The Moscow Experiment for Testing Innovative Medical Services opens the way to advanced technologies.
The technologies of AI are actively developing and offered solutions for many areas of economics and social life. The special need for digital assistants was manifested during the coronavirus pandemic in the medical industry. Many developers guide their efforts to create projects for the use of clinics in practical work. But from theoretical studies, new ideas to the "binding of them on the ground", that is, for practical use, the doctors have always had a large temporary distance. For the successful development of digital medicine, it was necessary to reduce this period, to develop good opportunities to test their proposals with practitioners, on the basis of real data on patients, in the face of the specific medical institution.
Pioneer in providing such opportunities to introduce new technologies in the practice of medicine was Moscow.
The Moscow health care system provides an experiment on the use of computer vision technologies in radiology. The initiators and organizers of the experiment were the complex of the social development of the capital, IT Department and the Department of Health. Medical IT services based on artificial intelligence (II) helped radiologists in the analysis of pictures and diagnoses.
The AI algorithms were integrated into a single medical information and analytical system (EMIAS). They analyzed the following studies:
- X-ray diagnostics;
- Computer tomography;
- Fluorography
- Mammography.
The machine handled the snapshots very quickly, literally in a few minutes I gave conclusion and put preliminary diagnoses. It greatly facilitated the work of the doctors and unloaded them from the routine work. During the Pandemic period of COVID-19, the volume of radiologists has grown several times. And II's help turned out to be like the way.
Sergey Morozov, director of the Center for Diagnostics and Telemedicine, notes that the experiment was large-scale. More than 20 companies took part in it. In total, more than one and a half million medical pictures and pictures were processed and analyzed. 1.5 million studies. During the experiment, the process of automatic medical data processing was worked out.
For the metropolitan experiment, developers have created 38 services based on AI for analysis of 10 research species. Computer view services suggested Care Mentor AI (Carentoreiai LLC). The two developments are already used in the practice of the work of Moscow hospitals.
- Radioscreening of the pathologies of the chest organs makes it possible to diagnose pneumonia, cancer, tuberculosis. These are the most dangerous social diseases that are important to see in the initial stages for subsequent effective treatment. The service analyzes the pictures obtained from the X-ray apparatus of the clinic. Then localizes the pathological plot of the organ, indicates the exact location and size of pathology. These results are returned to a single information system. Processing occurs in just 8 seconds. The car sees even small changes in the body, which human eye is not capable of fixing. It distinguishes many grades of gray. The machine makes a color mask on the image, which are a tip for a radiologist. A specialist sees a pathological site and conducts further diagnosis.
- The service "CT COVID-19" helps to see early signs of infection in the pictures of CT. It determines the location, the size of pathology and the percentage of lung damage, including according to the CT-0 CT-4 classification approved by the Ministry of Health of Russia. The developer company was awarded the Moscow Mayor Prize "Novator Moscow" in the nomination "Artificial Intelligence and IT Technologies". Service for 2-3 minutes analyzes and issues the results of the image processing. It also imposes color masks on images on each slice. The doctor studies the tips of the car and takes the final decision on the diagnosis.
Ilya Plisci, General Director of Carentoreiai LLC, reminds that the AI is not a replacement for a doctor. He becomes an assistant specialist, focuses the attention of a person on the field of pathological damage to the body. The final decision on the diagnosis and the choice of treatment techniques remains behind the doctor. When working on the service, the company conducted extended consultations with experts of the Research Institute of Moscow and St. Petersburg. To avoid mistakes in identifying pathologies, the markup of each study for neural network training is carried out at least 3 experienced specialists who have experience more than ten years in a clinic and hospital.
In addition to the developments of Karentoreiai, the COVID-MULTIVOX service developed by specialists from Gammamed-Soft was involved in the Moscow experiment. With its help, Moscow radiologists on CT images analyze the pathology of the lungs at COVID-19.
This technology allows you to determine the share of healthy lung tissue. It gives not only volumes, location of pathological changes of the organ. The machine can estimate the level of tissue damage, highlighting the following gradations: fibrous fabric, dense matte glass, matte glass. Such detail allows you to assign more efficient treatment for each individual case.
The developers of the Gammamed-Software amounted to a service program in collaboration with the doctors of the city clinical hospital number 52, scientists of nuclear physics named after D.V. Skobeelsyna. The work took place in a limited time, and doctors instead of recreation advised developers in breaks between their work in the Red Zone of the Covered Office.
