Published on : Nov 20, 2018
Researchers have invented a new blood test that combines liquid biopsy with artificial intelligence to detect cancer in very early stages so treatment can be started much before symptoms start to appear. The results, which were published in the journal Nature, could prove to be highly promising for arresting the development of cancer before it spreads and becomes morbid.
One of the key issues associated with making early detection a plausible possibility is the problem of how that extremely hard to find cancer-specific mutation can be traced in the blood, especially in the early stages of cancer when the blood has very small amount of tumor DNA.
The researchers behind the new test focused on epigenetic alternations and not on mutations to identify thousands of modifications specific to specific cancers. Machine learning was then applied to the data and classifiers capable of identifying the presence of cancer DNA in the blood samples and to determine the type of cancer.
The type and origin of cancer was tracked by the scientist by comparing tumor samples from 300 patients across seven cancer sites (colorectal, pancreatic, lung, bladder, kidney, breast, and leukemia) and blood samples from healthy donors. Cell-free floating DNA in the blood plasma was analyzed for the purpose.
The team has now successfully profiled and matched over 700 tumor and blood samples from more varieties of cancer. To further validate the approach, the researchers plant to analyze data from larger health research studies that are already being undertaken, where blood samples from patients were gathered over the course of years to month before the diagnosis of cancer.