Sovremennye tekhnologii v meditsine最新文献

{"title":"Neural Network Technologies for Age Estimation in Children from Orthopantomograms (a Pilot Study).","authors":"M P Poletaeva, Yu V Vasilevsky, D K Valetov, N Angelakopoulos, G V Zolotenkova","doi":"10.17691/stm2026.18.2.02","DOIUrl":"10.17691/stm2026.18.2.02","url":null,"abstract":"<p><p><b>The aim of the study</b> was to investigate the potential of using artificial intelligence technologies for age estimation in children from dental radiographs.</p><p><strong>Materials and methods: </strong>A retrospective study was conducted, analyzing orthopantomograms of 322 children (173 female, 149 male) aged 4-16 years. Fourteen permanent mandibular teeth were annotated on each radiograph. Neural network training was performed by splitting the data into training and test sets at a ratio of 80:20; 5-fold cross-validation was used. Age estimation was approached as a regression task. The neural network training and validation were conducted in Python using the PyTorch library. The accuracy of age prediction was assessed using the coefficient of determination (R²), mean squared error (MSE), and mean absolute error (MAE).</p><p><strong>Results: </strong>The study showed that the developed machine learning model was highly accurate in age estimation in children. The mean absolute error across cross-validation was 0.92 years, which was significantly lower than the error associated with traditional manual methods.</p>","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 2","pages":"15-21"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13178086/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147966838","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A New Quantitative Method for Assessing the Ultrastructure of Cardiomyocyte Mitochondria of the Right Atrial Appendage.","authors":"E A Kuzheleva, A A Garganeeva, O V Tukish, E E Syromyatnikova, B B Dorzhieva, M Yu Kondratiev, S L Andreev","doi":"10.17691/stm2026.18.2.05","DOIUrl":"10.17691/stm2026.18.2.05","url":null,"abstract":"<p><p><b>The aim of this study</b> was to develop a quantitative method for assessing the ultrastructure of cardiomyocyte mitochondria in the right atrial appendage of patients with chronic heart failure (CHF).</p><p><strong>Materials and methods: </strong>A single-center prospective study included 39 patients aged 67 [58; 71] years who underwent coronary artery bypass grafting. The inclusion criteria were as follows: heart failure with a left ventricular ejection fraction of <50%, atherosclerotic plaques in two or more large coronary arteries of 70% or greater, and the cardiac team's decision to perform coronary artery bypass grafting. To perform electron microscopy, biopsy samples were collected from the right atrial appendage during coronary artery bypass grafting. Two calculated parameters were used for the analysis: \"total area of interfibrillar mitochondria\" and \"ratio of outer to inner membrane length\". Based on these, the total index was calculated by the formula: the \"total area of interfibrillar mitochondria\" / the \"ratio of outer to inner membrane length\".</p><p><strong>Results: </strong>The median values of the calculated parameters characterizing the ultrastructure of cardiomyocyte mitochondria in the right atrial appendage were as follows: 43.7 [35.9; 54.3]% for the \"total area of interfibrillar mitochondria\", 31 [25; 37]% for the \"ratio of outer to inner membrane length\", and 1.4 [0.95; 2.00] for the total index accounting for both ultrastructural characteristics of mitochondria: the \"total area of interfibrillar mitochondria\" / the \"ratio of outer to inner membrane length\".To assess the clinical significance of the proposed total index, associations between the obtained values and atrial fibrillation (AF) were studied. Mitochondrial ultrastructural changes were more pronounced in patients with both AF and CHF. However, some parameters had no statistically significant differences: the total area of interfibrillar mitochondria was lower in patients with AF (42% vs 49%, p=0.224), and the ratio of outer to inner membrane length was higher in patients with AF (35% vs 31%, p=0.125). The total index was significantly lower in patients with AF (0.96 vs 1.75, p=0.021). ROC analysis was performed to identify a connection between total index and AF, and AUC was 0.773 (p=0.021).</p><p><strong>Conclusion: </strong>The proposed total index for analyzing cardiomyocyte micrographs, calculated as the \"total area of interfibrillar mitochondria\" / the \"ratio of outer to inner membrane length\", provides a comprehensive characterization of mitochondria that accounts not only for their number and size but also for their internal structure.