Abstract

PURPOSE: This study aims to compare the use of 18F-FDG-PET/CT, CT, brain MRI, and tumormarker S-100B in chemotherapy response assessment of stage IV melanoma patients. METHODS: In 25 patients with stage IV melanoma, FDG-PET/CT and S-100B after 2-3 months (three cycles) of chemotherapy was compared with baseline PET/CT and baseline S-100B. Retrospectively, the response was correlated with the outcome. In patients with clinical suspicion for brain metastases, MRI or CCT was performed. RESULTS: There was agreement between FDG-PET/CT and CT regarding response to chemotherapy in all patients. There was a clear trend to a longer OS of PET/CT responders (n = 10) compared with PET/CT non-responders (n = 15; p = 0.072) with remarkably better 1-year OS of 80% compared to 40% (p = 0.048). There was a significant longer PFS of PET/CT responders compared with PET/CT non-responders (p = 0.002). S-100B was normal at baseline in eight of 22 patients where it was available. Chemotherapy response assessment with S-100B failed to show correlation with OS or PFS. Eleven patients developed brain metastases during treatment, first detected by PET/CT in two and by MRI or CCT in nine of 11 patients. Appearance of brain metastases was associated with a poor survival. CONCLUSIONS: 18F-FDG-PET/CT and CT alone are equally suitable for chemotherapy response assessment in melanoma patients and clearly superior to S-100B. PET/CT responders have better early survival, but this is shortlived due to late therapy failure-often with brain recurrence. Additional brain MRI for therapy response assessment in such high-risk patients is mandatory to detect brain metastases missed by PET/CT.