A test using artificial intelligence (AI) to detect tumour DNA in a patient’s blood has shown to be much more effective at predicting cancer recurrence, and has the potential to improve cancer care and extend lifespans by detecting cancer recurrence earlier.
Beating cancer the first time around is a remarkable achievement, one to be celebrated for sure. But this is often not the end of the line, as many cancers can return time and time again.
The best chance for successful treatment for cancer recurrence lies in how early that recurrence can be detected. And now, researchers from Weill Cornell Medicine in the US have harnessed AI to analyse the blood of cancer patients in remission to pick up the presence of cancer far earlier than current methods.
While blood tests for cancer recurrence are not new – known as a liquid biopsy – it’s the addition of AI analysis that takes the test to the next level.
Improve cancer care
The new technology has the potential to improve cancer care with the very early detection of recurrence and close monitoring of tumour response during treatment.
In the study, published in the journal Nature Medicine, the researchers showed that they could train a machine learning model – a type of AI – to detect strands of tumour DNA (known as ctDNA) in patient blood tests with a very high level of accuracy.
They successfully demonstrated the technology could pick up recurrences of several different types of cancer including with lung cancer, melanoma, breast cancer, colorectal cancer and precancerous colorectal polyps
Dr Dan Landau, co-author of the study, says the technology could revolutionise how cancer is monitored and improve life expectancy for cancer patients.
“We were able to achieve a remarkable signal-to-noise enhancement, and this enabled us, for example, to detect cancer recurrence months or even years before standard clinical methods did so,” he says.
How does it work?
In the past, liquid biopsy technology was fairly limited in scope, mainly looking for relatively small sets of cancer-related gene mutations. But Dr Landau says these are often too sparsely present in blood to be reliably detected, which leads to many recurring cancers being missed until they have progressed further.
A few years back, Dr Landau and team developed a different approach to liquid biopsies, focusing on sequencing the whole DNA genome in the blood sample, rather than focusing on a small part.
This allowed them to get a much broader picture of what’s happening in the blood, providing them with more DNA ‘signal’ to work with.
In this new study, this DNA sequencing has been analysed by a machine learning AI the team calls MRD-EDGE, specifically looking for subtle patterns that indicate cancer.
For example, in one of their tests, the researchers got MRD-EDGE to analyse the blood 15 colorectal cancer patients post-treatment. The system predicted that nine of them had cancer remnants circulating in their blood.
No false positives
Months later, five of those patients were indeed found to have recurring cancer using more traditional testing methods. Less than the nine predicted sure, but what was really important was that there were no false positives – none of the samples MRD-EDGE said were cancer-free went on to develop a recurrence.
The system showed similar levels of sensitivity in tests on early-stage lung cancer and triple-negative breast cancer patients, with early detection of all but one recurrence.
Dr Landau says their system fills a gap in cancer diagnosis and treatment and should have a positive impact on patient diagnostic outcomes.
“On the whole, MRD-EDGE addresses a big need, and we’re excited about its potential and working with industry partners to try to deliver it to patients,” he says.
Have you ever had a cancer return? What other conditions might this kind of AI analysis be able to predict? Let us know in the comments section below.
Also read: AI tool successfully used to warn of heart attacks
