Background: Postoperative skull bone defects are one of the urgent problems of neurorehabilitation. Skull bone defect limits the scope of rehabilitation measures, complicates patient care, and leads to secondary complications. Significant risks of postop complications demands making a decision about surgery individually. Surgery timing varies widely and remains controversal. Aims: to formulate the features of skull bone defects reconstruction in patients at various stages of rehabilitation based on the analysis of the frequency and structure of postoperative surgical complications. Materials and methods: The retrospective analysis of cranioplasty results was performed in the 129 patients treated in Federal Scientific and Clinical Center of Intensive Care Medicine and Rehabilitology from 2018 to 2022 at various stages of rehabilitation (intensive care, inpatient, outpatient). The features of surgery, frequency and structure of surgical complications dependent on rehabilitation stage were analyzed. Results: A total of 129 patients were included in the study: 84 men (65%) and 45 women (35%). The average age of the patients was 43.213.9 years. The average timing of cranioplasty surgery was 79 days [IQR 60; 133]. Seventy two patients (56%) were operated on at the intensive care stage of rehabilitation, forty (31%) and seventeen (13%) patients were operated on at the inpatient and outpatient stages, respectively. In total, complications occurred in 16 patients (12%). Intensive care patients required careful preoperative preparation, correction of homeostasis and metabolism disorders. In our series, postoperative complications were observed in 12 patients on intensive care stage (17%); all cases of hydrocephalus occurred only in intensive care patients. In patients operated on at the inpatient stage of rehabilitation, complications occurred in 4 cases (10%), while there was no statistically significant difference in the incidence of complications in patients from the intensive care and inpatient subgroups (p=0.334). Complications in patients at the outpatient stage of rehabilitation were not observed in our series. Conclusions: Cranioplasty surgery is possible even in somatically burdened patients at the intensive care stage of rehabilitation. It allows to expand the scope of rehabilitation measures and facilitate medical care. When planning surgical intervention in the early stages after the cranioplasty surgery, it is important to take into account the increased risk of hydrocephalus manifestation.

Large generative models (LGMs) have significant potential for healthcare and medical science. While publications are growing exponentially, LGM studies lack quality and breakthrough findings. Research articles call for standardized approaches to ensure safe and effective integration of LGMs into clinical practice. Currently, the Moscow healthcare system is testing LGMs as tools for supporting medical decision-making, which has required development of specialized methods and techniques for assessing LGM quality. This paper presents two methods for assessing the quality of large generative models. Both methods are based on analysis of literature data (over 200 sources), results from comprehensive testing of 204 LGMs, and hands-on experience in assessing model quality using a sample of more than 12,000 cases. Designed for two main LGM application scenarios, the methods incorporate a dedicated approach to building test samples, tailored and validated questionnaires, testing methodologies, and unified requirements for the composition and structure of quality assessment outputs.

Background. Management of patients in acute period of aneurysmal subarachnoid hemorrhage Hunt—Hess grade IV-V is still a complex and unsolved problem, especially regarding surgical treatment and various complications. Objective. To analyze postoperative outcomes in patients undergoing surgery in acute period of aneurysmal subarachnoid hemorrhage Hunt—Hess grade IV—V between 2006 and 2020, as well as to assess various factors influencing treatment outcomes. Material and methods. We analyzed 163 patients who underwent surgery within 21 days after aneurysmal subarachnoid hemorrhage Hunt—Hess grade IV—V. All patients were divided into 2 groups depending on the period: 2006—2011 (group 1) and 2012—2020 (group 2). Results. Mortality reduced from 28.6% in group 1 to 8.3% in group 2. At the same time, incidence of vegetative state (GOS grade 2) increased from 4.8% to 17.4%. Incidence of outcomes GOS grade 3 — 5 was similar. Conclusion. Large or giant aneurysm, repeated preoperative subarachnoid hemorrhage, intraoperative aneurysm rupture and prolonged temporary clipping can impair postoperative outcomes in patients with hemorrhage Hunt—Hess grade IV—V. External ventricular drainage, intraarterial injection of verapamil, intracisternal administration of calcium channel blockers and active surgical tactics improve postoperative outcomes and reduce mortality.

