Scientists from Nizhny Novgorod have developed an artificial intelligence capable of diagnosing the type of brain tumor. The development was carried out by scientists from the Research Center in the field of artificial intelligence and the Institute of Aging Biology of the university.
The system shows a high level of accuracy in the diagnosis of gliomas, the most aggressive type of brain tumors. It determines the subtype of the tumor, predicts patient survival, and helps in choosing treatment tactics.
Ivanchenko added that traditional diagnostic methods often lead to errors. The fact is that classical diagnostics of gliomas does not always allow to quickly and accurately determine the type of tumor. The new system is based on 13 key genes, which significantly improves results compared to outdated histological methods.
Our tool evaluates the tumor based on transcriptomic analysis data, when we see which genes are turned off, which are turned on, and we can assess the level of activity of these genes in a particular patient.
Ivanchenko explained that the doctor can double-check the decision of the neural network, agree or disagree with it. This opens up new opportunities for personalized medicine and allows in the future to create clinical test systems for rapid assessment of the level of gene expression.
