Researchers at University Health Network in Canada have devised a new diagnostic test powered by artificial intelligence (AI) technology to detect and classify cancer at the earliest stages.
The approach combines liquid biopsy, epigenetic alterations and machine learning into a blood test.
Early identification of cancer is expected to aid in better treatments, even before the appearance of symptoms.
University Health Network Princess Margaret Cancer Centre senior scientist Daniel De Carvalho said: “A major problem in cancer is how to detect it early. It has been a ‘needle in the haystack’ problem of how to find that one-in-a-billion cancer-specific mutation in the blood, especially at earlier stages, where the amount of tumour DNA in the blood is minimal.”
To develop the test, the team carried out profiling of epigenetic alterations and was able to identify several modifications that were unique to each cancer type.
A big data technique was then used to apply machine learning for creating classifiers that could detect a cancer-derived DNA presence in blood samples and determine the cancer type.
In order to track the cancer origin, the researchers analysed cell-free circulating DNA in the blood plasma of 300 patient tumour samples from seven disease sites and compared them with samples from healthy donors.
Findings showed that the circulating plasma DNA in every sample matched the tumour DNA. Since then, the team profiled and matched more than 700 tumour and blood samples from various cancer types.
The findings have been published in scientific journal Nature.
Currently, the researchers are planning to further evaluate the AI-based diagnostic test in analysis of data from large population research studies, where blood samples were obtained months to years prior to cancer diagnosis.
These validation studies will then be followed by prospective studies for cancer screening.