APPLICATION OF AN ARTIFICIAL NEURAL-NETWORK IN EPILUMINESCENCE MICROSCOPY PATTERN-ANALYSIS OF PIGMENTED SKIN-LESIONS - A PILOT-STUDY

In vivo epiluminescence microscopy (ELM) is a non-invasive technique w hich improves the clinical diagnosis of naevi and malignant melanoma b y providing diagnostic criteria that cannot be appreciated by the nake d eye. The present study investigated whether ELM criteria pattern ana lysis can be employed in an objective, observer-trained, computer-aide d diagnostic system, and whether artificial neural networks (ANN) can be applied to the diagnosis of pigmented skin lesions (PSL). The ELM c riteria patterns of 200 PSL oil immersion images (60 common naevi, 60 dysplastic naevi, and 80 malignant melanomas) were analysed using a st andardized questionnaire. One hundred randomly assigned PSL were used as a training set for an ANN, the remaining 100 PSL serving as the tes t set. The ANN was trained by backward propagation according to the hi stological diagnosis, and its performance was compared with that of hu man investigators. Out of the test set the human investigators correct ly diagnosed 88% of PSL and the ANN 86%. In a dichotomized model compa ring common, compound, and dysplastic naevi vs. malignant melanoma, i. e. benign vs. malignant PSL, the sensitivity and specificity of human diagnosis was 95 and 90%, respectively, whereas the sensitivity and sp ecificity of the ANN diagnosis was 95 and 88%. Our data indicate that artificial neural networks can be trained to diagnose PSL at a human e xpert level, based on patterns provided by ELM criteria. We suggest th at this technique offers a new approach to the diagnosis of PSL.