What cognitive skills do patients need to detect melanoma and how can they be improved?
Background
The incidence of cutaneous malignant melanoma (MM) is increasing with a doubling time of around 10 years, although rises in mortality have been more modest. No surgical or systemic treatment has been shown to improve overall melanoma survival if the cancer has spread beyond the primary site. By contrast it is widely believed that early diagnosis—meaning diagnosis and excision before a primary tumour has metastasised — has accounted for some of the considerable reduction in case-fatality rates seen over the last 50 years. The majority of melanomas are detected by patients themselves, or their relatives or friends.
Patient delay and lack of knowledge in recognising the early features of melanoma remains an impediment to early and hence effective treatment. Early self-detection campaigns have mostly been based on rule-based strategies such as the ABCD(E) criteria. However, there is little experimental work supporting such cognitive strategies, and the idea that one can extend strategies used by experts to novices is questionable. There is also limited evidence to suggest that rule-based approaches such as the ABCD may have undesirable properties and be inferior to approaches that make use of images. Outwith the context of melanoma we have recently shown that 'matching' using a structured image database might be a better strategy, and one that is ultimately scalable using the world wide web.
Support
Cancer Research UK are currently supporting our research into finding an alternative and more effective way for people to identify potentially suspicious lesions on their skin using images rather than a rule-based approach such as ABCD.
Approach
Over the next 2 years we will be carrying out several studies using lesion images from our extensive departmental database. Each study will address different aspects of computer and image-based learning: examples may include investigating whether the ways in which a lesion image is presented onscreen affects the way we would 'feel' about the lesion, or aim to identify which kinds of lesions 'worry' people the most.
Future
We hope that the information gained from these experiments could be used to guide us in designing an interactive, web-based tool to assist the public in the deciding whether or not a skin lesion merits medical intervention.
Volunteering for our research
Our current study is now available online for those wishing to take part in our research. It takes approximately 20 minutes to complete and can be accessed by clicking on the following link:
The study is anonymous, meaning that no personal identifiable information will be asked for or kept by us.
Please ensure you have enough time to complete the whole study before you start as you cannot return to finish an uncompleted version at a later time, and partially completed studies are deleted from our system.
For further information about our research, please contact Karen Robertson at the following address:
email: karen.robertson@ed.ac.uk
telephone: 0131 5362047