Novel biomarkers distinguished melanoma from dysplastic nevi
Four novel molecular biomarkers could aid in clinically diagnosing melanoma from dysplastic nevi, according to study results.
Using biopsies obtained from a 1990 to 1998 archives, researchers studied immunohistochemical expressions of 12 molecular biomarkers (pAkt, Bim, BRG1, BRMS1, CTHRC1, Cul1, ING4, MCL1, NQO1, SKP2, SNF5 and SOX4) in 122 melanomas (28 on the head and neck, 94 from sun-protected sites) and 33 dysplastic nevi on tissue microarrays. Receiver operating characteristic (ROC) curve and artificial nural network (ANN) analysis were used for determining the accuracy of single marker and optimal combinations. Markers simultaneously optimizing melanoma accuracy were examined by classification and regression tree.
Univariate and multiple logistic regression analysis showed that Bim, BRG1, Cul1 and ING4 had statistically significant expression differences in comparing melanoma with dysplastic nevi (P<.01). In combination, the four markers attained 84.3% area under the ROC curve, 94.3% sensitivity and 81.8% specificity for discriminating melanoma from dysplastic nevi. In a separate output unit used to confirm 4-marker differentiation between dysplastic nevi and melanoma, ANN correctly classified 82.3% of training and 81.2% of testing. ING4, Cul1 and BRG1 had the most important classification parameters with a cross-validation error of 0.03 when ranking top-performing biomarkers.
“Specific markers for discrimination of melanoma from dysplastic nevi were scarce,” the researchers said. “The misdiagnosis of melanoma is the second most common reason for cancer malpractice claims in the United States. Four markers described here could be used to assist the histological diagnosis of melanoma, thereby providing important information to clinical pathologists.”