St. Petersburg State University: Neural network for the treatment of patients with heart disease

The main articles are:

• Big Data
• Artificial Intelligence in Medicine
• Neural networks (neural networks)

2022: Creating a neural network

The doctor Clinics of high medical technologies named after N.I. Pirogov St.Petersburg State University Gleb Kim created, neuronet which, on the basis databases encrypted of patients' medical histories, can predict the development of cardiovascular diseases, as well as give recommendations for their treatment. St. Petersburg State University announced this on July 13, 2022.

Using the system "Second opinion. Neural networks and aortic pathology in cardiac surgery "specialists from all regions of the country will be able to receive expert assistance in the management of patients with a life-threatening condition - aneurysm or aortic dissection. It will be enough for the doctor to provide a certain amount of data from the patient's medical history: clinical data, his anthropometry, results of computed tomography of the aorta and echocardiography, after which the neural network will give its forecasts and recommendations.

The conclusions of the expert system are formed as a result of its training on the basis of encrypted data from patient records. The collection of big data - large amounts of data necessary for effective training of neural networks, is planned to be carried out in cooperation with cardiac surgeons and cardiologists from all regions of Russia.

To implement the project, a team was created, which included cardiac surgeons of the N. I. Pirogov Clinic of High Medical Technologies of St. Petersburg State University Gleb Kim and Viktor Pyagai and programmers of St. Petersburg State University. Work on the creation of the Second Opinion medical neural network has been carried out with the support of the University for about a year and a half. As of July 2022, nine models of neural networks have been created and tested. The main task of developing development is to accumulate data - this is what allows neural networks to learn and form accurate recommendations.