Image: New AI technology may help detect melanoma skin cancer earlier than current methods and to help reduce the number of unnecessary biopsies (Photo courtesy of Deposit Photos).
Newly developed technology uses artificial intelligence (AI) to help detect melanoma skin cancer earlier than current methods and to help reduce the number of unnecessary biopsies. The AI-based method employs machine-learning software to analyze images of skin lesions and to provide doctors with objective data on telltale biomarkers of melanoma.
"This could be a very powerful tool for skin cancer clinical decision support," said Alexander Wong, professor at University of Waterloo (Waterloo, ON, Canada), "The more interpretable information there is, the better the decisions are." Prof. Wong developed the technology in collaboration with Daniel Cho, former PhD student at Waterloo, David Clausi, professor at Waterloo, and Farzad Khalvati, adjunct professor at Waterloo and scientist at Sunnybrook.
Currently, dermatologists largely rely on subjective visual examinations of skin lesions (e.g. moles) to decide if patients should undergo biopsies to diagnose the disease. The new system deciphers levels of biomarker substances in lesions, adding consistent, quantitative information to assessments currently based on visual appearance alone. In particular, changes in the concentration and distribution of eumelanin (gives color to skin) and hemoglobin are strong indicators of melanoma.
"There can be a huge lag-time before doctors even figure out what is going on with the patient," said Prof. Wong, "Our goal is to shorten that process." The AI system was trained using tens of thousands of skin images and their corresponding eumelanin and hemoglobin levels. It gives doctors objective information on lesion characteristics to help them identify or rule out melanoma before deciding if to take more invasive action. The technology could be available to doctors as early as 2018.
The research was recently presented at the 14th International Conference on Image Analysis and Recognition (ICIAR 2017, July 5-7, 2017, Montreal, Canada).
University of Waterloo