Validation of a Russian automated CT perfusion processing system for ischemic stroke

Full Text

List of abbreviations

ICA — Internal Carotid Artery

PCA — Posterior Cerebral Artery

CT — Computed Tomography

CTP — CT Perfusion

ACA — Anterior Cerebral Artery

MCA — Middle Cerebral Artery

A1–2 — A1–A2 segments of the Anterior Cerebral Artery

M1–3 — M1–M3 segments of the Middle Cerebral Artery

P1 — P1 segment of the Posterior Cerebral Artery

rCBF (Relative Cerebral Blood Flow) — relative cerebral blood flow

rCBV (Relative Cerebral Blood Volume) — relative cerebral blood volume

rMTT (Relative Mean Transit Time) — relative mean transit time

Tmax (Time to Maximum) — time to maximum of the residue function

TTP (Time to Peak) — time to peak contrast concentration

BACKGROUND

At present, automated segmentation of perfusion computed tomography (CTP) data in patients with ischemic stroke is widely used in clinical practice and has become an integral part of the diagnostic workflow as well as a critical tool for decision-making regarding reperfusion therapy [1]. Automated CTP post-processing systems such as RAPID (iSchemaView/RapidAI, USA) and Olea Sphere (Olea Medical, France) are now widely implemented in clinical practice and have enabled a large number of clinical studies, which substantially extended the indications for intravenous thrombolysis and mechanical thrombectomy into extended time windows [2]. These systems provide rapid, objective, and minimally operator-dependent quantitative assessment of infarct core and penumbra, which is crucial for optimizing patient selection for reperfusion therapies.

The clinical relevance of automated CTP analysis has been demonstrated in pivotal randomized clinical trials. As early as 2015, the EXTEND-IA and SWIFT PRIME trials used automated CTP processing with RAPID to select patients with acute ischemic stroke for mechanical thrombectomy in the early therapeutic window (up to 6 hours); in both trials, CTP-assisted selection was associated with high rates of functional recovery, with 71% and 60% of patients, respectively, achieving functional independence (modified Rankin Scale score 0–2) at 90 days [3–5]. Subsequently, the DAWN (2018) and DEFUSE 3 (2018) trials showed that the use of RAPID for automated core volume estimation allows safe extension of the treatment window for reperfusion therapy (mechanical thrombectomy) up to 24 hours: in DEFUSE 3, functional independence was achieved in 45% of patients in the intervention group versus 17% in the medical-therapy group [6, 7]. The RAPID platform was also used as a selection tool in EXTEND (thrombolysis up to 9 hours) and DEFUSE 1–2, collectively forming the evidence base underlying current recommendations for reperfusion therapy [8].

Importantly, the international neurology and neuroradiology community recognizes the use of automated CTP analysis systems as a standard of care in acute ischemic stroke. The updated 2026 guidelines of the American Heart Association/American Stroke Association (AHA/ASA) on the early management of patients with acute ischemic stroke formalize an approach based on quantitative assessment of penumbra and core rather than solely on time-based criteria, and assign a central role to perfusion imaging for patient selection in extended time windows for both thrombolysis (up to 9 hours) and mechanical thrombectomy (up to 24 hours) [9]. The global expansion of automated CTP platforms reflects this trend: for example, in the Chinese multicenter iStroke (Beijing, China) study, a domestically developed automated CTP analysis system demonstrated strong correlation with RAPID in estimating infarct core and critically hypoperfused tissue volumes (ρ = 0.68 and ρ = 0.66, respectively), with comparable prognostic performance for 90-day clinical outcomes, which supported its implementation across Chinese stroke centers [10].

Thus, having a locally validated automated CTP analysis system is a critical prerequisite for full implementation of modern reperfusion therapy standards. However, in the Russian Federation there is currently no domestically developed CTP post-processing system that has undergone sufficient clinical validation. In this context, the present study is devoted to the validation of the Russian solution ArchiMed PRO Chronos through comparison with one of the most widely used CTP processing systems, Vitrea (Canon Medical Systems, Japan), which is approved for clinical use in the Russian Federation.

Study aim — to validate the Russian ArchiMed PRO Chronos system by comparing it with the Vitrea (Canon Medical Systems) platform in the quantitative assessment of ischemic core and critically hypoperfused tissue volumes.

METHODS

Study design

This was a retrospective study based on imaging data from patients with ischemic stroke who were hospitalized in a single medical center.

Study setting

The study was conducted between January and February 2026 and included patients consecutively admitted from April 2024 to January 2025 at the City Clinical Hospital named after S.S. Yudin, Moscow Department of Health, Moscow, Russian Federation.

