Automated evaluation system of dermoscopic images of longitudinal melanonychia: proposition of a discrimination index for detecting early nail apparatus melanoma.

The prognosis of nail apparatus melanoma is generally poor because of difficulty in early stage diagnosis. Most nail apparatus melanomas occur as longitudinal melanonychia, and criteria and algorithms for dermoscopy diagnosis of longitudinal melanonychia have only recently been proposed. However, as with any clinical diagnosis, the diagnosis based on dermoscopy is to some extent subjective. Our goal is to develop an automated dermoscopic screening system for longitudinal melanonychia and to propose a novel objective and quantitative index for discriminating early nail apparatus melanoma from benign longitudinal melanonychia including melanocytic nevus. We propose an automatically calculated index representing degrees of color variegation in dermoscopic images of longitudinal melanonychia. Dermoscopy images of six cases of early stage nail apparatus melanoma and 25 cases of benign longitudinal melanonychia were analyzed with our screening system and a threshold of melanoma discrimination index was determined. This single melanoma discrimination index diagnosed early nail apparatus melanoma with 100% sensitivity and 92% specificity. The automatically calculated index proposed in the present study is valuable for managing longitudinal melanonychia. The results suggest that the degree of color variegation is essentially different between early nail apparatus melanoma and benign longitudinal melanonychia including melanocytic nevus of the nail apparatus.