In a cutting edge study, the research team have demonstrated the feasibility of multi-cancer early detection using patient blood samples, known as liquid biopsies. TriOx uses a novel method for analysing these liquid biopsies that integrates numerous genetic features through machine learning to predict the likelihood of a cancer diagnosis.
Liquid biopsies have become an important focus of research in recent years as a potential adjunct to cancer screening programmes. Current screening programmes are limited to certain cancers that make up less than 30% of all cases. Uptake of screening initiatives is often low, particularly where the tests are invasive.
Liquid biopsies could overcome these limitations and enable the detection of several different cancers at their earliest stages. They are based on the detection of tiny fragments of circulating cancer DNA in the blood that carry specific markers, making them distinct from the DNA coming from healthy cells. However, most liquid biopsy tests currently in the clinic only look at a small fraction of the cancer genome, which limits how well they can detect cancer.
A new approach to liquid biopsy testing from Prof. Anna Schuh and her team instead looks at multiple cancer signals across the entire genome, in order to make the tests more sensitive. To do this, the team used a cutting-edge DNA sequencing technique called TAPS, paired with a machine learning-based test called TriOx. Together, these technologies analyse and integrate various cancer signals to give an overall score reflecting the amount of cancer DNA found in the blood.
To test the diagnostic accuracy of the approach, researchers analysed blood samples from a cohort of symptomatic patients referred to an NHS rapid diagnostic clinic for cancer testing. The blood test was shown to be highly sensitive, catching cancer signals from all 6 different types of cancer present in the cohort, from the earliest stages of disease. The combination of DNA features in TriOx was shown to better detect cancer, compared to testing each feature individually. Importantly, those who did not have cancer reliably tested negative.
While the test is still in the development phase, the study demonstrates the promise of blood-based early cancer detection, a technology which could revolutionize screening and diagnosis practices. When people present to the GP with potential non-specific cancer symptoms, it can be difficult to determine their cancer risk based on these symptoms alone, leading to high rates of referrals for further investigation. Only 7% of NHS GP referrals result in a cancer diagnosis. A simple blood test like TriOx could help to guide who should receive further testing, reducing rates of unnecessary screening and helping to catch cancer cases earlier. Notably, earlier diagnosis significantly increases the likelihood of cancer being cured, as well as being much cheaper for healthcare systems.
This is the first study to explore whether combining numerous DNA features in a single test can improve the detection of cancer from blood samples and provides valuable evidence to inform further research. The test will now undergo in-depth clinical testing in a much larger group of patients to validate its performance in different cancers and at different stages. They will also investigate if including additional DNA features helps to further increase the sensitivity of the test.
The paper, ‘Multimodal cell-free DNA whole-genome TAPS is sensitive and reveals specific cancer signals’, is published in Nature Communications.