Meet the HSC student who built an AI model to detect breast cancer

2 months ago 6

The eureka moment arrived just before midnight. Michaela Loukas, 17, wrapped in pyjamas and in the thick of the HSC, was tinkering at her science extension project when the results came in.

She had built an AI model to detect cancerous breast tissue. There was a twinge of excitement as she read the results, but she tamped it down and ran the test again. The same result blinked back at her.

Michaela Loukas from Marist Catholic College, Penshurst, took home the title of young scientist of the year for her project identifying malignant tissue with 98 per cent accuracy.

Her model had identified malignant tissue with 98 per cent accuracy.

“It just kept repeating the result … I just couldn’t believe it. I was like, hold up – did I just pick up something pretty significant?” Michaela said.

HSC science extension teacher Rachel Thompson described Michaela as a “unicorn” and the sort of student you cannot help rooting for.

“It was all her; she knew exactly what she wanted to do from the get-go. She was brilliant,” Thompson said.

A recent graduate of Marist Catholic College, Penshurst, Michaela was named the 2025 young scientist of the year by the Science Teachers’ Association NSW for her assessment task research project. It was titled Assessing the Accuracy and Interpretability of a Recurrent Neural Network for Breast Cancer Classification and Molecular Subtyping using Ribonucleic Acid Sequencing Data.

Or in layman’s terms, “I trained my project for breast cancer classification.”

“It achieved 98 per cent accuracy in detecting and distinguishing between malignant and healthy, and if it was malignant, it also was able to distinguish between the four most common breast cancer subtypes.

“For that, it only achieved 70.6 per cent accuracy, which is somewhat high, but definitely not ready for clinical deployment.”

Michaela’s model also yielded a finding that was supported by recently emerging research.

“My AI model detected the gene transferrin [essential for moving iron through the body] as having quite a potentially significant impact when it comes to breast cancer development and gene dysregulation involved in breast cancer. I found all this secondary research that validated the results yielded by my model.

“I couldn’t believe it. I was like, ‘Oh my god, oh my god.’ I woke my Mum up!”

The project took nine months. “I call it my little baby,” Michaela said.

Michaela has a lifelong passion for STEM. When she was eight, she moved a boys’ periodic table T-shirt to the girls’ section of a clothing store because, as she said, “Why are we entrenching these gender biases?”

In primary school she joined a lunchtime coding club, organised by a female teacher.

“I kind of saw myself in her … It really inspired me,” she said.

In high school, she realised she could combine her software skills with her love of science, particularly biology. She was partly inspired by her father, who has had four different cancers over his life.

“I was just confused because we’ve got all this science and theory behind cancers … but why are we so stagnant with our treatment options?” she said.

Michaela’s father is doing well now, but he teases her about her focus on breast cancer.

Michaela’s award-winning research project.

“My Dad says to me, Michaela, out of all the cancers I have had, of course you choose the one I don’t have.”

She shrugs. “It was the data set that was most easily accessible,” but she was also motivated after watching her friend’s mother wait months to find out what type of breast cancer she had, and what treatment would be best.

“By the time you actually receive the information as to whether or not radiation or chemo will be best for her, the cancer’s already progressed, and it’s so costly,” she said.

“Unless we can find a way to deliver these treatments in a way that can ensure equitable access globally, we are not going to see the advances in healthcare and in our human health as a society that we strive for.”

Michaela will find out her HSC results next week and hopes to study a bachelor of engineering in bioinformatics at UNSW.

Start the day with a summary of the day’s most important and interesting stories, analysis and insights. Sign up for our Morning Edition newsletter.