We reviewed our experience with positron imaging tomography (PET) scanning from 1996-1998 in 45 patients (51 scans) with high risk (n = 15) or recurrent melanoma (n = 30). Positive or negative PET findings were considered correct if confirmed or refuted by other radiologic modalities or biopsy within 6 weeks, or clinical follow up within 6 months. PET scans were usually obtained to make a decision regarding surgical management (29/51 scans, 57%) or to confirm that a patient was disease free (12/51 scans, 23%). PET results facilitated clinical decision making in 15/45 (33%) patients, often due to the detection of distant disease making a surgical procedure inappropriate (5/15, 33%). There were 3 false positive (6%) and 7 false negative (14%) PET scans. False negative scans occurred in patients with small volume disease of the lung (4 patients) or CNS (2 patients, 3 scans). Three patients had false positive scans. These were secondary to benign hepatic tumors and sarcoidosis (one patient each). One patient with stage IV melanoma had PET findings not confirmed by CT scanning or clinical follow-up. PET scanning was clinically useful in a third of patients in our initial experience, and false positives were uncommon. Small volume disease in the lung or CNS may be missed by PET, and additional imaging modalities should be obtained when patient management would be impacted.