Artificial intelligence (AI) and machine learning software are being used by researchers to help detect the earlier stages of skin cancer.
Teams at the University of Waterloo, in Ontario, Canada, and the Sunnybrook Research Institute are using AI to detect melanoma skin cancer through the analysis of thousands of images.
The machine learning software was trained using a large number of skin images and information on the corresponding eumelanin and haemoglobin levels. This means the AI system can now give doctors objective data on telltale signs of skin cancer.
Currently, dermatologists rely on their own, potentially subjective, visual examinations of lesions such as moles, to assess whether patients should have a biopsy to properly diagnose the disease. Skin cancer is fatal if detected too late, but can be treated quite easily if caught at an early stage.
Unnecessary biopsies are a large healthcare cost, as well as being an invasive procedure. Using the AI technology to assess the skin to a higher standard and produce more accurate results helps to counteract these issues. The technology could be made available to doctors as soon as next year.
Using quantitative analysis, the AI looks at several biomarkers in lesions, including changes in the distribution and concentration of eumelanin and haemoglobin. Alexander Wong, professor of systems design engineering at Waterloo, said: ‘This could be a very powerful tool for skin cancer clinical decision support. The more interpretable information there is, the better the decisions are.’
The key is helping doctors to make quicker decisions, says Wong. ‘There can be a huge lag time before doctors even figure out what is going on with the patient. Our goal is to shorten that process.’
This is one of a growing number of examples of AI being used in healthcare. Google-owned DeepMind teamed up with the NHS in 2016 to develop an app called Streams, which helps to detect acute kidney injuries.