The Moscow clinics began to use AI for detection of cancer on mammograms

2020

Use of a system for detection of cancer on mammograms

At the end of July, 2020 in the Moscow clinics began to use artificial intelligence for detection of cancer on mammograms. It is about the Tsels system which works at a basis of neural networks and analyzes digital medical images, revealing on them the slightest suspicious changes in structure of tissues of mammary gland.

Service is created to help radiologists to interpret results of researches, including in the course of screening of a breast cancer. We are sure that the new technology will increase quality of diagnosis of oncological diseases — the Deputy Mayor of Moscow Anastasia Rakova whose words are cited by TASS explained to journalists.

The Moscow clinics began to use AI for detection of cancer on mammograms

All procedure of recognition of researches lasts no more than 10 minutes. The algorithm has hypersensibility and reveals any little change. The doctor interprets finds and solves whether they have diagnostic value and whether it is possible to suspect in this case a breast cancer or concerns are vain.

According to on the official site of the Mayor of Moscow, results of screening of a breast cancer undergo double control — two doctors should interpret researches. Usually experts who well understand pathological changes of a mammary gland participate in it. "Цельс" helps specialists to notice and estimate the most minor changes — mikrokaltsinata, sections of reorganization of structure of fabric, asymmetry.

According to the chief specialist on beam and instrumental diagnostics, the director of the Center of diagnostics and Sergey Morozov's telemedicine, artificial intelligence will not replace doctors, but will be able to be to them good help, helping doctors to notice pathology in time.

By the end of July, 2020 AI services are connected to 48 hospitals and 193 clinics and their branches in Moscow.^[1]

System implementation of "Tsels" in departments of radiodiagnosis in Moscow

An AI system for the analysis of medical images "Tsels" begins work in departments of radiodiagnosis of the city of Moscow in two directions — fluorography and mammography. On July 14, 2020 the company Medical screening of a system which developed this technology and signed the contract with the Moscow Department of Healthcare reported about it.

For an experiment which will last till December 31, 2020 the Moscow authorities approved issue of 16 grants to developers of the medical software for the total amount of 147.3 million rubles to approve methods of decision support on the basis of analysis results of data.  These company funds can get in three months a job or at achievement of the minimum quantity of the processed researches. Financing of all project from own means is estimated by the company at 45 million rubles.

The Moscow medical institutions implement AI for the analysis of photofluorograms for 45 million rubles

Solutions of service passed calibration test and got admission to application in clinical practice — now doctors will use service for setting of real diagnoses. "Цельс" automates a number of processes in work of the radiologist, thereby increasing the accuracy and speed of diagnostics. The neuronet finds symptoms of pathologies in pictures and ranges researches on priority that the doctor first of all browsed pictures of those patients which with the largest probability have a pathology. It allows to carry out quicker additional examination, to make the diagnosis and to begin treatment.

Use of AI services takes place the Experiment framework on use of the innovative technologies for the analysis of medical images started by the Moscow Department of Healthcare and the Center of diagnostics and telemedicine. The experiment estimates possibilities of use of neural network technologies in work of radiologists. For participation in it services need to take place several stages of check on compliance to the stated parameters on the basis of the real verified data.

For us the Experiment is an opportunity to receive a feedback from doctors and to improve our solution. We analyze data on work of service, on the basis of it we finish and we develop a system. For July, 2020 the command of service conducts work on other directions of diagnostics. MVP (the product having the minimum functionality — a comment) systems for classification of patients with COVID-19 is already ready. We set a goal to implement functionality of segmentation on KT COVID in very short time, until the end of summer of 2020, and by the end of 2020 we are going to make the serious competition to existing solutions in this direction, - the operating officer of Tsels service Nikita Nikolaev

In addition to artificial intelligence for the analysis of photofluorograms, in departments of radiodiagnosis of the Moscow medical institutions also  implemented  technology of voice input of the medical conclusions. The cost of connection of one microphone to a workplace of the radiologist is estimated at 20 thousand rubles, and the quality of speech recognition is 93%.^[2]

Notes