Astro Technology Applied to Dermatology
Oct 05, 2011
The same techniques astronomers use to parse starlight could be used to assess deadly melanomas, according to astronomy professor Andrew Sheinis. Based on a hyperspectral imager used to get information about the size and composition of distant celestial objects, Sheinis has compressed this technology into a camera-sized device that can be integrated into a microscope to get a three-dimensional snapshot of potentially cancerous moles. He was interviewed by Science & Vie magazine this summer.
Moles Like the Stars
By Andrew Sheinis
Science & Vie, August 2011
SV: How did you come across this idea?
AS:It’s been more than 10 years. I was at the dermatologist, whom I visit every 6 months. Because I spent a lot of time surfing under the California sun as a child, my skin is covered in moles and I have a major risk of developing skin cancer. I was struck by how the dermatologist took few photos. His diagnostic relied on a method named ABCDE (A for asymmetry, B for irregular borders, C for color, D for diameter, and E for evolution): the more asymmetric the mole, with irregular borders, a color that changes, a big diameter, the more it will evolve and risk hosting cancerous cells. Therefore, one must remove them. But the decision depends enormously on the experience and talent of the physician. For me, this subjective exam resembled astronomy observations from hundreds of years ago. At the time, I was working on developing a ‘hyperspectral’ imager for the enormous Keck telescope in Hawaii, and I wondered to myself whether we could adapt this technology to the medical follow-up of moles.
SV: What did this consist of exactly?
AS:Hyperspectral imaging allows us to record simultaneously a luminous signal in several hundred wavelengths. And we can do this for each pixel of an image. At the beginning, this technology was developed for use by the Department of Defense for satellite observations. In astronomy, we use it to classify the size and composition of stars. It allows us to register very subtle changes in size, shape and color.
SV: What was needed to adapt this to dermatology?
AS:First, we had to miniaturize it so that we could build a portable imager, a few centimeters on each side. By the end, it was coupled with an objective lens, which made it similar to a camera, linked to a computer. This would register the data that they could then compare with those gathered in the previous test, and signal to the doctor the high-risk moles.
SV: What is the happening with the project at the moment?
AS:When the idea came to me, I hadn’t found anyone to finance the project. But 18 months ago, the University of Wisconsin gave me the opportunity to work on it. I contacted doctors and biologists. With an industrial partner, we developed the machine. At the moment, we’re comparing the hyperspectral results with standard visual results on approximately 50 patients. We’re testing the feasibility before embarking on clinical trials.