In association with
Omrop Fryslân
NOS News•
Two 17-year-old secondary school students in Leeuwarden have developed a program that can help detect lung cancer. The lung department of the Medical Center Leeuwarden (MCL) is very enthusiastic about the project of the teenagers, even though it is not new what they have come up with.
Ted de Reus and Yahya Oufkir (both 17 years old) researched for their profile paper at the gymnasium how artificial intelligence can help to recognize cancer cells.
The secondary school students wanted to do “something social” for their profile paper, they say Omrop Fryslân. Ted de Reus: “We have had several family members who have died of lung cancer. We noticed that it often ends fatally and that society could use help to reduce the number of deaths.”
Artificial intelligence
Through Ted’s older brother they came into contact with artificial intelligence (AI), or artificial intelligence. They decided to put 25,000 microscopic images of pieces of lung tissue into a computer program. “The program started training for that,” says Yahya. “The system can now distinguish healthy cells from lung cancer cells based on that image.”
“It’s very clever of two 17-year-old boys,” says Professor of Computational Pathology Jeroen van der Laak. He is involved in the application of artificial intelligence in pathology on a daily basis. “However, the technique used by the students is not unique. At Radboudumc alone, we have a group of 35 people who do nothing else.”
The boys received help from the lung department of the hospital in Leeuwarden. “In a very short time they have made something that I can use to speed up the process immediately in my practice. That is almost never possible with such a study,” says pulmonologist Wouter Van Geffen.
10 for profile work piece
The doctor would like the students’ program to be further developed. “The next step may be that we ask permission from the ethics committee if we can also put test material from MCL patients in here. Then we will see if it works just as well as on the material that the boys have processed.”
Van der Laak cannot imagine that the boys’ program will soon be used in the MCL. “Of course they only examined one dataset. European legislation would really not allow their program to be used clinically in the near future.”
Useful in practice or not, the students received a 10 for their profile paper. “I would have been very pleased if they had developed some kind of Mickey Mouse program that would recognize four pixels. But this exceeds all expectations,” says their teacher Erik Wessels.