Quantitative measurements of formalin-induced fluorescence for differential diagnostics of melanomas and lesions of human skin.

The usefulness of formaldehyde-induced fluorescence (FIF) for detection of melanoma cells has been suggested by several investigators during the last 40 years. FIF can be easily excited and observed in microscopic sections of formalin-fixed paraffin-embedded skin samples. However, such an approach has never been widely used in melanoma diagnostics for reasons including lack of clear diagnostic criteria, considerable inconsistencies in both the protocols used and qualitatively analysed results reported by different groups. This study aimed at determination of the spectral bands optimum for detecting melanoma cells. The study involved three sets of the excitation and emission bands: gammaex=366 nm, gammaem>425 nm; gammaex=450-480 nm, gammaem>515 nm; gammaex=450-480 nm, gammaem=510-550 nm. Microscopic digital imaging was used to quantitatively determine the fluorescence intensity of 53 primary melanomas and 32 benign lesions. Best classification of melanomas with algorithm based on fluorescence intensity threshold was obtained for gammaex=450-480 nm, gammaem=510-550 nm. Receiver operating characteristics (ROC) analysis of the algorithm yielded area under the curve=0.84 +/- 0.05 for melanocytic cells present in the stratum corneum. Our results clearly indicate that the FIF emitting molecules (most probably 5-S-cysteinyldopa) are present in melanomas at the concentration significantly higher than in benign lesions. In terms of the ROC analysis, the diagnostic performance of the test based on the FIF intensity is as good as of many other commonly used diagnostic tests.

Skin layer-specific Melan-A expression during progression of human cutaneous melanoma: implications for diagnostic applications of the marker.

Melan-A is widely used in the diagnostics of human melanoma. The immunogenicity of this glycoprotein makes it a potential target in immunotherapy and several authors have suggested its potential as a prognostic factor. Up to now there has been no clear direct evidence of changes of Melan-A expression during the progression of melanoma. We have performed objective immunohistochemical assessment of the expression of Melan-A in benign naevi and melanomas at different stages of progression. Our results show a complex pattern of changes in the expression of Melan-A in melanomas depending on the location of melanoma cells within individual skin layers. The expression of the antigen during tumour progression significantly decreases for melanoma cells located in the granular/spinous layer (r=-0.94, P=0.02) and increases for the papillary layer (r=0.99, P=0.002) and reticular layer (r=0.89, P=0.04). It should also be emphasized that from the Clark II level of progression the melanomas can be detected with high sensitivity and specificity using a simple cut-off test based on the determination of Melan-A expression in tumour cells located within the papillary layer.