RATIONALE AND OBJECTIVES: The aim of the study was to evaluate the use of high resolution magnetic resonance imaging (MRI) for differentiation of melanin-containing skin tumors. METHODS: Twenty-seven melanocytic nevi and 18 malignant melanomas were examined by high resolution MRI. Signal intensities and signal-to-noise (SNR) and contrast-to-noise ratios (CNR) of the tumors were determined in enhanced (T1, T2, water-suppression, and fat-suppression sequences) and contrast-enhanced images (T1 and fat-suppression sequences). The differences were tested for significance by a Wilcoxon test. RESULTS: Malignant melanomas differed from melanocytic nevi in that they displayed a higher SNR in T2-weight and unenhanced and contrast-enhanced fat-suppression scan. Malignant melanomas exhibited a higher CNR than did benign lesions in unenhanced and contrast-enhanced fat-suppression sequences with dermis as reference tissue (P < 0.05) and in enhanced fat-suppression sequences with subcutis as reference tissue (P < 0.05). CONCLUSIONS: The usefulness of SNR and CNR analysis on MRI for the differentiation of malignant skin tumors from benign skin tumors of the melanin-containing system is limited. Clinical and histologic examinations are, further, the important step in evaluation of melanin-containing skin tumors.
RATIONALE AND OBJECTIVES: The aim of the study was to evaluate the use of high resolution magnetic resonance imaging (MRI) for differentiation of melanin-containing skin tumors. METHODS: Twenty-seven melanocytic nevi and 18 malignant melanomas were examined by high resolution MRI. Signal intensities and signal-to-noise (SNR) and contrast-to-noise ratios (CNR) of the tumors were determined in enhanced (T1, T2, water-suppression, and fat-suppression sequences) and contrast-enhanced images (T1 and fat-suppression sequences). The differences were tested for significance by a Wilcoxon test. RESULTS:Malignant melanomas differed from melanocytic nevi in that they displayed a higher SNR in T2-weight and unenhanced and contrast-enhanced fat-suppression scan. Malignant melanomas exhibited a higher CNR than did benign lesions in unenhanced and contrast-enhanced fat-suppression sequences with dermis as reference tissue (P < 0.05) and in enhanced fat-suppression sequences with subcutis as reference tissue (P < 0.05). CONCLUSIONS: The usefulness of SNR and CNR analysis on MRI for the differentiation of malignant skin tumors from benign skin tumors of the melanin-containing system is limited. Clinical and histologic examinations are, further, the important step in evaluation of melanin-containing skin tumors.