"Metal Detector" Algorithm Hunts Down Vulnerable Tumors

Scientists have developed an algorithm capable of functioning as a "metal detector" to identify vulnerable tumors, marking a significant advancement in personalized cancer treatment. This breakthrough could eventually be used to pinpoint cancer patients most likely to benefit from immunotherapy.

The algorithm, named PRRDetect, was created by researchers funded by Cancer Research UK and the NIHR at the University of Cambridge (Cambridge, UK). It works by recognizing specific patterns of mutations that emerge in cancer cells when they cannot repair DNA errors. PRRDetect represents a major leap in utilizing genomics – the study of all genetic information in a person’s DNA – to gain deeper insights into cancer. Until now, genomic sequencing (DNA reading) tests have typically focused on identifying particular mutations that drive cancer, such as those in BRCA genes, which doctors can target with specific drugs. However, this new algorithm goes beyond identifying isolated mutations by detecting broader mutation patterns, or ‘mutational signatures,’ that reveal deeper information about the cancer’s genetic makeup.

In this study, researchers focused on analyzing patterns in DNA that undergo indel mutations, a type of mutation where nucleotides (the basic units of DNA) are either inserted or deleted incorrectly in the DNA sequence. By examining nearly 5,000 tumors, the researchers identified unique mutation patterns indicative of ‘post replicative repair dysfunction’ (PRRd), a condition in which cells have faulty repair mechanisms. This data was then used to develop PRRDetect, a tool that scans genome sequences for PRRd-related mutations. Since tumors with PRRd are more responsive to immunotherapy, which activates the body's immune system to target cancer cells, the research team believes this tool could help translate genomic findings into more effective treatments for patients.

The study explored genomic data from a range of cancer types, including lung and brain tumors, as well as bowel, endometrial, skin, bladder, and stomach cancers. While there was existing evidence that PRRd was more prevalent in these cancers, PRRDetect may be the first effective tool to identify it. The research team is now working on clinical trials to evaluate how well PRRDetect can predict a patient's response to immunotherapy. Additionally, they plan to expand their genomic analysis to cover 20 cancer types, potentially leading to even more groundbreaking discoveries. Published in the journal Nature Genetics, the study also revealed new insights into the potential causes of cancer. Ten of the indel mutation patterns were linked to known carcinogens, such as tobacco use and UV light exposure, while 19 others were previously unidentified. These novel patterns could suggest unknown causes of cancer or growth mechanisms that could be targeted by new treatments.

“Cancers with faulty DNA repair are more likely to be treated successfully,” said Professor Serena Nik-Zainal, who led the first study into the new algorithm. “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.”