Semi- automated non- invasive diagnostics method for melanoma differentiation from nevi and pigmented basal cell carcinomas

摘要

The incidence of skin cancer is still increasing mostly in in industrialized countries with light- skinned people. Late tumour detection is the main reason of the high mortality associated with skin cancer. The accessibility of early diagnostics of skin cancer in Latvia is limited by several factors, such as high cost of dermatology services, long queues on state funded oncologist examinations, as well as inaccessibility of oncologists in the countryside regions - this is an actual clinical problem. The new strategies and guidelines for skin cancer early detection and post-surgical follow-up intend to realize the full body examination (FBE) by primary care physicians (general practitioners, interns) in combination with classical dermoscopy. To implement this approach, a semi- automated method was established. Developed software analyses the combination of 3 optical density images at 540 nm, 650 nm, and 950 nm from pigmented skin malformations and classifies them into three groups- nevi, pigmented basal cell carcinoma or melanoma.