Abstract

The term 'common mole', often used to describe a subset of benign pigmented skin lesions, is traditionally defined on the basis of morpho-chromatic features. In recent years, certain research groups have developed equipment and methods, such as digital dermoscopy analysis, that enable objective evaluation of pigmented skin lesions. In this study we use a digital dermoscopy analyser trained for the recognition of pigmented skin lesions to compare the subjective definition of 'common' and the mathematical concept of 'close to the mean of measurements'. A subset (100) of digital images of flat pigmented lesions, obtained in daily practice, were classified by trained and non-expert clinicians as common moles (60) or clear-cut melanoma (40), and processed with a DB-Mips analyser. The resulting parameters, validated by a classifier, were used to evaluate Hotelling's T2 multivariate distances from the mean. 'Common' moles could not be clearly defined in terms of closeness to the means of objectively evaluated parameters. Their diagnosis indudes many other evaluations and clusters of variables. The clinical semantics of the term 'common' does not conform to any unambiguous mathematical definition.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call