Kyrgyzstan to implement MRI-based tumor classification project proposed by ASI Digital Solutions finalist

The Futogy developers team from the Vologda Region came third in the second stage of Agency for Strategic Initiatives (ASI)’s World AI&Data Challenge, presenting a system for efficient classification of brain tumors through MRI. The system is intended to simplify and speed up the detection of tumors while minimizing the routine work and making diagnostics more precise. The pilot solution is currently being implemented in Kyrgyzstan.

According to Yana Kovalenko, Project Director, ASI Center for Digital Development, the solution is being deployed at one of the MRI clinics in the country, while a local IT park provides support. Furthermore, negotiations with the Ministry of Health of the Kyrgyz Republic are underway to industrialize the method and replicate it throughout the entire country.

“This is a great example of not only Russian regions, but foreign countries joining a project at its every stage. Kyrgyzstan brought an objective to the competition, and they got a solution. Now the Futogy team has an opportunity to implement their project abroad as a test, and then replicate it in Russia and other CIS countries,” Yana Kovalenko noted.

Model learning was based on data arrays including over 4,000 images of three types of tumors: meningioma, glioma, pituitary adenoma (class 3); 1st stage glioma, highly malignant glioma (class 2), tumor/no tumor (class 1). The model precision is over 70%. The solution also provides for DICOM files reading (the industry standard for creating, storing and transmitting digital medical images and patient examination documents).

“The choice of our objective was contingent on its importance now. Solutions like that, enabling us to detect life-threatening brain pathologies, are virtually absent at the market at the moment, despite the global nature of the problem. At the moment, pilot implementation of the project is underway in the customer region. Later on, we’re planning to partner with the local clinics and adjust our solution to deploy it in our own region,” Denis Lashenkov, Futogy Team Head, said.

The Team’s plans include expanding the solution to pathologies caught on mammography, CT and chest x-ray images, in addition to MRI. Currently, they are in consultations with representatives of the Vologda regional healthcare department, Severstal Health Clinic, Cherepovets Out-Patient Clinic No.1, Mallenom Systems, discussing acquisition of data needed to improve their model.

Mirbek Okenov, the tasksetter, Ulut Soft General Manager, noted that doctor have to spend considerable time on initial analysis of images, and that their analysis is often subjective.

“Our goal is to provide state healthcare institutions with an opportunity to use advanced technologies. Those technologies are expensive; at the same time, the DataMasters project empowers us to implement solutions like that at no extra cost, using open source code. Our task now is to implement the pilot solution with the Ministry of Health of the Kyrgyz Republic and to discuss possible improvements an adjustments with the team,” Mirbek Okenov clarified.

Any region can join the implementation process by leaving an application on the project website.