At first, the base of CT research was created. On the basis of it, the neural network was trained qualitatively and differentiated to determine the volume of lung tissue disease. The result of machine processing was the chart built to her. They show doctors the dynamics of changes in the patient's health. Based on this data, the specialist corrected or changes the prescribed treatment.
AI Intellect provides the doctor for the following data:
- The remaining volume of healthy lung fabric
- Percentage of body pathology
- The degree of defeat of each pathological section of the lung
- Dynamics of changes in the volume of healthy and damaged areas of the organ
- The effectiveness of the treatment technique
In addition, the machine makes recommendations to a doctor about whether it is necessary to translate the patient to intensive care whether it is required to connect it to the IVL apparatus.
Service measurements have high accuracy. The absolute volume of healthy and affected tissues is indicated in cubic centimeters. The relative size of pathology is measured as a percentage in relation to the entire volume of the lungs. Having obtained exact indicators, the severity of the disease is accomplished.
The final conclusions about the disease, the program is based on the CT-0 CT-4 classification. The machine provides a report with a description of quantitative values and brief comments. All information is stored in the database, ensuring the continuity of the patient's treatment.
Service testing took place at N.V. ambulance Sklifosovsky and in the medical scientific and educational center of Moscow State University named after M.V. Lomonosov, as well as other clinics of the capital. The car analyzed more than 120 thousand CT studies. Test test results have shown that the service can be used in the federal regional system of telemedicine consultations.
Andrei Gavrilov, the supervisor of the development of LLC Gammamed-Soft, is particularly emphasized by the efficiency of the service. The standard snapshot processing time is 10 minutes. AI analyzes the image and provides a report to a doctor 2 times faster - for 5-6.5 minutes. Instead of two studies per hour with the help of the service, 6 studies can be held. This gives prompt and reliable support for doctors, loaded by large volumes of work during the coronavirus pandemic and epidemics of other dangerous and seasonal infections. During the Moscow experiment, the authors of the program were able to improve it, make a more useful doctor for practical work. The Moscow government has provided financial support to developers as a gratitude for their quality product.
Experiments on the use of technologies II are planning to carry out other regions of the country. For example, on the basis of the first Russian Institute of Artificial Intelligence at the University "Innopolis" in the Republic of Tatarstan, testing of the Airadiology service began.
This program analyzes the radiography of the chest organs, recognizes and measures the dimensions and depths of the lung damage. The machine gives a doctor a report consisting of two files:
- The Structure Report Research Protocol defines the likelihood of pathological changes in the body.
- Primary X-ray with the imposition of a thermal map on it, showing areas of pathological changes of the organ.
The machine handles snapshots in just 30 seconds. And in the report gives clear images.
Ramil Kuleyev, director of the Institute of the University of Innopolis, emphasizes the presence of huge statistics, indicating the level of load of radiologists. In Russia, over the course of the year, approximately eighty millions of radiological studies of the chest bodies are carried out, every day more than 220 thousand pictures are made in our country. All this volume should be treated doctors. If technology will be used in their work, the results of the results will occur much faster and more accurate. The liberation of the doctor from routine work will allow to conduct a more in-depth study of the complex cases, the operational implementation of urgent studies.
Ramil Kuleyev appreciated Moscow Health Infosystem as one of the best in Europe. It is the leader in Russia and allows you to work out new services, programs and technologies for subsequent introduction of them in the practice of the work of metropolitan and regional medical centers. For this there are all the necessary regulations and techniques. They were developed by the Center for Diagnostics and Telemedicine and the Information and Analytical Center in the Health of Moscow. Now developers of new services can test their projects on a qualitatively built IT infrastructure.
To participate in the Moscow experiment, the entrance is open to all health services based on computer vision technologies. Its applications for the participation of IT companies can be submitted to the center of diagnosis and telemedicine. Its site hosts a list of service requirements and a list of necessary support documents.
Those developments that meet the stated requirements will participate in test and tests on the features of the functionality and the accuracy of algorithms. In this case, each service must be integrated with the Eris. Successfully passed testing services receive grants from the Government of Moscow. Developments that have successfully passed testing are included in the industrial circuit. They become available to medical professionals.
Inna Frost, Director of the Development of Karentoreiai, the usefulness of the company's testing of the company within the Moscow Experiment is highly assessed. Examination helped assess the technical maturity, functionality and practical usefulness of development. High ratings were given a pulse for new searches and improving the products developed.
Source mos.ru.