Eligibility criteria

Inclusion criteria: a clinical diagnosis of ischemic stroke with symptom onset no more than 24 hours before presentation to the emergency department; age ≥18 years; completion of a standard multimodal CT protocol of the brain, including non-contrast CT, CT angiography of the brachiocephalic arteries, and CT perfusion; presence of occlusion of one of the following arteries: internal carotid artery (ICA), M1–M3 segments of the middle cerebral artery (MCA M1–3), A1–A2 segments of the anterior cerebral artery (ACA A1–2), or the P1 segment of the posterior cerebral artery (PCA P1).

Exclusion criteria: presence of intracranial hemorrhage or any other major brain pathology unrelated to ischemic stroke; inadequate image quality of at least one examination precluding reliable post-processing; lack of coverage of the vascular territory of the occluded artery within the CTP scan volume; absence of occlusion of one of the following arteries: ICA, MCA M1–3, ACA A1–2, or PCA P1.

Description of eligibility criteria: the eligibility criteria were defined according to the aim of the study, which was to compare the results of automated CT perfusion post-processing in acute ischemic stroke within the territories of major cerebral arteries using two software packages (Vitrea, Canon Medical Systems, and the ArchiMed PRO Chronos system). Patients aged 18 years or older with clinically verified ischemic stroke and known or presumed symptom onset within 24 hours before admission, who underwent non-contrast CT, CT angiography, and brain CTP in the acute phase, were included. The presence of ICA, MCA M1–M3, ACA A1–2, or PCA P1 occlusion on CT angiography was mandatory. Patients with intracerebral hemorrhage or other significant brain pathology (mass lesions, vascular malformations, etc.) that could substantially affect interpretation of perfusion maps and ischemic lesion volumes were excluded. To minimize systematic error, cases with inadequate image quality in any modality (e.g., motion artifacts, pronounced beam-hardening artifacts) and those with incomplete CTP coverage of the territory supplied by the occluded artery were also excluded, as this would make quantitative estimation of infarct core and critically hypoperfused tissue unreliable. The choice of this patient category and the specified limits (age ≥18 years, therapeutic window up to 24 hours) reflects current concepts of the target population of ischemic stroke patients for whom perfusion imaging is indicated to guide reperfusion treatment and is consistent with the inclusion criteria of major trials that validated perfusion thresholds and software solutions for CTP analysis.

Grouping of participants: no formal matching was performed, as the study design implied within-subject comparison of CTP post-processing results obtained with two software packages in the same patients, i.e., each patient simultaneously served as a “case” and a “control” for the two methods. A single retrospective cohort of 52 consecutively included patients who met the eligibility criteria was formed; for each of them, the same baseline CTP dataset was processed in parallel on Vitrea and PRO Chronos workstations using the standard threshold settings for these systems. Thus, the ratio of observations between the compared methods was 1:1, and comparability of groups was ensured by the within-subject design itself, without the need for additional matching on other variables.

Description of the intervention

The study evaluated CT perfusion (CTP) data processing using two different approaches: the Vitrea workstation (Canon Medical Systems, Japan) and the ArchiMed PRO Chronos software package (Russia). Qualitative and quantitative assessment of infarct core and critically hypoperfused tissue volumes in acute ischemic stroke was performed using the standard threshold settings of both systems (rCBF < 30% and Tmax >6 s for Chronos, and rCBV < 38% with TTP >5.3 s or rMTT >55% for Vitrea, where rCBF is Relative Cerebral Blood Flow; rCBV is Relative Cerebral Blood Volume; Tmax is Time to Maximum of the residue function; TTP is Time to Peak; and rMTT is Relative Mean Transit Time). A total of 52 patients were included in the study.

Study endpoints

Primary endpoint. The primary endpoints were the volume of the ischemic core on CTP (in milliliters) automatically calculated by the PRO Chronos and Vitrea software, and the inter-software difference in core volume estimation (absolute and relative difference, and Pearson correlation coefficient r between the values obtained by the two methods). This quantitative parameter was considered key, since infarct core volume is the principal variable used both for selecting patients for endovascular treatment and for assessing the comparability of different CTP post-processing algorithms. The primary endpoint was treated as a continuous variable.

Secondary endpoints. Secondary endpoints included the volume of critically hypoperfused tissue on CTP, automatically calculated by each software package as a continuous quantitative variable.