</p>","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 2","pages":"54-60"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13178087/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147966801","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Immunohistochemical Detection of Microglia Using Iba-1 as a Marker in Health and Disease.","authors":"V V Guselnikova, O V Kirik, D A Sufieva, V A Razenkova, A A Beketova, D E Korzhevskii","doi":"10.17691/stm2026.18.1.03","DOIUrl":"10.17691/stm2026.18.1.03","url":null,"abstract":"&lt;p&gt;&lt;p&gt;&lt;b&gt;The aim of the study&lt;/b&gt; was to evaluate the possibilities and limitations of using the immunohistochemical detection of Iba-1 protein for morphofunctional analysis of microglia under different conditions.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Materials and methods: &lt;/strong&gt;In the study we used brain samples from Wistar rats at different ages: 7 days of postnatal development (n=18), 14 days of postnatal development (n=18), 4-6 months (n=22); the brain samples from 3-6-month-old SHR (spontaneously hypertensive rats) male rats (n=4); human cerebral cortex samples (n=10). Rabbit polyclonal antibodies against calcium-binding protein Iba-1 (Biocare Medical, USA) were used to detect microglia immunohistochemically followed by light and confocal laser microscopy. A previously developed original technique was applied to simultaneously detect microglial cells and amyloid plaques.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Iba-1 protein in microglial cells was shown to be present in grey and white matter in all cerebral regions under study. Microglial cells in different regions were noted to be characterized by pronounced structural and functional features when stained for Iba-1. In addition to microglial cells, in the rat brain there were found other Iba-1-immunopositive macrophages, which were tissue cerebral macrophages differing from microgliocytes by specific morphology and localization. Iba-1 protein was demonstrated to be present in microglial cells in all investigated stages of postnatal development that makes Iba-1 an appropriate marker for comparative ontogenetical studies of microglial cells. The localization of Iba-1 in the bodies and processes of microgliocytes enables to more clearly identify the complex ramified cell morphology and make three-dimensional reconstructions. In the brain of spontaneously hypertensive rats (SHR line rats) the Iba-1-immunopositive microglia were shown to have a number of structural and functional features indicating the moderate activation of microglia. 48 h after ischemic injury, an anomalously great number of large Iba-1-immunopositive cells with amoeboid morphology were detected in the gray matter of the striatum in an ipsilateral hemisphere near the injured area in the rat brain. Microglial cells in human cerebral cortex were noted to be localized in the majority of the detected amyloid plaques and characterized by amoeboid morphology indicating their severe activation.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusion: &lt;/strong&gt;The presented findings suggest Iba-1 protein to be a reliable and universal microglial marker. The immunohistochemical detection of the protein enables to identify and analyze qualitatively and quantitatively microglial cells in different brain regions in human and laboratory animals under normal and pathological conditions. The limitations of using Iba-1 as a marker of microglia include the inability to determine the microglial activation vector and the difficulty in distinguishing between microglia and infiltrating macrophages of th","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 1","pages":"34-42"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13003591/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147501518","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TEM-based Study of the Phenotype of Astrocytes Differentiated from Induced Pluripotent Stem Cells from a Healthy Donor and a Patient with Parkinson's Disease.","authors":"K A Kutukova, M V Ivanov, E V Novosadova, A V Brydun, E L Arsenyeva, L V Novosadova, I V Kokorev, I A Grivennikov, V S Sukhorukov, S N Illarioshkin","doi":"10.17691/stm2026.18.1.01","DOIUrl":"10.17691/stm2026.18.1.01","url":null,"abstract":"&lt;p&gt;&lt;p&gt;&lt;b&gt;The aim of this study&lt;/b&gt; was to study the role of transmission electron microscopy (TEM) in assessment of the phenotype of astrocytes obtained with the directed differentiation technique from induced pluripotent stem cells (iPSCs) from a healthy donor and from a patient with a hereditary form of Parkinson's disease (PD).