Purpose: The objective of this study was to evaluate the value of (18)F-FDG PET/CT and S-100B tumour marker for the detection of liver metastases from uveal melanoma in comparison to liver metastases from cutaneous melanoma.
Methods: A retrospective evaluation was conducted of 27 liver metastases in 13 patients with uveal melanoma (UM) (mean age: 56.8, range: 30-77) and 43 liver metastases in 14 patients (mean age: 57.9, range: 40-82) with cutaneous melanoma (CM) regarding size and FDG uptake by measuring the maximum standardized uptake value (SUV(max)). S-100B serum tumour markers were available in 20 patients. Cytology, histology, additional morphological imaging and follow-up served as reference standard. In nine patients liver metastases were further evaluated histologically regarding GLUT-1 and S-100 receptor expression and regarding epithelial or spindle cell growth pattern.
Results: Of 27 liver metastases in 6 of 13 patients (46%) with UM, 16 (59%) were FDG negative, whereas all liver metastases from CM were positive. Liver metastases from UM showed significantly (p < 0.001) lower SUV(max) (mean: 3.5, range: 1.5-13.4) compared with liver metastases from CM (mean: 6.6, range: 2.3-15.3). In four of six (66.7%) patients with UM and liver metastases S-100B was normal and in two (33.3%) increased. All PET-negative liver metastases were detectable by morphological imaging (CT or MRI). S-100B was abnormal in 13 of 14 patients with liver metastases from CM. S-100B values were significantly higher (p = 0.007) in the CM patient group (mean S-100B: 10.9 microg/l, range: 0.1-115 microg/l) compared with the UM patients (mean: 0.2 microg/l, range: 0.0-0.5 microg/l). Histological work-up of the liver metastases showed no obvious difference in GLUT-1 or S-100 expression between UM and CM liver metastases. The minority (36%) of patients with UM had extrahepatic metastases and the majority (86%) of patients with CM had extrahepatic metastases, respectively. There was a close to significant trend to better survival of UM patients compared with CM patients (p = 0.06).
Conclusion: FDG PET/CT and serum S-100B are not sensitive enough for the detection of liver metastases from UM, whereas liver metastases from cutaneous melanoma are reliably FDG positive and lead regularly to increased S-100B tumour markers. The reason for the lower FDG uptake in UM liver metastases remains unclear. We recommend to perform combined contrast-enhanced PET/CT in order to detect FDG-negative liver metastases from UM.