Content |
History
2024
Launch of neural network to assess rejection of transplanted kidney
Russian scientists have developed an artificial intelligence model to automatically assess kidney transplant rejection. The creation of an innovative neural network became known on October 17, 2024.
The development is the result of the joint work of specialists of MedTech AI LLC (a joint venture of PJSC VimpelCom and Sechenov University) and doctors from NCC No. 1 RNTCH named after B.V. Petrovsky. The model is designed to automatically assess fibrosis and interstitial infiltration in histological studies of kidney transplant.
The neural network is able to identify and assess signs of rejection of the transplanted organ, which can become an important tool of the "second opinion" to support medical decision-making. This development is especially relevant in light of the fact that in 2023 in Russia 1789 kidney transplantation operations were carried out, which is about two-thirds of all transplants.
Konstantin Romanov, Director for Artificial Intelligence and Digital Beeline Products, General Director of MedTech AI LLC, noted that artificial intelligence will help standardize research and increase the accuracy of determining further therapy.
The model demonstrates high accuracy in predicting clinical scores according to the Banff classification, making it a promising aid tool for pathologists. According to the head of the Laboratory of Digital Microscopic Analysis of the Institute of Regenerative Medicine, NTPB, Sechenov University Aleksei Fayzullin, the metrics calculated by the neural network make it possible to distinguish between borderline cases and cases with acute cellular rejection with an accuracy of more than 95%.
The head of the pathological department of the NCC No. 1 of the RNCh named after Academician B.V. Petrovsky Svetlana Solovyova emphasized that automatic determination not only of the presence, but also of the area of pathological changes in the tissue can help increase the accuracy and reproducibility of diagnostics.
It is planned to conduct clinical trials of the model and its further registration for use in medical institutions.[1]
Creating a Company
On February 20, 2024, VimpelCom and Sechenov University announced the creation of a joint company to develop medical systems with artificial intelligence. The new structure was called "MedTech AI."
According to the press service of the university, the joint venture will develop and implement software with AI algorithms, as well as research activities. The company will also register developments as medical devices and enter them into the register of Russian software, participate in tenders for government investments and grants, conduct consulting, and attract industrial and technological partners.
In particular, MedTech AI will create AI solutions for orthopedic tasks. As noted in Sechenov University, the use of AI allows you to more accurately and quickly determine diseases of the musculoskeletal system, including scoliosis, coxarthrosis and arthritis. AI can analyze medical images, such as radiographic and MRI images, and detect hidden anomalies that would be invisible to the human eye. New projects will improve the diagnosis, treatment and prediction of diseases of the musculoskeletal system, thereby significantly improving the quality of life of patients.
Konstantin Romanov, Director General of MedTech AI, has been appointed Director of Artificial Intelligence and Digital Products at VimpelCom. According to him, the team on artificial intelligence and big data at VimpeCom has been collaborating with Sechenov University since 2020. During this time, joint efforts have created international-class AI solutions for the early diagnosis of hip pathologies in MRI studies, an AI model for the diagnosis and analysis of histological kidney preparations for the transplantology industry, and the development of solutions in the field of pathomorphology and remote monitoring of patients.[2]