Down to Earth: Spotting Skin Cancer

24 August 2017

Interview with

Stuart Higgins, UCL

Down To Earth takes a look at tech intended for space which has since found a new home down here on Earth... This week - how searching for stars could help detect a deadly illness…

Stuart - What happens when the science and technology of space comes Down to Earth.

This is Down to Earth from the Naked Scientists where we look at how the technology and science developed for space is used back on planet Earth.

I’m Dr Stuart Higgins, and today we’re talking about how the mathematics used to spot X rays coming from exploding stars can also be used to catch the early signs of skin cancer.

On June the 1st 1990, the German Aerospace Centre, with the UK and US, launched ROSAT the ROentgen satellite. It was designed to allow researchers to search for X rays coming from across the universe. However, scientists needed a way to identify the faint traces of X rays from the data the satellite was producing, a bit like trying to hear a voice under a background of static. They needed to separate the signals from the noise.

This led researchers at the Max Planck Institute for Extraterrestrial Physics in Germany to develop the scaling index method. It’s an analysis technique which identifies structures from within a dataset.

Imagine a digital photograph of the seaside; a beautiful golden beach, the sea lapping at the shore. The scaling index method would take each pixel in that image, draw a circle around it, and check to see whether it’s neighbouring pixels look similar. The algorithm then uses that information to work out whether the pixel belongs to - say - a straight line like the horizon, a flat plane like the sky, or a point object like a seagull up above. Now imagine the same scene on a greying cloudy day; it’s tricky to see that seagull. The algorithms power is that it can still pick out the seagull from the noise of the grey clouds.

But this algorithm could be applied to more than just astrophysics or, indeed, seagull spotting. The same researchers in Germany teamed up with medical doctors to invent a system for spotting the early signs of skin cancer. Skin cancer can look like a darkening in the pigmentation of the skin and can be challenging to properly identify. In their system, doctors can take magnified digital images of a patient’s skin and apply the scaling index method to identify subtle changes in the colour and variation. The algorithm is able to score how likely the variation is due to melanoma, a type of skin cancer. The big advantage is that the doctor doesn’t need to be an expert in melanoma in order to use the system. The algorithm does the work helping the doctor make an accurate diagnosis.

So that’s how the mathematics of spotting X rays from exploding stars is being used to help identify skin cancers back on Earth.

 

Add a comment