Detectability of liver metastases in malignant melanoma:: prospective comparison of magnetic resonance imaging and positron emission tomography

Purpose: We evaluated the diagnostic accuracy of magnetic resonance imaging (MRI) and positron emission tomography (PET) for detection of liver metastases in malignant melanoma. Material and methods: Thirtyfive patients with 39 combined unenhanced MRI and fluorine-18 deoxyglucose (F-18 FDG) PET scans were prospectively studied. In discordant imaging findings final diagnosis was proven by clinical follow-up > 6 months and demonstration of progressive liver metastases by at least one imaging method. Sensitivities and specificities were compared and the influence of lesion size and melanin content on diagnostic accuracy was determined. Results: MRI and PET were concordantly negative for presence and number of liver metastases in 28 patients and positive in four patients. PET and MRI were false positive in one patient each. In one patient MRI showed a single metastases not seen by PET and in one patient MRI demonstrated more metastases at the first examination. In follow-up investigations MRI revealed more metastases than PET in both patients. The sensitivities for lesion detection were 47% (16/34) for PET and 100% for MRI. Lesion detectability by PET was related to lesion size (P < 0.0001) but not to melanin content. Conclusion: MRI is more sensitive in the detection of liver metastases in patients with malignant melanoma. Small lesions are easily missed by PET, while melanin content does not influence detectability by PET. (c) 2004 Elsevier Ireland Ltd. All rights reserved.