E-poster Presentation 2014 World Cancer Congress

Risk prediction models for melanoma: a systematic review. (#558)

Fiona M Walter 1 2 3 , Juliet A Usher-Smith 1 , Angelos Kassianos 1 , Jon Emery 1 2 3
  1. University of Cambridge, University of Cambridge,, Cambridge, UK
  2. University of Melbourne, Melbourne, Australia
  3. University of Western Australia, Perth, WA, Australia

Background:

Melanoma incidence is rising rapidly worldwide among white skinned populations.
Earlier diagnosis is the principal factor that can improve prognosis. Defining high-risk populations using risk prediction models may help targeted screening and early detection approaches.

 Aim:

To identify and compare existing risk prediction models for melanoma.

 Methods:

In this systematic review we searched Medline, EMBASE and the Cochrane Library for primary research studies reporting or validating models to predict risk of developing cutaneous melanoma. To be included, models needed to have been developed using a step-wise method, include a combination of risk factors and be applicable to the general population. Screening of studies for inclusion and data extraction were completed independently by at least two researchers.

 Results:

4141 papers were identified from the literature search and six through citation searching. 25 risk models were included. Between them, the models considered 144 possible risk factors, including 18 measures of number of naevi and 26 of sun/UV exposure. Those most frequently included in final risk models were number of naevi, presence of freckles, history of sunburn, hair colour and skin colour. Despite the different factors included and different cut-offs for sensitivity and specificity, almost all models yielded sensitivities and specificities that fit along a summary ROC with AUROC of 0.755, suggesting most models had similar discrimination. Only 2 models have been validated in separate populations and both also showed good discrimination with AUROC values of 0.79 (0.70-0.86) and 0.70 (0.64-0.77).

 Conclusions:

A large number of risk prediction models for melanoma have been developed. Comparison between them is difficult due to the lack of validation studies and number of different variables considered. Further research should focus on validating existing models rather than developing new ones.