Tumours vulnerable to immunotherapy could be hunted with algorithm developed in Cambridge

An algorithm described as acting like a metal detector has been created by Cambridge researchers to hunt down vulnerable tumours.

The development could eventually have major implications for the treatment of cancer.

Scientists analysed the full DNA sequence of 4,775 tumours from seven types of cancer, using data from Genomics England’s 100,000 Genomes Project.

They used the data to create an algorithm able to identify tumours with faults in their DNA that make them easier to treat.

Called PRRDetect, the algorithm could ultimately enable medics to work out which patients are more likely to have successful treatment, leading to more personalised and effective approaches.

The scientists, at the University of Cambridge and NIHR Cambridge Biomedical Research Centre, published their findings in Nature Genetics

Lead author Prof Serena Nik-Zainal, from the Early Cancer Institute at the University of Cambridge, said: “Genomic sequencing is now far faster and cheaper than ever before. We are getting closer to the point where getting your tumour sequenced will be as routine as a scan or blood test.

“To use genomics most effectively in the clinic, we need tools which give us meaningful information about how a person’s tumour might respond to treatment. This is especially important in cancers where survival is poorer, like lung cancer and brain tumours.

“Cancers with faulty DNA repair are more likely to be treated successfully. PRRDetect helps us better identify those cancers and, as we sequence more and more cancers routinely in the clinic, it could ultimately help doctors better tailor treatments to individual patients.”

The researchers sought patterns in DNA created by ‘indel’ mutations, in which letters are inserted or deleted from the normal sequence.

Cancers with faulty DNA repair mechanisms - known as ‘post-replicative repair dysfunction’ or PRRd - featured unusual patterns of indel mutations, which the team used to develop PRRDetect.

PRRd tumours are known to be more sensitive to immunotherapy - a treatment in which the immune system is used to attack cancer cells.

The hope is the new algorithm will act like a metal detector to find patients more likely to benefit from it.

Prof Nik-Zainal previously carried out an examination of genomes of tens of thousands of people that revealed previously unknown patterns of mutations linked to cancer.

The latest research focused on cancers with a higher proportion of tumours with PRRd, including bowel, brain, endometrial, skin, lung, bladder and stomach cancers.

Some 37 patterns of indel mutations were found across the seven cancer types, 10 of which were already linked to known causes of cancer, such as smoking and exposure to UV light and eight of which were linked to PRRd.

The remaining 19 patterns were new and the researchers say they could be linked to causes of cancer not fully understood yet or mechanisms within cells that can go wrong when a cell becomes cancerous.

The research was funded by Cancer Research UK (CRUK) and the National Institute for Health and Care Research (NIHR).

Dr Iain Foulkes, executive director of research and innovation at CRUK, said: “We can now get full readouts of tumour DNA much more easily, and with that comes a wealth of information about how an individual’s cancer can start, grow and spread.

“Tools like PRRDetect are going to make personalised treatment for cancer a reality for many more patients in the future. Personalising treatment is much more likely to be successful, ensuring more people can live longer, better lives free from the fear of cancer.”

NIHR scientific director Mike Lewis said: “Cancer is a leading cause of death in the UK so it's impressive to see our research lead to the creation of a tool to determine which therapy will lead to a higher likelihood of successful cancer treatment.”

Prof Matt Brown, chief scientific officer at Genomics England, said: “Genomics is playing an increasingly important role in healthcare and these findings show how genomic data can be used to drive more predictive, preventative care leading to better outcomes for patients with cancer.

“The creation of this algorithm showcases the immense value of whole genome sequencing not only in research but also in the clinic across multiple diverse cancer types in advancing cancer care.”