Abstract

PURPOSE: To evaluate the usefulness of PET/CT in melanoma patients with an elevated serum S-100B tumour marker level.

METHODS: Out of 165 consecutive high-risk melanoma patients referred for PET/CT imaging, 47 had elevated (>0.2 microg/l) S-100B serum levels and a contemporaneous 18F-FDG PET/CT scan. PET/CT scans were evaluated for the presence of metastases. To produce a composite reference standard, we used cytological, histological, MRI and PET/CT follow-up findings as well as clinical and S-100B follow-up.

RESULTS: Among the 47 patients with increased S-100B levels, PET/CT correctly identified metastases in 38 (30 distant metastases and eight lymph node metastases). In one patient with cervical lymph node metastases, PET/CT was negative. Eight patients had no metastases and PET/CT correctly excluded metastases in all of them. Overall sensitivity for metastases was 97% (38/39), specificity 100% (8/8) and accuracy 98% (46/47). S-100B was significantly higher in patients with distant metastases (mean 1.93 microg/l, range 0.3-14.3 microg/l) than in patients with lymph node metastases (mean 0.49 microg/l, range 0.3-1.6 microg/l, p=0.003) or patients without metastases (mean 0.625 microg/l, range 0.3-2.6 microg/l, p=0.007). However, 6 of 14 patients with a tumour marker level of 0.3 microg/l had no metastases.

CONCLUSION: In melanoma patients with elevated S-100B tumour marker levels, FDG-PET/CT accurately identifies lymph node or distant metastases and reliably excludes metastases. Because of the significant number of false positive S-100B tumour marker determinations (17%), we recommend repetition of tumour marker measurements if elevated S-100B levels occur before extensive imaging is used.