One of the reasons why cancer is fatal is that it is most often diagnosed at a late stage. Therefore, one of the tasks facing scientists is to develop a simple, fast, inexpensive and effective technology for diagnosing cancer. Artificial intelligence can help with this, and some success has already been achieved in this direction. Chinese scientists have created a system that detects certain types of cancer from a drop of blood in a few minutes.
Cancer diagnosis using artificial intelligence
A new tool created by scientists makes it possible to detect malignant tumors of the pancreas and stomach, as well as colorectal cancer. These types of cancer are the most common and at the same time the most deadly, especially pancreatic cancer, which is generally incurable, as well as colorectal cancer.
Unfortunately, effective methods for diagnosing these diseases still do not exist. In particular, there are no technologies yet that would allow us to diagnose diseases using a blood test, much less do it ourselves. To detect cancerous tissue, doctors rely mainly on imaging or surgical procedures, which means that cancer is detected at a fairly advanced stage.
However, in the United States, the diagnosis of pancreatic cancer is currently approved using one biomarker, known as CA19-9. However, existing tests detect the disease only 80% of the time, and only after patients develop symptoms of the disease. Before symptoms appear, the effectiveness of this method does not exceed 13%.
The new technology, developed by Chinese scientists, is based on a machine learning system that analyzes many metabolic byproducts, or metabolites, in the blood. We have already told you that metabolism changes during cancer. By the way, this becomes one of the reasons that people begin to lose a lot of weight.
Thus, the metabolites contained in the blood serve as a kind of “biomarkers” that can be used to determine the presence of cancer in a person. An important advantage of this diagnostic method is that cancer can be detected in the early stages, since certain metabolites in cancer appear immediately, even when a person does not yet have any symptoms.
New effective technology for diagnosing cancer
In a study conducted by scientists, a new AI-based diagnostic tool was able to distinguish blood donors without cancer from donors with cancer. The effectiveness of the test was in the range of 82-100%. The authors report this in the journal Nature Sustainability. For diagnosis, the system requires less than 0.05 milliliters of blood.
The most interesting thing is that this instrument can carry out diagnostics using both liquid and dried blood. In one test, the system was able to detect pancreatic cancer even from dried blood spots in 81.2% of cases. Dried blood can be stored and transported at much lower cost. According to scientists, this would make tests available throughout the world, including poor countries in Africa.
In addition, according to the researchers, the technology they developed will reduce the estimated proportion of undetected cancer cases by 20-50% in rural China. Of course, we are talking about the three types of cancer mentioned above. This will increase survival rates or at least prolong life for many people.
But, unfortunately, in practice, new The method will not be available for use anytime soon. Previously, it must undergo many tests using a large number of analyses. The current study tested the system on only a few hundred blood samples. In addition, people whose diagnosis had already been confirmed took part in the work.
Do not forget to subscribe to our Zen and Telegram channels so as not to miss the most interesting and incredible scientific discoveries!
Finally, let us remind you that this is not the first time that scientists use AI to diagnose cancer. Previously, we talked about a method that is based on the analysis of “cancer DNA”. You can read more about this at the link.