Diagnosing cancer with artificial intelligence

“Inside each of us, cancer cells are growing,” says Te Herenga Waka—Victoria University of Wellington PhD student Qurrat Ul Ain.

“Luckily, it’s not as bad as it sounds, because most of us have healthy and active immune systems. However, identifying cancer early is still really important—if melanoma is diagnosed at any early stage, for example, the survival rate increases to 92%.”

“Melanoma is the most serious type of skin cancer anyone can get,” she says. “It can progress quickly and be life-threatening. Over 4000 New Zealanders are diagnosed with melanoma every year, and around 300 New Zealanders die as a result.

“Also, tumour cells are clever—over time, they can change their appearance, allowing them to avoid the immune system and spread to other parts of the body—which can make them hard to find and treat.”

To counteract these problems, Qurrat is working on developing a computer-aided diagnostic system for skin cancer.

“My research focuses on using Genetic Programming, which is an Artificial Intelligence technique, to identify melanoma, as well as eight other types of skin diseases,” Qurrat says. “This involves using common melanoma or skin disease features like colour, texture, and lesion shape to teach the AI to identify images of melanoma amongst images of normal skin or other skin diseases.”

So far, Qurrat’s AI has had an accuracy rate of 98% in identifying melanoma in images.

“This is exciting progress,” she says. “There have been a lot of other developments in AI recently which mean that automatic early skin cancer detection could soon be a real possibility.

“The detection methods I have developed can directly help dermatologists identify cancer quickly in skin images. This technology could also be extremely useful in rural areas where people may not have easy access to cancer care, as people could use images taken on a smartphone or camera with this AI software to identify any visual signs of skin cancer.”

Qurrat always wanted to pursue a research area that would help solve real-world problems, so this project was a natural fit for her, working alongside Dr Bing Xue, Professor Mengjie Zhang, and Dr Harith Al-Sahaf from the University’s School of Engineering and Computer Science. She also enjoys the opportunity to tutor students at the University, helping them learn and achieve and hopefully also contribute to solving real-world problems in their own way, Qurrat says.

“It can be frustrating to do research sometimes when your research doesn’t quite work out, but in a lot of ways that’s the point of research—constantly searching for new answers and methods,” Qurrat says. “And when your algorithm works and helps solve a real-world problem like cancer detection you feel over the moon!”

Qurrat also faced a number of other challenges during her research, including moving all the way from Pakistan to New Zealand and looking after her new-born daughter. She says the support of her husband and her PhD supervisors, as well as the hard-work and passion she inherited from her parents, greatly helped her out on her PhD journey.

“I’ve found a beautiful home for my family and the chance to pursue my ambition of spreading humanity through kindness and humility, along with learning to advance in my technological goals with opportunities that know no bound,” Qurrat says.

Qurrat plans to continue her research in cancer detection, as well as continue teaching.

“I will continue doing research by developing AI solutions for cancer detection, specifically focusing on to breast cancer detection and helping doctors diagnose cancers at an early stage,” she says.