Abstract:
Introduction. Artificial intelligence (AI) is a new technical discipline that
uses computer technology to research and develop the
theory, method, technique and application system for
simulating, and extending human intelligence. According
to "pubmed.ncbi.nlm.nih.gov" data, only in the last 5
years, around 100,000 researches in the field of AI in
medicine have been posted. This impressive number only
demonstrates the importance and topicality of the topic,
and denotes the speed with which AI enters medicine.
Purpose. This research aims to better understand this
technology and how it transforms medicine, what is
the role of artificial intelligence-based systems in
performing different medical activity in
specializations and what are the results nowadays.
One of the main goal was to investigate the role of
artificial intelligence-based systems in performing
medical work in specialties including radiology,
oncology, cardiology, pediatry etc.
Material and methods. The main resources for "searching articles" were pubmed.ncbi.nlm.nih.gov and
cyberleninka.ru, so the sources that were analyzed in the research provide as
objective a picture as possible of the role of AI that has undergone criticism and
analysis by specialists from many corners of the world. The dynamic development
of scientific progress in solving the topic of AI in medicine and new discoveries in
this field play a dominant role in the work.
Results. A tool, called the Molecular Prognostic Score (mPS), has recently been developed
that is able to accurately predict the prognosis of breast cancer patients and
comprehensively identified 184 genes related to breast cancer prognosis without
any biological information. Unlike previous tools, it can be applied even to patients
with estrogen receptor-negative breast cancer. In addition, the score provides
useful information to avoid overtreatment.(Figure1). In anesthesiology, artificial
intelligence and spectral analysis techniques are used to more directly analyze
electroencephalographic signals in order to estimate the depth of anesthesia. A
group of researchers reached such results as: the accuracy in using
electroencephalography features was 88.4%, while the accuracy of the BIS index
was 84.2%. Another study in this field is directly focused on the rapid diagnosis of
myocardial infarction (MI) using electrocardiography (ECG). A total of 412,461 ECGs
were used to develop a variational autoencoder (VAE). The performance of the
neural network, measured as the area under the receiver operating characteristic
(ROC) curve, was 0.887 (0.845-0.922).The way CNN works is showed in fig 2-3.
Conclusions. Their ability to learn from historical examples, analyze non-linear data, handle
imprecise information and generalize by allowing the model to be applied to
independent data has made them a very attractive analytical tool in the field of
medicine. The author is of the opinion that the most important results still await us
in the future.
