Elias Fizesan (17) presents his device for analysing coughing patterns to diagnose and monitor asthmatic patients FJSL

Elias Fizesan (17) presents his device for analysing coughing patterns to diagnose and monitor asthmatic patients FJSL

Fizesan is among the laureates representing Luxembourg at several international competitions, most recently presenting his project at the world’s biggest youth science competition, Regeneron International Science and Engineering Fair (ISEF), where he won the third prize in the biomedical engineering category. 

Can you tell us about the device you created, and why this project matters to you? 

Elias Fizesan: Basically, I created a device that can be attached to the chest of a patient that a doctor suspects might have asthma. This device monitors the patient’s breathing vibrations and can diagnose the patient by analysing these vibrations through some algorithms. Aside [from] all the worrying respiratory ailments associated with the covid-19 pandemic, I got interested in the subject after reading a research paper citing one in three asthmatic patients as misdiagnosed. This represents around 100m (33%) misdiagnosed individuals out of an estimated 300m asthmatic patients who believe they have asthma and may be consuming medication for a non-existent condition.

What did the data collection process entail?

I couldn’t find any data online, so I was left with two options: to quit or start my own data collection process. So I created a website called recordacough.com with an eight-second timer to collect three different audio cough samples that is then sent to a database used to train ML [machine-learning] algorithms, which often require tons of data to improve accuracy. My data collection process was similar to that of researchers at the EPFL, who used a website to retrieve data on covid patients with the aim to improve diagnosis. I believed this could also be used for asthmatic patients, and it was a safe way to collect data in a pandemic. Audio cough samples are still collected via the website to boost the machine-learning algorithms with over 3,000 audio samples already submitted, including multiple entries by the same participant, but this only increases the accuracy of the algorithms since each cough is unique. Most of the participants were contacted through word of mouth.

From winning a silver medal for your AI greenhouse project to a health-related project: why did you make this shift, and did you have to learn new skills to do so?

The great thing about the AI greenhouse is that it involved computer science, in which I already have some background, with over seven years in programming and five years working on ML, so it was quite easy for me to transition. Although I had to learn more about biology, respiratory conditions, the human body, asthma's effect on the lungs... but computer science helps you learn about different fields. I was able to learn, explore and implement many algorithms during this project. It took me two weeks to build the website.

Elias Fizesan

Elias Fizesan (17) FJSL laureate and I-FEST2 silver medal recipient (middle). Photo: courtesy FJSL 

 

You presented your project at the Regeneron ISEF. How did your interview go, and how would you want the world to react to your findings?

I made four presentations to four judges. The first was a general presentation while the second focused more on technical aspects such as my formular for the codes and improving the algorithms. The last two sessions with biomedical engineers as judges was the most interesting as we discussed the application of the device for child diagnosis using specific algorithms for children considering my data was trained using participants over 16 years old and children cannot really express their symptoms to doctors. An ideal reaction would be the use of this device by doctors alongside additional tests to increase diagnosis accuracy. Some of the routine tests doctors perform include collating family history and although this alone is not the most accurate, it does contribute to some accuracy. With a 90% diagnosis accuracy, combining the device with other tests will increase accuracy to near perfect diagnosis.

What’s next? And do you have any last word for other young aspiring scientists?

I plan to apply to university in the UK or US in October, probably in the field of computer science. Also I would like to keep exploring new areas that involve computer science and biology or a completely new subject. For aspiring scientists, I would say, start as soon as possible and designate time for research. There’s so much information on the internet for free! Start with the most difficult task of the day and build a routine. I dedicated 10 to 15 minutes to research each day, and I discovered quite a lot in that time. Keep taking small steps until you find a field you’re really passionate about, and passion will drive you.

Read more about the FJSL 2021 laureates here.