Scientists in China have developed a new method of testing for cancer, powered by AI, that could help diagnose the disease with a single drop of blood – even before patients have symptoms
A brand new method of testing powered by AI could one day be used to diagnose the early stages of three of the deadliest cancers – using just a single drop of dried blood.
Chinese researchers have revealed that in primary testing, the tool was able to detect cancer in patients already diagnosed around 82% to 100% of the time, by recognising certain chemicals in the blood droplet samples. It was also able to distinguish between patients who didn’t have cancer and those with pancreatic, gastric or colorectal cancer.
The new tool is powered by a form artificial intelligence known as machine learning which is able to detect metabolites – substances left over in the liquid part of the blood, the serum, after metabolism. These by-products, depending on their type and quantity, act as ‘biomarkers’, potentially flagging the presence of cancer cells in the body. Writing in the journal Nature Sustainability, the researchers claim that by detecting these specific chemicals in the blood, the test has the capability to accurately identify cancer in patients approximately 82% to 100% of the time.
This use of metabolites has been proposed as a potential new and effective way to detect and diagnose cancer in the early-stages of the disease when chances of survival for the patient are higher and they might not yet be experiencing symptoms. Despite pancreatic, colorectal and gastric cancers being among the most deadly cancers in the world, none have a specific diagnostic blood test and are currently usually identified through the use of medical imaging or surgical procedures.
According to Cancer Research UK, pancreatic and bowel cancers are amongst the top five most common causes of cancer deaths in the country and are predicted to kill 14,620 men and and 13,791 women between 2023 and 2025. In theory, this test would require just a single, tiny spot of blood in order to to diagnose these diseases, according to the scientists in China who have developed it.
Dr. Chaoyuan Kuang, an assistant professor at the Albert Einstein College of Medicine and an oncologist told Live Science that these samples of blood serum can be “collected, stored and transported at much lower cost and with much simpler equipment”, helping to “democratise the availability of cancer early detection testing across the world”.
The new test is being proposed as a sustainable and environmentally friendly way of detecting cancer. The study shows that the method is more effective at testing the small dried blood samples that liquid blood. In one experiment, using the dried blood spots enabled them to detect 81.2% of cases of pancreatic cancer, compared to 76.8% using liquid blood samples.
Although Kuang, who is not involved in the research, suggests that the testing won’t be available to patients on a widespread basis for a number of years and requires more testing, scientists behind the method say that if used in widespread cancer screening programmes in the future, it could make a huge difference.
Extensive trials of the method are still needed in larger diverse populations, as currently, only small proof-of-concept tests have been carried out using a few hundred samples. The researchers exclusively evaluated the machine learning model on individuals already diagnosed with cancer and have not yet assessed its efficacy as a genuine diagnostic tool.
Due to the low cost nature of the test, it could be used in remote areas where access to general medicine and cancer testing is limited and resources are limited, helping those disproportionately-high barriers to medical care. By 2030, its expected that 75 per cent of all cancer deaths will occur in low- and middle-income countries.