Using AI to Spot Early Signs of Oral Cancer
Oral cancer is one of the most common cancers in the world with approximately 300,000 new cases diagnosed each year. As with many cancers, early detection is vital which is where artificial intelligence may be able to help.
Researchers from the University of Kingston, London have just secured funding to explore how AI could be trained to detect the early signs of oral cancer using a mobile phone app.
Oral cancer typically appears as a growth or sore in the mouth that does not go away and includes cancers of the tongue, lips, cheeks, the hard and soft palate and the floor of the mouth. Symptoms include unexplained bleeding, persistent sores and the development of red and white patches in the mouth.
Early Warning Signs
Professors Sarah Barman and Paolo Remagnino from Kingston University will work alongside a team from the University of Malaya to train an artificial intelligence system to spot early warning signs using images taken of the inside of the mouth by mobile phones.
“We’re basically training a system to detect abnormalities in the mouth that could be the early indications of oral cancer,” said Professor Barman. “Our challenge is to develop deep learning models that demonstrate high accuracy and prediction of disease.”
Cancer Research Malaysia has already developed a phone app called MeMoSA (Mobile Mouth Screening Anywhere) that can capture images of the oral cavity that can be interpreted remotely. However, it relies on the availability of medical specialists to review and interpret the images. The hope is that the new research project will develop an AI system to speed up the process.
Training Algorithms
Algorithms will be trained on thousands of images of normal mouths as well as those with abnormalities.
The research team will also use a type of artificial intelligence called a generative adversarial network, or GAN. This creates realistic images based on what it has learned.
If successful, the AI-powered mobile phone app will not only save lives but also time and money, freeing up resources within public healthcare systems.
Add your Comment
[LOGIN FIRST] if you're already a member.fields are required.