Dec 28, 2020
0 0

In Moscow, a neural network was trained to diagnose lung damage by a blood test

The neural network “CT calculator”, created by scientists from Moscow State University. Lomonosov, determines the state of the patient’s lungs by analyzing blood, saturation (oxygen level in the blood) and the general condition of the patient, said Moscow Mayor Sergei Sobyanin.

By analyzing the data, artificial intelligence determines the state of the lungs in accordance with the possible indicators of computed tomography: CT 0-1, CT 2 and CT 3-4. Based on the results of the assessment, the doctor decides whether the patient needs computed tomography and whether it is possible to refuse this type of study if the neural network showed a mild course of COVID-19. If a serious condition is diagnosed, the doctor may decide to immediately hospitalize the patient for detailed diagnosis and intensive treatment.

As noted by Sobyanin, the neural network was trained based on data from more than half a million patients who have been treated in Moscow for COVID-19 since March. As new data becomes available, the digital service will be improved.

The CT-calculator is already built into the Moscow UMIAS system. Access to it is open to doctors not only in Moscow, but also in other regions, ”the head of the city noted.

During the pandemic, Moscow began to actively introduce digital technologies into the healthcare system. So, for 10 months now, Moscow hospitals have been using a special program for analyzing CT images of the lungs. In October, a medical decision support system (MDSS) developed by Sberbank specialists was introduced in Moscow adult clinics. Artificial intelligence provides “tips” for doctors to prescribe treatment, thereby making their work easier. For further training of the system, the developers will have to combine large amounts of data – clinical protocols and prescriptions, which were worked on by about 200 leading doctors in Moscow.

Article Categories:

Leave a Reply