Artificial Intelligence can see genetic mutations in the cancer cells that escape the human eye

Myelodysplastic syndrome (MDS) is a disease, which can develop into acute leukaemia. It starts in the stem cells in the bone marrow, disturbing the maturing and differentiation process of blood cells.

Now scientists at the University of Helsinki have demonstrated that artificial intelligence (AI) can effectively spot  genetic mutations in the cancer cells of MDS patients.

Dysplastic megakaryocyte in bone marrow of a patient with myelodysplastic disorder. This disease often evolves into acute leukemia. Image credit: Ed Uthman via Wikimedia (CC BY 2.0)

MDS is diagnosed from a bone marrow sample, which is then analysed for genetic changes. They may not be easy to spot in some cases, despite the methodology improving over time. One of the biggest advantages of AI is that it is able to look through huge pools of data very quickly and efficiently. That is why scientists tested it as an image analysis technique to screen microscopic images of MDS patients’ bone marrow samples.

The machine learning algorithm managed to analyse those samples and accurately identify the most common genetic mutations affecting the progression of the syndrome. In essence, AI is able to see genetic mutations related to MDS in tissue samples. In fact, AI sees the features that elude the human eye, making the diagnosis that much more accurate. Furthermore, this approach helps tracking the progression of the disease, because collecting quantitative data on cellular changes and their relevance to the patient’s prognosis becomes easier.

Scientists believe this is just one of the examples how AI could benefit medical diagnosis. It could help choosing the right treatment. AI can notice details that escape the human eye. And it can look through a lot of data quickly, spotting patterns and various hidden details. In the case of MDS, AI can spot cancer earlier and also help categorize patients into groups to determine the nature of the disorder in more detail.

Oscar Brück, one of the scientists in this study, said: “Image analysis helps us analyse large quantities of biopsies and rapidly produce diverse information on disease progression. The techniques developed in the project are suited to other projects as well, and they are perfect examples of the digitizing medical science”.

Myelodysplastic syndrome used to be called pre-leukemia or smoldering leukemia. Cancer develops in about a third of MDS patients. However, death often occurs without cancer diagnosis due to bleeding or infections. Scientists are trying to develop new and improved diagnosis and treatment options, but so far MDS can only be cured in some cases by a bone marrow transplant.


Source:  University of Helsinki