Cartographic presentation of information is especially in demand and visual for covering global and regional events, as well as assessing the location of stationary and non-stationary objects in the study area. Geoinformation systems (GIS) in epidemiological surveillance of parasitic diseases provide collection, storage, processing, access to information, display and dissemination of spatially coordinated data that can be used to solve scientific and applied problems: analysis, monitoring, assessment, forecasting the spread of parasites and maintaining sanitary epidemiological well-being of the population.

 Aim. To study the main stages of the history of GIS development in the world considering the possibilities of GIS application in healthcare.

 Materials and methods. Scientific publications of domestic and foreign authors on the problem from 2000 to 2023 were used. The contextual search included the keywords “geographic information”, “GIS” (in Russian and English). Search tools included the scientific electronic library eLIBRARY.RU and search engines Google and Yandex. 31 sources were found, of which 17 were domestic. Descriptive, analytical methods, the method of retrospective historical analysis, and content analysis were used.

 Results. The article presents the history of GIS development in the world, consisting of four stages — the pioneer period, the period of government initiatives, the period of commercial development, the user period. The areas and examples of the use of GIS in epidemiology and in the planning of healthcare systems are given.

 Conclusion. Improvement of GIS technologies and software in healthcare makes it easier to analyze data on stationary and non-stationary objects, as well as it improves quality and validity of analytical information.

The application of machine learning in healthcare, as one of the more general artificial intelligence technology, has shown enormous potential for improving diagnostic and treatment outcomes for various conditions. However, success of AI-based software largely depends on the availability of high-quality medical datasets and the infrastructure built to streamline its management. Creating relevant, representative and accurately labeled datasets is a complex and expensive task that requires diverse expertise and a robust roadmap for dataset building in radiology.

 This paper presents a dataset creation methodology in radiology that establishes principles and protocols to ensure a standardized approach to dataset building, secures a convenient infrastructure for data management, and provides a framework to automate the creation of high-quality datasets.

 With our experience in implementing the methodology presented in this paper for routine diagnostic imaging, we demonstrate typical errors that arise when preparing radiology datasets and offer ways to avoid them.

The Unified national medical nomenclature (UNMN) has been under development since 2022 with using the Unified Medical Language System (UMLS) Metathesaurus and other sources. UNMN is a terminological system based on ontological approach and potentially applicable in Russian language medical text annotating. Currently, terms from different clinical branches are being added to UNMN utilizing both automatized and expert ways. Often in medicine abbreviations allow expressing the meaning of the concepts in a rapid way. However, their recognition in unstructured text is not trivial issue. The development of software for automated abbreviations recognition from research articles could enrich UNMN and accelerate clinical decision support systems development.

 The aim of this study was to create the automated algorithm for UNMN terms abbreviations recognition from text of Russian language research articles.

 Methods. Validation and testing dataset included unstructured abstracts of Russian language research articles aggregated from eLIBRARY. Fulltext wordings of extracted abbreviations have been corrected with bilingual (RussianEnglish and EnglishRussian) translation.

 Results. Final version of the algorithm based on semantic rules demonstrated ~93% sensitivity and ~99% specificity in abbreviations and their fulltext wordings extraction. Large percentage (~87%) of terms has been successfully corrected and presented in the initial form after bilingual translation. Half (~49%) of abbreviations has been mapped with 100% accuracy to UNMN terms. Processing of 168 000 abstracts using the developed algorithm lead to creation of the Unified medical abbreviations thesaurus with UNMN terms (exceeding 6600 unique entries).

Implementation of Artificial Intelligence (AI) is considered one of the most promising directions in the digital transformation of healthcare. Such systems can improve the quality of therapeutic and diagnostic processes and the efficiency of planning and managing the healthcare industry. However, the potential of AI to enhance public health indicators and improve the functioning quality of the healthcare system is inextricably linked to ethical issues arising from the specific aspects of their creation and implementation, as well as their direct impact on the life and health of communities, individual patients, and medical personnel.

 It is necessary to form and increase trust from the medical community, patients, regulatory and supervisory bodies, and other interested parties in order to implement AI in healthcare. For this purpose, it is advisable for AI developers and other involved parties to follow a set of unified ethical principles. Based on leading work in the field of AI ethics, 19 principles for the creation of AI in healthcare were developed. They were directed to public discussions and adjusted in consideration of the feedback. Adherence to the published principles by participants in the processes of creation, testing, validation, market launch, and postmarketing support can significantly increase trust in AI and contribute to the successful implementation of ethical AI systems in healthcare.

