What is known?
Melanoma risk-prediction models are statistical models that aim to estimate an individual’s risk of developing melanoma based on the person’s known risk factors, such as presence of moles (or naevi), sunburn history and hair colour. These models can be used, for example, to help clinicians pinpoint people at high risk who may benefit from personalised preventive advice and other measures.
What is new?
Twelve different melanoma risk-prediction models published between 1989 and 2015 were assessed to estimate how accurately they predicted the risk of melanoma in a sample of 1164 people.
All of these people had participated in the population-based Australian Melanoma Family Study and information was available for each person on their melanoma risk factors and whether they had in fact developed melanoma.
Overall, the study found that each of the 12 models performed well in distinguishing between people of the same age with and without melanoma, based on their melanoma risk factors alone.
What does this mean?
The 12 risk prediction models for melanoma were found to perform well enough such that they would allow doctors to correctly stratify their patients according to their future risk of melanoma, providing a useful clinical tool for doctors to target and discuss personalised melanoma prevention strategies with their high-risk patients.
Contact: Joanne Aitken
Reference: Vuong K, Armstrong BK, Espinoza D, Hopper JL, Aitken JF, Giles GG, Schmid H, Mann GJ, Cust AE, McGeechan K. An independent external validation of melanoma risk prediction models using the Australian Melanoma Family Study. Research Letter. British Journal of Dermatology. 2020. doi: 10.1111/bjd.19706.