degree of malignancy of a given tumor (i.e., grading). Consecutively, a relationship of survival prognosis and the extent of tumor burden as well as degree of FDG accumulation--determined by FDG-PET--has been suggested in various tumors. The aim of this study was to assess the potential of fluor-18-FDG-PET in order to evaluate the survival prognosis in melanoma. Patient data (n=95) were retrospectively analyzed, and the results of functional FDG-PET staging was correlated with survival data. Time of staging (diagnosis of primary versus recurrence) had no statistically significant effect on survival prognosis when patients were matched for pertaining node metastasis (NM) stages. Differences in survival were owing to the presence of metastatic disease rather than time of staging. Tumor (T)-stage (T1-T4) alone had no effect on survival prognosis when patients were matched for NM stages. Differences in survival were also due to higher rates of lymph node (LN) and organ metastases in higher T-stages. Detection of LN metastases (N1M0) had a statistically significant and predominant impact on 5-year survival (N0M0 80% versus N1M0 45%; p<0.01). Additional presence of distant metastases in LN-positive patients (N1M1) had only a statistically insignificant further impact on survival (5-year survival in N1M0 45% versus N1M1 29%; p>0.05). Exclusive presence of organ metastases (N0M1) showed a statistically significant drop of survival with a 5-year survival of 61% in N0M1 versus 80% in N0M0, respectively (p<0.03). Further, the combined presence of LN and distant metastases had the worst prognosis (5-year survival in N1M1 29% versus N0M1 61%; p<0.02). Based on a qualitative 4-point scoring system, patients with malignancy-typical FDG uptake showed an overall 5-year survival of 38%, as compared to patients with malignancy-suspicious lesions (71%; p <or= 0.02) or patients without evidence of disease (80%; p<0.001). When PET and computed tomography (CT) findings were compared, survival was best in patients with both studies (CT and PET) being negative (5-year survival, 83%), worst when PET and CT were positive (5-year survival, 61%; p<0.02), and showed an intermediate survival when PET was positive, but with CT still negative (5-year survival, 73%). In patients staged by FDG-PET, significant risk factors--as identified by univariate und multivariate analyses--were (i) malignancy-typical FDG uptake, (ii) detection of LN, and (iii) organ metastases. In conclusion, FDG-PET offers valid prognostic information, demonstrating a good relationship of functional TNM-stage and actual survival prognosis. The results obtained in one PET study seem to be as suitable as the combined results of conventional staging, including clinical and morphologic (e.g., CT) methods and clinical follow-up. FDG-PET was more accurate in tumor detection and seemed to detect tumor spread earlier, as compared to CT.