Tatyana Kobyakova, a graduate of the Institute of Physics and Engineering in Biomedicine at MEPhI (PhysBio), has developed a project that will help in the diagnosis of brain metastases. Her project on automatic analysis of MRI images using artificial intelligence won a prize at the recently concluded Russian engineering competition, and in the future could become a real solution for oncology treatment.

Brain metastases are secondary tumors that occur when cancer spreads to the brain from other organs. They are diagnosed in 20-40% of cancer patients. Metastases cause serious symptoms: headaches, convulsions, memory and coordination disorders. Without timely diagnosis, they can lead to serious consequences. Traditional diagnostics of metastases is difficult: radiologists spend up to 40 minutes analyzing a single MRI scan to identify and evaluate tumors. Small lesions (a few millimeters) often go unnoticed, and the human factor increases the risk of errors. Even in modern clinics, this process remains labor-intensive.

Tatyana Kobyakova created a computer product that reduces the time of MRI analysis up to 5 minutes and increases the accuracy of diagnostics. Her program uses a neural network that automatically detects and highlights metastases, including the smallest ones. The system compares before and after treatment images, tracks tumor changes, and generates reports that can be easily integrated with medical systems or electronic patient records. This solution saves doctors time, minimizes errors, and creates opportunities for research, such as studying the effect of radiation doses on metastasis control.

The success of the project depends on the support of the Russian Engineering Competition, operated by MEPhI. Competition helps young scientists turn ideas into finished products. At the acceleration stage, the PhysBio student collaborated with experts from the Business Center of Neurosurgery clinic, who helped to adapt the program to the clinics’ requirements and also provided images of 600 patients for training the neural network. The competition's trainings and master classes allowed to refine the algorithms and prepare the system for implementation. Doctors highly praised the project for its ability to speed up diagnostics and increase its reliability.

Tatyana's project makes the diagnostic process fast, accurate and accessible. It demonstrates how artificial intelligence can serve medicine, and how the ideas of young scientists can find practical application.

In the future, Tatyana plans to improve the system by adding more data to enhance its accuracy and implement it in clinics across Russia. «There have never been such systems addressed specifically for brain tumors. There was image analysis and tumor segmentation, but no one was generating statistics depending on time. And it is important to see the dynamics in oncology. This is the strength of my program. It is now being licensed and will be implemented in clinics on a commercial basis» – explains Tatyana Kobyakova.

Thanks to the Russian Engineering Competition, the MEPhI student gained access to collaboration with leading medical centers and support for further research. This project shows how technology and science can come together to solve complex medical problems, giving patients a chance for more effective treatment.