8566 Background: The incidence of malignant melanoma has risen dramatically over the past decades. Prognosis for patients with metastatic melanoma is poor and treatment options are limited. Complete surgical resection is frequently front line therapy for patients with resectable metastases. However this is only an option if relapse is diagnosed early. PET imaging using 2-flourine-18, 2-fluro-2-deoxy-D-glucose (FDG) has been used successfully for staging of metastatic melanoma. However it's usefulness in surveillance for patients at high-risk of relapse (resected stage III and IV disease) has not been well studied. We conducted a retrospective review of our institutional experience using PET scanning as part of regular follow up in patients with resected stage III or IV melanoma for the purpose of detection of early recurrence. Methods: Retrospective review of the medical records of 667 consecutive patients seen at Mayo Clinic from 12/13/1999–1/1/2007, with the diagnosis of malignant melanoma for whom PET imaging was performed or reviewed at our institution. Eligibility criteria included: resected stage III or IV melanoma with at least 2 PET scans (<1 year apart) as part of regular clinical follow-up (surveillance). Frequency of PET scanning was variable and at physician discretion. Results: Of the 667 patients, 110 patients were deemed eligible for review. There were 98 recurrences among 65 patients. Of these 37 recurrences (38%) were found by PET scanning alone (asymptomatic); 61 recurrences (62%) were found by other methods (physical exam, CT, MRI, etc). Of the 37 recurrences found by routine PET scanning 17 (46%) were completely resected; of the other 61 recurrences found by other methods 33 (54%) were completely resected. Conclusions: PET scanning as regular surveillance for patients with high risk melanoma, may be a useful tool in early identification of asymptomatic melanoma relapse amenable to complete surgical resection (additional 1/3 of cases). Considering the potential clinical benefit of completely resected melanoma relapse, and the retrospective nature of this review, these data are hypothesis-generating and warrant further investigation. No significant financial relationships to disclose.