&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Materials and methods: &lt;/strong&gt;Monolayer astrocyte cultures differentiated from iPSCs from a healthy donor and a PD patient having the G2019S mutation in the &lt;i&gt;LRRK2&lt;/i&gt; gene were used in the study. The obtained glial cultures were characterized using real-time PCR and immunocytochemical staining for glia-specific genes and proteins. TEM was used to examine astrocyte ultrastructure.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;PCR analysis and immunocytochemical staining demonstrated that cell lines received from a healthy donor and a PD patient expressed the required pattern of glia-specific genes and synthesized astrocyte-specific proteins. However, some glia-specific genes were expressed at reduced levels by mutant cells. One of the most typical ultrastructural features of astrocytes received from iPSCs from a PD patient was destructive changes in mitochondria, including mitochondrial clearing, swelling, and cristae destruction. In many cells, mitochondria were completely absent after a long culturing. Another characteristic feature of cells with a mutation in the &lt;i&gt;LRRK2&lt;/i&gt; gene was the accumulation of vacuoles with contents of varied electron density. Distinct changes in the ultrastructure of nuclei, protein-synthesizing organelles, and cytoskeletal elements were also seen in cultured astrocytes with a PD-associated &lt;i&gt;LRRK2&lt;/i&gt; mutation. Here, the morphometric study did not reveal any differences in the average cell area, nuclear area, cytoplasm area, or nuclear-cytoplasmic ratio between astrocytes of the control line and the PD mutation line.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusion: &lt;/strong&gt;Reprogramming and obtaining of astrocytes from iPSCs received from a donor with a PD-associated mutation in the &lt;i&gt;LRRK2&lt;/i&gt; gene allow to assess the nature and dynamics of pathological morphochemical and ultrastructural changes caused by the mutation during gliogenesis. The use of combined techniques (PCR, immunocytochemistry, TEM) to compare cell cultures differentiated from iPSCs allow to assess, on the one hand, general culture parameters, such as the dynamics of culture differentiation based on changes in the expression level of specific genes and immunocytochemical markers, and on the other hand, morphofunctional changes at the level of individual cells. TEM demonstrates significant potential for studying cell cultures differentiated from iPSCs. This technique is instrumental for phenotyping the resulting cells based on their ultrastructure, assessing the degree of their morphological maturity, and identifying minor ultrastructural changes in cells, both pathological and differentiation-associated. The results of this TEM-based s","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 1","pages":"5-20"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13003593/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147501589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Biocompatibility Assessment of a Hybrid Material Based on Epoxy-Treated Pericardium and Polyvinyl Alcohol Cryogel in a Rat Vascular and Subcutaneous Implantation Model.","authors":"E A Ovcharenko, T V Glushkova, E A Senokosova, K Yu Klyshnikov, T N Akentieva, A A Klyueva, A V Mironov, A E Kostyunin","doi":"10.17691/stm2026.18.3.03","DOIUrl":"10.17691/stm2026.18.3.03","url":null,"abstract":"<p><p><b>The aim of the study</b> was to evaluate the biocompatibility of a hybrid material based on epoxy-treated bovine pericardium and cryostructured polyvinyl alcohol (PVA) after implantation into the rat aortic wall and subcutaneous tissue.</p><p><strong>Materials and methods: </strong>The study evaluated hybrid material patches based on epoxy-treated bovine pericardium and cryostructured PVA; unmodified \"KemPeriplasNeo\" pericardial patches served as controls. The samples were implanted into the abdominal aortic wall and subcutaneously in male Wistar rats. The vascular implantation periods were 5 and 20 days, and the subcutaneous implantation periods were 60 and 120 days. The specimens were excised from the aorta and studied histologically using Russell-Movat staining and immunostaining for neutrophil myeloperoxidase (MPO) and the macrophage marker CD68. The calcium content in subcutaneously implanted samples was assessed by spectrophotometry and alizarin red S staining. The quantitative data were presented as the median, percentiles, minimal and maximal values (Me [25%-75%; min-max]).