Description

High density of pigmented skin lesions (i.e. common skin moles) is the strongest known risk factor for cutaneous malignant melanoma. Although the majority of children are born with none, pigmented skin lesions begin to form and develop in childhood, and about 50% of malignant melanomas originate from pre-existing pigmented lesions. Many epidemiologic studies of melanoma, therefore, have been focusing on pigmented lesion development, and pigmented lesion enumeration is an integral part of these studies. In order to standardize the enumeration, an interdisciplinary team of skin cancer investigators developed an enumeration protocol which was published by the International Agency for Research on Cancer (IARC) in 1990. The protocol also specifies that pigmented lesions should be documented, by photographing the back of the participant as the subject stands with arms by his or her sides.

The enumeration protocol reduces variation in lesion counts to an acceptable level when a study involves a small group of examiners and a relatively short time period. However, the protocol may have problem in controlling counting variation if a study involves a large number of participants, a large group of examiners, and a long time period. In addition, manual counting is time consuming and expensive. It is desirable to automate the enumeration by analysing photographs taken according to the IARC protocol using a computer program. In this project, we will develop such a computer program to enumerate pigmented skin lesions on the back of a patient.

Principal Investigator

• Tim Lee

Research Team

• Stella Atkins
• David McLean
• Richard Gallagher