Early diagnosis of skin cancer (melanoma) is essential to improve survival. Melanoma frequently develops from existing moles on the skin. Current diagnostic practice relies on expert clinicians (dermatologists) being able to successfully identify any new or changing moles in individuals who may have many (> 60) irregularly shaped moles. This ‘sequential monitoring’ of skin moles constitutes the gold standard method for early melanoma detection. Astrophysicists face a similar challenge when they map the night sky to detect new events, such as exploding stars. Using computer algorithms, based on two or more sequential images, astrophysicists can detect new events, identify their changes, and accurately predict how they will appear subsequently.
MoleGazer is a collaboration between astrophysicists and dermatologists to apply these astrophysical analysis techniques of sequential images of the night sky, to the automated detection, characterisation, and monitoring of skin lesions. This ERC PoC grant had the aim of constructing a new and unique database of ‘total body photography’ (TBP – high-quality images of the entire skin). The PoC project enabled the TBP of the same patients attending skin cancer screening clinics at multiple time points. These patients included both 50 high-risk patients attending screening clinics for sequential three-monthly TBP imaging and clinical assessment, and several hundred patients in clinics who have undergone TBP as standard care.
The PoC grant used these TBP images to develop an algorithm that accurately registers the images and identifies the same moles in sequential imaging. These moles are then presented for characterisation by both astronomical source detection algorithms and by expert dermatologists. These data characterise the properties and evolutionary path of naevi to form a training sample of moles that will next be used to inform the adaption of existing astronomical algorithms. These algorithms will be used in the development of the final MoleGazer algorithm, designed to ‘map moles’ over a patient’s lifetime to detect changes, with the eventual aim of using this technique to detect melanoma as early as possible.
