Russian-designed program module COVID-MULTIVOX, capable of analyzing the degree of lung damage on the basis of tomography scan images without human involvement has successfully passed the first tests, the press-service of the M.V. Lomonosov Moscow State University, one of the participants in the project, said on Thursday.
The module is a joint product of the Research Institute of Nuclear Physics and the fundamental medicine department of the Moscow State University, Moscow’s state clinical hospital N. 52 and the group of companies Gammamed.
«The computerized module COVID-MULTIVOX, designed for assessing lung damage on the basis of CT scan images without a diagnostics specialist taking part has undergone successful tests and produced the first results… Researchers are going to integrate this module with clinical and laboratory data in order to create a system of support for making medical decisions and eventually to share this product with regions,» the news release runs.
The module uses a combination of various artificial intelligence methods, including computer vision, neural networks, decision-making trees and special methods of processing medical images, including morphometry and vascular analysis. The program is capable of visualizing and creating a 3D image of the lungs.
Specialists at hospital No. 52 say the program will allow for optimizing the interpretation of CT scan images. It is capable of promptly and quickly detecting the damaged areas. Also, it can be used to forecast pneumonia’s likely trends and outcome, as well as to compare the effectiveness of antivirus therapy methods and respiratory support. The product is currently in the implementation phase at hospital N. 52 and the MSU’s Medical Science and Education Center.
«We worked as a team. Doctors at hospital N. 52 — the initiators of this product — set tasks, while the Research Institute of Nuclear Physics and Gammamed delivered software solutions within tight deadlines. Doctors then carried out checks and coordination — during the breaks between their shifts in the «red zone», in fact without taking a minute for rest. A handy diagnostic tool is the end product. Yet, I would like to stress that the module is merely an assistant — competent, impartial and reliable. But it is up to the doctor to make a final solution in the end,» the chief of the Research Institute of Nuclear Physics’ laboratory of medical computer systems, Andrei Gavrilov, is quoted in the news release as saying.
In the longer term, researchers plan to integrate the COVID-MULTIVOX module with clinical and laboratory data with the am of creating a system of support for medical decisions. Also, there are plans for pairing this system with telemedicine technologies the MSU’s Research Institute of Nuclear Physics works on in cooperation with partners.