</p><p><strong>Results: </strong>After implantation into the abdominal aortic wall, the hybrid material samples showed a lower tendency toward thrombotic deposit formation on the surface compared to the pericardium. On day 5 of implantation, the thrombus thickness was 49.1 [34.7-64.6; 27.4-71.6] μm in the experimental group vs 170.3 [158.1-210.3; 124.4-217.0] μm in the controls (p<0.001). By day 20 of implantation, the macrophage density was lower in the peri-implant area of the modified samples compared to the controls: 219 [187-275; 112-362] vs 301 [244-338; 194-433] CD68⁺ cells per field of view (p<0.001), respectively. The neutrophil density at all experimental time points, as well as the macrophage density on day 5 of implantation did not differ significantly between the tested materials (p>0.17). No signs of calcification were detected in either group following the subcutaneous implantation over a period of 60 or 120 days. The control pericardium in the subcutaneous model showed the signs of cell-mediated degradation, whereas the hybrid samples were resistant to immune-cell infiltration and preserved their internal architecture.</p><p><strong>Conclusion: </strong>The modification of epoxy-treated bovine pericardium with cryostructured PVA cryogel improved its biocompatibility and enhanced resistance to biodegradation in a rat model after aortic-wall and subcutaneous implantation. The developed material may be used to create next-generation heart valve bioprostheses with potentially improved resistance to structural valve degeneration.</p>","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 3","pages":"25-36"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13410878/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148623463","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine Learning Application to Predict Bicycle Ergometer Test Results: a Prospective Cohort Study.","authors":"E V Berezina, K A Blinova, O A Dmitrieva, I E Mishina","doi":"10.17691/stm2026.18.2.06","DOIUrl":"10.17691/stm2026.18.2.06","url":null,"abstract":"<p><p><b>The aim of the study</b> was to develop an optimal technique to predict the results of a bicycle ergometry test (BET) based on the parameters recorded during a six-minute walk test (6MWT) using machine learning methods.</p><p><strong>Materials and methods: </strong>The study involved 56 patients who had experienced acute myocardial infarction and were undergoing the second stage of cardiac rehabilitation. The patients underwent a complete examination, including history taking, physical examination, anthropometric assessment, as well as a symptom-limited BET and 6MWT. During the 6MWT, we recorded the following: the distance covered, the heart rate, the blood pressure, the oxygen saturation, Borg rating of perceived exertion, the number of steps taken, and the electrocardiographic data. The algorithms for random forest, gradient boosting, k-nearest neighbors, and multiple linear regression were used to construct the machine learning models. The performance of the models was evaluated based on a determination coefficient, a mean absolute error, a mean square error, and a root mean square error. SHAP analysis was applied to interpret the findings.</p><p><strong>Results: </strong>The gradient boosting model provided the best prediction quality with a high determination coefficient (R² being around 0.99) and low error values for both target metrics: the distance walked in the 6MWT and the metabolic equivalent achieved during the BET. The significance analysis of features revealed the heart rate, age, and the body mass index to have the greatest impact on predicting the 6MWT distance, while for predicting the metabolic equivalent, the distance covered, the number of steps, and the body mass index were the most significant.</p><p><strong>Conclusion: </strong>The developed gradient boosting-based machine learning model demonstrated its high efficiency in predicting the results of the 6MWT-based BET. The suggested method can serve as a valuable auxiliary tool to plan cardiac rehabilitation programs, particularly in cases when BET is difficult or impossible to perform. The use of SHAP analysis helped to understand the contribution of each feature to the prediction, increasing the confidence in the model results.</p>","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 2","pages":"63-68"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13178085/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147966784","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Feasibility of Applying Proton Minibeam Technology on the Proton Therapy Complex \"Prometheus\".","authors":"O Yu Golovanova, G V Merzlikin, A D Kaprin, A N Soloviev, V O Saburov, S N Koryakin","doi":"10.17691/stm2026.18.4.01","DOIUrl":"10.17691/stm2026.18.4.01","url":null,"abstract":"<p><p>Spatial fractionation of proton beams is being considered as a new strategy to reduce the probability of side-effects in radiation therapy on the proton therapy complex (PTC) \"Prometheus\". By collimating the beams to submillimeter dimensions, the dose profile at the target surface consists of alternating high- and low-dose regions. This promotes sparing of normal tissues along the beam path through the activation of various biological mechanisms. <b>The aim of the study</b> was to implement the minibeam method on the PTC \"Prometheus\" and to experimentally assess the quality of the resulting dose distribution.