All articles by Arzamasov K. M.
-
2023 № 7 The use of artificial intelligence technologies as a way to ensure the quality of chest radiography.
When performing radiographic studies, errors may occur that reduce the diagnostic value of the radiographs and complicate their interpretation by radiologists and diagnostic software based on artificial intelligence technology. The creation of automated quality assessment systems will optimize this process, especially in conditions of increased workload of medical personnel.
Purpose: development of an automated quality control tool for chest radiographs, which allows for quality control of the patient’s positioning and the correctness of filling in meta-information about the study.
Material and methods. To train and test automated quality control models, were used 61505 chest radiographs, obtained from open datasets and the Unified Radiological Information Service of the Unified Medical Information Analysis System of the City of Moscow. To create models we used transfer training of deep neural network architectures VGG19 and ResNet152V2.
Results. 7 models were created: a model for determining the anatomical area of study, a model for determining projection, a model for determining photometric interpretation, models for determining incomplete visualization of the anatomical area on the frontal and lateral projections of the chest radiographs, a model for determining rotation on the lateral projection of the chest radiographs. All created models have diagnostic accuracy metrics above 95%, which allows them to be used in clinical practice. Based on the developed models, a web-based quality control tool of the chest radiographs was created, which allows analyzing the quality of X-ray datasets.
Conclusion. The active use of this quality control tool will optimize the process of assessing the quality of diagnostic studies and facilitate the processes of classification of studies and the formation of datasets. Also, this tool can be used to support the decision-making of an X-ray technician and assess the quality of the study before sending the study for processing to artificial intelligence-based services. -
2023 № 4 Medical datasets for machine learning: fundamental principles of standartization and systematization.
Backgraund: Active implementation of artificial intelligence technologies in the healthcare in recent years promotes increasing amount of medical data for the development of machine learning models, including radiology and instrumental diagnostics data. To solve various problems of digital medical technologies, new datasets are being created through machine learning algorithms, therefore, the problems of their systematization and standardization, storage, access, rational and safe use become actual.
A i m : development of an approach to systematization and standardization of information about datasets to represent, store, apply and optimize the use of datasets and ensure the safety and transparency of the development and testing of medical devices using artificial intelligence.
M a t e r i a l s a n d m e t h o d s : analysis of own and international experience in the creation and use of medical datasets, medical reference books searching and analysis, registry structure development and justification, scientific publications search with the keywords “datasets”, “registry of medical data”, placed in the databases of the RSCI, Scopus, Web of Science.
R e s u l t s . The register of medical instrumental diagnostics datasets structure has been developed in accordance with stages of datasets lifecycle: 7 parameters at the initiation stage, 8 – at the planning stage, 70 – dataset card, 1 – version change, 14 – at the use stage, total – 100 parameters. We propose datasets classification according to the purpose of their creation, a classification of data verification methods, as well as the principles of forming names for standardization and datasets presentation clarity. In addition, the main features of the organization of maintaining this registry are highlighted: management, data quality, confidentiality and security.
C o n c l u s i o n s . For the first time, an original technology of medical datasets for instrumental diagnostics structuring and systematization is proposed. It is based on the developed terminology and principles of information classification. This makes it possible to standardize the structure of information about datasets for machine learning, and ensures the storage centralization. It also allows to get quick access to all information about the dataset, and ensure transparency, reliability and reproducibility of artificial intelligence developments. Creating a registry makes it possible to quickly form visual data libraries. This allows a wide range of researchers, developers and companies to choose data sets for their tasks. This approach ensures their widespread use, resource optimization and contributes to the rapid
development and implementation of artificial intelligence. -
2023 № 9 Study of emphysematous changes in the population of Moscow using automated evaluation of radiological examinations.
Background. Emphysema, commonly seen in patients with chronic obstructive pulmonary disease (COPD), worsens the course of chronic cardiovascular and endocrine diseases and is also associated with an increased risk of lung cancer. Although the evaluation of COPD incidence is applied systematically, the prevalence of emphysema is often not known. One of the ways to offset that is automated analysis of chest CT scans using artificial intelligence technologies.
Goal. To study the prevalence of emphysema in the population of Moscow using automated analysis of radiological examinations.
Methods. The results of the chest CT scan of 116,216 patients were analyzed. All studies were performed between October 2022 and June 2023 in Moscow medical facilities. The Emphysema-IRA AI service (Intelligent Radiology Assistance Laboratories (AIRA Labs) LLC) used an automated mode to determine the presence of emphysematous changes in the lungs (binary classification – yes/no) and the percentage of emphysematous lesions in both lungs and each lung separately.
Results. The prevalence of pulmonary emphysema among the Moscow population was 0.614 per 1,000 people; the prevalence of clinically significant emphysema was 0.173 per 1,000 people. The majority of individuals presented with either pulmonary or clinically significant emphysema in CT belong to the elderly group (47.0% and 55.0%, respectively); the proportion of young people is also significant (9.0% and 5.0%). Men of all age groups have a significantly higher chance to get diagnosed with emphysema which suggests a higher incidence compared to the female population (Chi-square = 1000.0; p<0.001). Regardless of gender, a 5-year increase in age elevates the likelihood of both emphysema and clinically significant emphysema by 1.1 times.
Conclusions. Automated detection of signs of pulmonary emphysema on CT allows for a quick, population-wide, and objective assessment of the COPD prevalence. Thanks to the development of AI-based medical software, it has become possible to develop and implement ground-breaking digital technologies for healthcare management and public health studies. -
2023 № 4 Medical datasets for machine learning: fundamental principles of standartization and systematization.
Backgraund: Active implementation of artificial intelligence technologies in the healthcare in recent years promotes increasing amount of medical data for the development of machine learning models, including radiology and instrumental diagnostics data. To solve various problems of digital medical technologies, new datasets are being created through machine learning algorithms, therefore, the problems of their systematization and standardization, storage, access, rational and safe use become actual.
A i m : development of an approach to systematization and standardization of information about datasets to represent, store, apply and optimize the use of datasets and ensure the safety and transparency of the development and testing of medical devices using artificial intelligence.
M a t e r i a l s a n d m e t h o d s : analysis of own and international experience in the creation and use of medical datasets, medical reference books searching and analysis, registry structure development and justification, scientific publications search with the keywords “datasets”, “registry of medical data”, placed in the databases of the RSCI, Scopus, Web of Science.
R e s u l t s . The register of medical instrumental diagnostics datasets structure has been developed in accordance with stages of datasets lifecycle: 7 parameters at the initiation stage, 8 – at the planning stage, 70 – dataset card, 1 – version change, 14 – at the use stage, total – 100 parameters. We propose datasets classification according to the purpose of their creation, a classification of data verification methods, as well as the principles of forming names for standardization and datasets presentation clarity. In addition, the main features of the organization of maintaining this registry are highlighted: management, data quality, confidentiality and security.
C o n c l u s i o n s . For the first time, an original technology of medical datasets for instrumental diagnostics structuring and systematization is proposed. It is based on the developed terminology and principles of information classification. This makes it possible to standardize the structure of information about datasets for machine learning, and ensures the storage centralization. It also allows to get quick access to all information about the dataset, and ensure transparency, reliability and reproducibility of artificial intelligence developments. Creating a registry makes it possible to quickly form visual data libraries. This allows a wide range of researchers, developers and companies to choose data sets for their tasks. This approach ensures their widespread use, resource optimization and contributes to the rapid
development and implementation of artificial intelligence.