Aim. To demonstrate the special aspects of dataset creation for neuroimaging using the example of preparing a dataset with computed tomographic images of the brain with and without signs of intracranial hemorrhage.

 Methods. The creation of the dataset is based on the methodology developed by the Scientific and Practical Clinical Center for Diagnostics and Telemedicine (regulations for preparing the dataset), which is carried out in 4 stages: planning (selection of the necessary keywords for the initial selection of studies, determination of inclusion and exclusion criteria, source of medical information), selection (initial downloading of the text information - a brief patient history and description protocols from the Unified Radiological Information Service of the city of Moscow for 2020, anonymization of the received data, keywords analysis), labeling and verification (filling out the accompanying table with clinical and technical data, study selection by two radiologists and an expert verification by a neuroradiologist) and publication (publication of the dataset online, state registration).

 Results. In the process of creating a dataset, the special aspects, defined by the neuroradiology background, were noted and formulated, which should be taken into the account when executing the primary training, testing and additional training of artificial intelligence services for diagnosing brain diseases: the use of specific terms, the use of images with the least amount of noise and the highest contrast, as well as the use of ratios of subtypes of the target pathology corresponding to its ratio in the population. A dataset with computed tomography images containing signs of intracranial hemorrhage was prepared. The final version of the dataset included anonymized studies of 209 patients (109 with the pathology, 100 without the pathology): DICOM images, an accompanying text table with clinical features (gender, age, type(s) and number of hemorrhages, presence/absence of concomitant pathology) and technical parameters (slice thickness and reconstruction slice thickness).

 Conclusion. The special aspects of preparing datasets for training and testing neuroradiological artificial intelligence services were demonstrated.

Aim: To evaluate the experience of using software based on artificial intelligence technologies as part of the Moscow experiment on the use of innovative technologies in the field of computer vision for the analysis of medical images.

 Material and methods: A retrospective study was conducted. The work includes the conclusion outputs of 3 AI services on 822 thousand fluorographic studies for the period from 05.01.2022 to 29.12.2022. Pathology was present in 28,341 studies (3.4%). The assessment was carried out using quality metrics of binary classifiers and statistical methods. The metrics were assessed depending on the AI services threshold.

 Results: There was a pronounced imbalance between studies with norm and pathology. High values of imbalance-sensitive metrics and low values of imbalance-insensitive metrics were obtained, which was associated with a high rate of false positive and false negative results. By changing the threshold, it was possible to reduce the number of false negative results. For example, one of the AI services, with a threshold of 0.05, correctly identified 46.8% of studies with the norm, and with no false negative results.

 Conclusions: The number of false negative results for the studied versions of AI services is an obstacle to their autonomous implementation into routine practice, which requires their improvement. By optimizing the service threshold, it is possible to achieve error-free identification of 46.8% of studies with the norm, but due to the closed nature of AI services, this method is limited. Further options for optimizing services require additional study.

Aim. To study the possibilities of self-remote monitoring of blood pressure and heart rate in high-risk pregnant women.

 Material and methods. Women from 18 to 45 years old with hypertension (chronic and gestational), a history of preeclampsia, a high level of preeclampsia risk (above 1:100), body mass index (BMI) > 30 kg/m² were included in the study to test selfmonitoring of blood pressure using the MedSenger platform. The patients used a home tonometer to measure blood pressure and heart rate daily in the morning and evening and entered data into the personal account of the MedSenger application for 1 month. The doctor monitored the parameters of pregnant women, received notifications in case of deviations and decided on further tactics.

 Results. We planned to include 59 patients in the study, but 25 women (42.4%) refused to participate. 5 Out of 34 pregnant women (average age 30.1±2.3 years) did not activate the program and were excluded. Finally, 29 women took part in the study (49.1% of the initial number). The main reasons for refusing monitoring: no tonometer (30%), no trust in technology (23.3%), no email adress (16.7%), no free time (13.3%), no reason (16.7%).

The average systolic blood pressure in the study was 115±3.7 mmHg, the average diastolic blood pressure was 73±2.8 mmHg, the average heart rate was 84±3.5 per minute. 17 (58.6%) out of the 29 pregnant women completed the monitoring program on time. 12 women (41.8%) (average age 33.1±3.1 years) completed monitoring ahead of schedule.

 Conclusions. Self-remote monitoring of high-risk pregnant women is a feasible technology for remote monitoring.