</p><p><strong>Materials and methods: </strong>Collimation of the primary beam was achieved using brass multislit collimators. Mathematical modeling of the experiment was performed, and a phantom was irradiated from one and from two opposite directions. For effective summation of two fields and to achieve a homogeneous dose distribution in the target, intensity and energy modulation of the proton beams entering the collimator was carried out. To reduce errors caused by collimator deviation relative to the beam axis, a quality assurance system for collimator positioning on the treatment table was developed. Under the control of a scintillation detector and a laser positioner, positioning accuracy was within 0.2°. GafChromic EBT3 radiochromic films (Ashland, USA) were used as dosimeters.</p><p><strong>Results: </strong>When comparing calculated and experimental data, 90% of the distribution points matched the mathematical model according to the γ criterion (3%/3 mm). The quality of the minibeam field was assessed using the following parameters: valley-to-peak dose ratio; collimator throughput; ratio of the dose without the collimator to the dose with the collimator; and dose homogeneity index in the target. The obtained results met the quality assurance criteria for standard radiation therapy.</p><p><strong>Conclusion: </strong>Irradiation with proton minibeams using brass collimators is feasible on the PTC \"Prometheus\". A uniform dose distribution in the target area was experimentally achieved. The uniformity of coverage, the absence of local dose excesses, and the high dose gradient at the distal edge suggest that the described irradiation technique can be implemented on biological specimens.</p>","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 4","pages":"5-11"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13602024/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149042848","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Influence of Functional Magnetic Resonance Imaging Data Preprocessing Pipelines on the Accuracy of Schizophrenia Classification Using Machine Learning Methods.","authors":"A A Poyda, V A Orlov, A D Zhemchuzhnikov, S O Kozlov, S I Kartashov, L V Bravve, M A Kaydan, G P Kostyuk","doi":"10.17691/stm2026.18.3.02","DOIUrl":"10.17691/stm2026.18.3.02","url":null,"abstract":"<p><p><b>The aim of the study</b> was to analyze the influence of various pipelines for preprocessing raw functional magnetic resonance imaging (fMRI) data on the accuracy of classification of subjects into schizophrenia patients and healthy controls using machine learning methods, and to give recommendations for optimizing the data preprocessing pipeline for this task.</p><p><strong>Materials and methods: </strong>The study used fMRI data from 72 subjects acquired on a Siemens Magnetom Verio 3T MRI scanner (Siemens Healthineers, Germany) at the National Research Centre \"Kurchatov Institute\". Seven different preprocessing pipelines were used on each dataset. Feature vectors for classification were constructed from each preprocessed dataset. We applied the following three algorithms for feature vector construction: ReHo, FCM, and FHR. Classification was performed using 15 machine learning methods from the scikit-learn package for each feature set and each preprocessing pipeline. The final accuracy was determined as the maximum accuracy among all machine learning methods.</p><p><strong>Results: </strong>No single optimal preprocessing pipeline for all feature vector construction algorithms was found. Smoothing increased accuracy for the voxel metrics ReHo and FHR, but not for the regional metric FCM. Spatial smoothing was the only preprocessing step that affected classification accuracy for the FHR method. The FHR feature set without smoothing demonstrated accuracy close to chance level. The steps of frequency filtering and median normalization substantially increased the accuracy for both the ReHo and FCM methods, both with and without smoothing. The steps of inhomogeneity correction and slice timing correction did not increase accuracy for the FCM feature set but did improve it by several percent for the ReHo feature set. Application of ICA filtering changed accuracy within a range of 5%, in either a positive or negative direction.