Measurement of endpoints. All patients underwent non-contrast CT, CT angiography of the extra- and intracranial arteries, and brain CTP on a 160-slice CT scanner (Aquilion Prime SP, Canon Medical Systems, Japan) using the institution’s standard protocol. Raw dynamic CTP datasets were exported in DICOM format and independently processed on a Vitrea workstation (Canon Medical Systems, Japan) and in the ArchiMed PRO Chronos software (Russia) using the built-in algorithms for automated calculation of perfusion parameters and lesion volumes.

In PRO Chronos, the ischemic core was defined as tissue with a reduction in Relative Cerebral Blood Flow (rCBF) to < 30% of the value in the contralateral hemisphere; critically hypoperfused tissue was defined using the system’s standard threshold of Time to Maximum (Tmax) >6 s, consistent with values used in clinical trials. In Vitrea, the ischemic core was defined as tissue with a decrease in Relative Cerebral Blood Volume (rCBV) to < 38% of the contralateral values in combination with either Time to Peak (TTP) >5.3 s or Relative Mean Transit Time (rMTT) >55% of normal; critically hypoperfused tissue was delineated according to the corresponding built-in thresholds of the software. All volumes (core, hypoperfused tissue, and mismatch) were calculated automatically by the software and exported in tabular format for subsequent statistical analysis. Endpoint values were obtained directly from the primary CTP studies, without the use of secondary data sources from medical records.

Sensitivity analysis. No dedicated sensitivity analysis was planned or performed to test the robustness of the primary results to changes in the underlying assumptions of the study protocol. The study was retrospective, and the main focus was on assessing inter-software comparability of ischemic core and critically hypoperfused tissue volumes using standard, previously validated thresholds implemented in PRO Chronos and Vitrea. The results are presented without additional scenarios involving alternative thresholds, imputation or re-processing of missing data, or multivariable sensitivity checks in subgroups.

Statistics

Planned sample size. A formal a priori sample size calculation was not performed at the study planning stage. The retrospective cohort included all consecutively hospitalized patients during the study period who met the inclusion criteria and did not meet any of the exclusion criteria, resulting in a final sample of 52 observations. This approach was chosen to maximize the use of available clinical data while maintaining sample homogeneity with respect to key characteristics (stroke type, presence of large vessel occlusion, standard imaging protocol). No stopping rules (for example, early termination upon reaching statistical significance) were defined, as the analysis was conducted after inclusion of all eligible cases.

Statistical methods. Statistical analyses were performed using Python with the libraries numpy, pandas, statsmodels, scipy, seaborn, and matplotlib. Quantitative variables were described according to their distribution: for approximately normally distributed data as mean and standard deviation, and for clearly non-normal distributions as median and interquartile range (25th and 75th percentiles); categorical variables were summarized as absolute and relative frequencies (percentages). Paired quantitative measurements obtained from the two software packages (core volume and critically hypoperfused volume) were compared using parametric or non-parametric tests for paired samples, depending on data distribution; the strength of linear association between values derived from Vitrea and PRO Chronos was assessed using the Pearson correlation coefficient r. The level of statistical significance (p) for all tests was set at < 0.05, and all p-values were reported as two-sided. Missing data were not included in the final analysis, as studies with inadequate image quality or incomplete perfusion coverage had been excluded at the cohort formation stage; outlier handling and data transformation were performed only in the presence of obvious artefactual values and were not applied systematically.

RESULTS

Sample formation

A total of 52 patients who met the inclusion criteria were enrolled in the retrospective analysis (clinical diagnosis of ischemic stroke with symptom onset within 24 hours before admission, age ≥18 years, completion of non-contrast CT, CT angiography, and CT perfusion, and presence of occlusion of the ICA, MCA M1–M3, ACA A1–A2, or PCA P1 segment). All 52 patients selected according to the eligibility criteria were included in the primary statistical analysis without loss at the cohort formation stage, as the study did not involve prospective outcome follow-up and focused solely on the processing of baseline CTP data by two software systems.

Sample characteristics

The final cohort consisted of 52 patients with acute ischemic stroke due to large vessel occlusion who underwent non-contrast CT, CT angiography, and CTP according to a unified protocol on an Aquilion Prime SP scanner (Canon Medical Systems, Japan). All patients were older than 18 years and were imaged within a therapeutic window of up to 24 hours from symptom onset; additional demographic and clinical characteristics (age subgroups, sex, stroke severity, etc.) were not structured within this analysis and are therefore not reported in the Results section. Comparison between included and non-included patients could not be performed due to the retrospective nature of the study.

Main results

For quantitative analysis, the primary metrics were the infarct core volume and the volume of critically hypoperfused tissue, as directly obtained from segmentation of the time-to-perfusion maps, rather than penumbra volume, which is calculated as the difference between the critically hypoperfused volume and core volume and therefore depends on the quality of both perfusion maps. Measurements from both software packages were compared, and the correlation between them was assessed (Fig. 1).