Lung cancer is one of the most common and socially significant cancers worldwide and consists of two main subtypes: small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), which require different treatments. Computed tomography (CT) scans cannot reliably differentiate these subtypes, often necessitating invasive biopsies that carry significant risks. Radiomics offers a promising non-invasive alternative by quantitatively analyzing imaging data to extract detailed tissue characteristics beyond visual assessment. This pilot retrospective study analyzed 200 Moscow patients with histologically confirmed SCLC or NSCLC. Manual tumor segmentation on pretreatment CT scans allowed extraction of 107 radiomic features, from which 16 key features were selected to train four machine learning models. Models were evaluated using stratified 5-fold cross-validation, focusing on ROC AUC, accuracy, precision, and recall. All models demonstrated strong performance in distinguishing SCLC from NSCLC, with the gradient boosting model achieving the highest accuracy of 80.5% and ROC AUC of 0.888. These results highlight the potential of radiomics combined with machine learning to enable accurate, non-invasive differentiation of lung cancer subtypes. Further research is needed to expand feature sets, develop automated segmentation tools, and enhance clinical application of this approach.

Large language models (LLMs) are increasingly being used in radiology-related workflows, but their application to reference, regulatory, and methodological queries remains limited by hallucinations and the static nature of model knowledge. This study aimed to develop and evaluate a retrieval-augmented generation (RAG) system for radiologists designed to provide grounded responses to such queries. A knowledge base was created through a survey of practicing radiologists and expert validation of sources, resulting in a corpus of 1049 documents. The system incorporated structured document parsing, a two-level parent–child vector database, hybrid dense–sparse retrieval, reranking, and a local large language model. Performance was assessed through functional testing, automated LLM-as-a-judge metrics, and multireader expert evaluation by 16 radiologists using 400 technical queries. No hallucinations were detected in the 77-query functional testing set during expert review. On the full technical dataset, automated Contextual Precision, Contextual Recall, and Answer Relevancy were 0.735, 0.881, and 0.890, respectively. Expert evaluation showed high response accuracy (mean, 4.53/5) and high expert-assessed Contextual Precision (0.886). Inter-expert agreement was substantial to excellent for most Likert-scale criteria. These findings suggest that a hierarchical RAG architecture can provide reliable access to radiology-specific reference information, although external validation and automated updating of the knowledge base remain necessary.

The paper analyzes the experience of using epidural spinal cord stimulation in tonic and burst stimulation modes in 71 patients with the consequences of damage to the central nervous system, who underwent treatment and rehabilitation at the Research Institute of Rehabilitology named after professor I.V. Pryanikov, FSCC RR, from 2020 to 2021. The efficacy of epidural stimulation in the indicated modes for clinical manifestations of pyramidal syndrome in the extremities, chronic disorders of consciousness, neuropathic pain, and paresis of the extremities without spasticity has been studied. The use of invasive spinal cord stimulation in the burst mode in patients with spasticity and chronic disorders of consciousness is described for the first time. The impact of various factors on the effectiveness of stimulation has also been analyzed. The average age of patients in the study group was 47 ± 17 years, the average duration of test epidural stimulation was 4.3 ± 1.6 days. Twenty-eight patients (39.5%) were stimulated in the tonic mode, twenty-two patients (31%) were stimulated in the burst mode, and the remaining patients were stimulated in the combined mode. Results: positive dynamics after the use of epidural stimulation was observed in 39 patients (54.9%). The spasticity regressed in 32 patients (52.5%); in addition, an increase in the active range of motion was observed in 23 patients (37.7%) from this subgroup. The decrease in the severity of spasticity in the extremities ranged from 0.33 ± 0.1 to 0.63 ± 0.13 points (p<0.005). The signs of an increase in the level of consciousness as a result of the stimulation were noted in 8 patients (38.1%); the neuropathic pain syndrome regressed in 8 patients (72.7%). When applying this technique in patients with paresis without spasticity, the stimulation did not have an effect in any patients of this subgroup. There was no statistically significant influence of etiology, age of patients or duration of stimulation on its effect. The patients with spasticity, who were stimulated in the tonic mode, more often demonstrated the effectiveness of stimulation compared with the patients from the combined mode subgroup (p<0.013). Higher effectiveness of the burst mode stimulation compared to the combined mode stimulation in patients with spasticity was on the borderline of statistical significance (p=0.065). Thus, the use of epidural stimulation in patients with chronic neuropathic pain and spasticity showed the best results; the effect of epidural stimulation on the level of consciousness in chronic disorders requires additional studies and changes in their design.