</p><p><strong>Conclusion: </strong>Based on the obtained results, the following recommendations can be given for fMRI data preprocessing pipelines for binary classification of subjects into schizophrenia patients and healthy controls using machine learning methods and taking into account the feature vector construction algorithm. The most suitable for the ReHo metric pipeline includes motion correction, normalization, inhomogeneity correction, slice timing correction, frequency filtering, median normalization, and spatial smoothing. The most suitable for the FCM metric pipeline includes motion correction, normalization, inhomogeneity correction, slice timing correction, frequency filtering, and median normalization. Spatial smoothing is the most important factor for the FHR metric. The ICA filtering step does not provide a clear benefit and should therefore be applied with caution.</p>","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 3","pages":"14-22"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13410868/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148623516","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integration of Transformer-Based Architecture and Large Language Models for Optical Coherence Tomography Data Analysis to Improve the Accuracy of Differential Diagnosis of Retinal Diseases.","authors":"O V Konshina, A D Pershin, M K Kulyabin, V I Borisov","doi":"10.17691/stm2026.18.3.01","DOIUrl":"10.17691/stm2026.18.3.01","url":null,"abstract":"<p><p><b>The aim of the present study</b> was to improve the differential diagnosis accuracy of retinal diseases combining a transformer model to classify the biomarkers on OCT images and the large language model DeepSeek-V3.</p><p><strong>Materials and methods: </strong>Two datasets were collected and annotated: a training set (3288 central retinal OCT images annotated for 8 biomarkers) and a validation set (50 clinical cases from octcases.com). For biomarker classification, we compared ResNet, DenseNet, EfficientNet, and Vision Transformer (ViT-Tiny-Patch16-224) architectures. The ViT-Tiny model demonstrated the highest performance, its F1 macro - 0.84±0.03. An integration algorithm was developed to combine predicted biomarker labels with clinical history data via the DeepSeek-V3 LLM API.</p><p><strong>Results: </strong>Combining OCT biomarkers and the medical history significantly improved the diagnostic accuracy: Top-1 Accuracy - 78%, Top-3 Accuracy - 94%, MRR - 84%, representing 10-44% improvement over using either data type alone.</p><p><strong>Conclusion: </strong>The suggested approach enabled to automate the detection of biomarkers on OCT images and enhance the differential diagnosis accuracy of eye diseases, reducing the image interpretation time and supporting clinical decision-making.</p>","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 3","pages":"5-12"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13410867/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148623569","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development of an Artificial Intelligence-Based System for Predicting Fall Risk in Neurological Patients.","authors":"E S Ikonnikova, A E Slotina, G A Belyankin, M A Zabello, V Yu Vishnyakov, A V Lobanov, N V Bedeneu, A A Dobrovolsky, D O Zemlyanaya, A Yu Tropynina, O A Kirichenko, A V Baidukova, N S Suponeva","doi":"10.17691/stm2026.18.4.05","DOIUrl":"10.17691/stm2026.18.4.05","url":null,"abstract":"<p><p><b>The aim of the study</b> was to develop an artificial intelligence model for predicting the probability of falls based on the analysis of video recordings of gait patterns.</p><p><strong>Materials and methods: </strong>The study involved 187 patients (median age 59 years) with neurological diseases of various etiologies who complained of impaired balance and unsteadiness while walking. All participants underwent video recording of their movements using a smartphone during a 10-meter walk test and Timed Up and Go test. Based on the analysis of 2077 steps recorded in 122 patients, an algorithm for fall risk stratification was developed using the YOLO-NAS Pose M architecture and a two-layer machine learning model.</p><p><strong>Results: </strong>The final model when analyzing video data achieved 76% accuracy in risk prediction, with an average absolute error of time parameter prediction of 1.445 s. The following parameters were identified as key biomechanical predictors of fall risk: step base width at the ankle joint level, lateral trunk sway, step base width at the knee joint level, vertical foot clearance during the swing phase, and temporal gait characteristics.</p>","PeriodicalId":520289,"journal":{"name":"Sovremennye tekhnologii v meditsine","volume":"18 4","pages":"42-55"},"PeriodicalIF":0.0,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13602023/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149042901","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}