Novosibirsk Scientists Develop AI Brain Tumor Diagnosis Technology

The international Brain Tumor Segmentation (BraTS) Challenge - 2020 competition honored the work of a multidisciplinary team of Novosibirsk scientists. The team included researchers from the NSU Mechanics and Mathematics Department “Stream Data Analytics and Machine Learning Laboratory” (SDAMLL), Physicians at the Federal Center for Neurosurgery, and the Head of the Research Institute of Clinical and Experimental Lymphology, a branch of the Institute of Cytology and Genetics SB RAS. The Competition recognized the method and approach they developed for the analysis of medical data presented in the article, "Brain Tumor Segmentation and Associated Uncertainty Evaluation Using Multi-sequences MRI Mixture Data Preprocessing".

The BraTS competition has been conducted for several years and specializes in high and low grade gliomas (HGG and LGG). This year, the competition focused on three issues: segmentation of brain tumors, predicting overall patient survival based on preoperative MRI tomograms, and assessment of uncertainty measures in segmentation.

Bair Tuchinov, Head of the SDAMLL, described their work,

Our NSU_BTR team focused on the first and third issues. In the course of our research, we conducted more than 100 experiments. We used original methods that are based on an approach to processing MRI images using the basic sequences of T1-WI, T1-contrast, T2-WI, and T2-Flair. Our primary purpose for participating in international competitions is to compare our work with the best practices and to test our approaches.

During the Competition, the Team achieved positive results in the validation (verification) stage and the organizers invited them to publish their article in the Springer Nature edition. Their experience and best practices were applied in the major scientific project, "Development of technology for personalized diagnostics and development of recommendations for the treatment of neuro oncological diseases using neuroimaging methods based on artificial intelligence systems (deep machine learning)" that was supported by the Russian Foundation for Basic Research. Currently, seven articles have been published and three more are under review based on the results of the project. 

Professor Andrei Letyagin, Project Manager, renowned radiologist, and Doctor of Medical Sciences, summarized, “Our results can be used for personalized diagnostics in early detection and classification of neoplasms, predicting their status, and selecting a treatment.”