Artificial intelligence reveals prostate cancer is not just one disease

Research published in Cell Genomics today has shown that prostate cancer, which affects one in six men in the UK in their lifetime, includes two different subtypes of the disease, also known as evotypes.  

This discovery was made by using artificial intelligence (AI) to help unlock new discoveries about the evolution of prostate cancer.  

Just as humans – or Homo sapiens – have evolved into a species over time from our ancestors, the development of cancer is also an evolutionary process. By looking at the evolutionary tree of cancer, we can find out a lot about the disease, and it can even help us develop new treatments. 

“Our research demonstrates that prostate tumours evolve along multiple pathways, leading to two distinct disease types,” said lead researcher Dr Dan Woodcock, of the Nuffield Department of Surgical Sciences at the University of Oxford.   

“This understanding is pivotal as it allows us to classify tumours based on how the cancer evolves rather than solely on individual gene mutations or expression patterns.” 

The power of AI 

The study, funded by Cancer Research UK and Prostate Cancer Research, involved researchers working together as part of an international consortium, called The Pan Prostate Cancer Group, from the University of Oxford, the University of Manchester, the University of East Anglia and the Institute of Cancer Research, London.  

The researchers analysed the DNA of prostate cancer samples in 159 patients using whole genome sequencing – a comprehensive way of looking at the entirety of someone’s genetic material.  

They then used an AI technique known as neural networks to compare the DNA of the different samples.  

This identified two distinct cancer groups among these patients.  

These two groups were then confirmed using two other mathematical approaches applied to different aspects of the data. And importantly, this finding was also validated in other independent datasets from Canada and Australia.  

The researchers were then able to integrate all of this information to generate an evolutionary tree showing how the two subtypes of prostate cancer develop, ultimately converging into two distinct ‘evotypes’. 

“This study is really important because until now, we thought that prostate cancer was just one type of disease. But it is only now, with advancements in artificial intelligence, that we have been able to show that there are actually two different subtypes at play,” said Professor Colin Cooper, from UEA’s Norwich Medical School. 

“We hope that the findings will not only save lives through better diagnosis and tailored treatments in the future, but they may help researchers working in other cancer fields better understand other types of cancer too.” 

Revolutionising the future of prostate cancer treatment

It’s hoped that these findings could revolutionise how prostate cancer is diagnosed and treated in the future. 

Crucially, the team’s collaboration with Cancer Research UK aims to develop a genetic test that, when combined with conventional staging and grading, can provide a more precise prognosis for each patient, allowing tailored treatment decisions. 

Prostate cancer is the most common cancer in men in the UK with around 55,000 cases a year. While prostate cancer is responsible for a large proportion of all male cancer deaths, it is more commonly a disease men die with rather than from. Research like this is key, as knowing when to avoid unnecessary treatment is also important, to help spare men from related side-effects such as incontinence and impotence.  

“We simply don’t know enough about what a prostate cancer diagnosis means at present – there are many men who have disease which is or may become aggressive and being able to treat aggressive disease more effectively is critical,” said Dr Naomi Elster, Director of Research at Prostate Cancer Research. 

“But on the other side of the coin are the too many men who live with side effects of cancer treatment they may never have needed.  

“These results could be the beginning of us being able to take the same ‘divide and conquer’ approach to prostate cancer that has worked in other diseases, such as breast cancer.”