Machine-Learning Classification of Port Wine Stain With Quantitative Features of Optical Coherence Tomography Image

Abstract

Port wine stain (PWS) is the benign congenital capillary malformation of skin, occurring in 0.3% to 0.5% of the population. In this paper, we build two automated support vector machine (SVM) based classifiers by extracting quantitative features from normal and PWS tissue images recorded by optical coherence tomography (OCT). We use both full feature set and simplified feature set for training. Accuracy of 92.7%, sensitivity of 92.3% and specificity of 93.8% were obtained for classifier with full feature set. Accuracy of 92.7%, sensitivity of 94.9% and specificity of 87.5% were obtained for classifier with simplified feature set. Our results suggest that extracting quantitative features from optical coherence tomographic images could be a potentially powerful method for accurately and automatically identifying PWS margins during laser therapy.