 

Fig. 1. Volume of the ischemic core (a) and critically hypoperfused tissue (b) measured using the Chronos and Vitrea systems.

 

Pearson correlation analysis showed a statistically significant correlation for infarct core volume, with a correlation coefficient of 0.95, and for critically hypoperfused volume, with a correlation coefficient of 0.85.

Normality of the distributions was assessed using the Shapiro–Wilk test; since the significance level was >0.05, the non-parametric Wilcoxon signed-rank test for paired samples was chosen. When comparing infarct core volumes and critically hypoperfused volumes between the two software packages, p-values were >0.05, indicating no significant differences between the methods in the estimation of these volumes.

All automated segmentations were visually reviewed by an expert and classified into two categories: adequate segmentation without extraneous regions and inadequate segmentation, in which the segmented area extended beyond the vascular territory of the occluded large artery (examples shown in Fig. 2).

 

Fig. 2. Example of inadequate segmentation using the Chronos (a) and Vitrea (b) systems. The segmentation maps show areas of ischemic core outside the vascular territory of the occluded large artery (red), in both cases involving the right middle cerebral artery.

 

The number of studies with adequate segmentation was 44/52 (84.6%) for Chronos and 44/52 (84.6%) for Vitrea. The total number of studies with simultaneously adequate segmentation by both methods was 40/52 (76.9%). The measurement results after exclusion of cases with inadequate segmentation are shown in Fig. 3.

 

Fig. 3. Volume of the ischemic core (a) and critically hypoperfused tissue (b) measured using the Chronos and Vitrea systems after exclusion of studies with inadequate segmentation.

 

A similar correlation analysis was then performed using Pearson’s test, demonstrating statistically significant correlations for infarct core volume (r = 0.98) and critically hypoperfused volume (r = 0.93).

Normality was again assessed with the Shapiro–Wilk test, and as the significance level remained >0.05, the Wilcoxon signed-rank test was used for paired samples. The p-values for differences in infarct core volume and critically hypoperfused volume between the two software packages were >0.05, again indicating no significant difference in their estimates when using the two CTP processing methods.

DISCUSSION

Summary of the main findings

This study demonstrates that the Russian automated CTP processing system ArchiMed PRO Chronos provides a high level of agreement with the Vitrea Canon Medical Systems platform in the quantitative assessment of ischemic core and critically hypoperfused tissue volumes in patients with ischemic stroke. Strong linear relationships between volumes calculated by the two systems were observed both in the overall cohort and after exclusion of cases with inadequate segmentation, with no statistically significant differences between quantitative estimates. The comparable frequency of adequate automated segmentation in both systems indicates a similar performance profile under real-world clinical conditions.

Interpretation of the findings

The present results show strong correlations between Chronos and Vitrea for both ischemic core volume (r = 0.95) and critically hypoperfused volume (r = 0.85). It should be noted that even the current gold standard RAPID has a sensitivity of only 40.5% for detecting ischemic lesions on CTP in the overall population of patients with acute ischemic stroke, and this limitation largely reflects intrinsic constraints of the CTP technique itself [11]. For example, for core volumes >70 mL, RAPID and Olea show sensitivities and specificities of 73.7% and 81.2% versus 73.7% and 68.3%, respectively [11]. Major limitations of CTP include the difficulty of data analysis in the posterior fossa, motion and beam-hardening artifacts, and insufficient spatial resolution for detecting small and lacunar infarcts. In view of these limitations, the present study focused only on patients with occlusion of a large artery in the carotid circulation or the posterior cerebral artery, with the target vascular territory located supratentorially.

It is important to emphasize the differences in data-processing algorithms between the systems under investigation. Chronos uses a deconvolution approach similar to that implemented in RAPID, as well as standard thresholds for defining ischemic core and critically hypoperfused tissue, with rCBF < 30% and Tmax >6 s, respectively, which correspond to widely accepted criteria [12]. In contrast, Vitrea employs a more advanced deconvolution method based on a Bayesian framework, but uses different thresholds for segmenting core and critically hypoperfused tissue (rCBV < 38% with TTP >5.3 s or rMTT >55%, respectively) [13]. Different deconvolution algorithms can lead to substantial discrepancies in absolute perfusion parameter values, and optimization of these thresholds across software platforms remains an ongoing process [14]. With respect to RAPID as the gold standard compared with Vitrea, previous work has demonstrated strong correlation in quantitative core estimates between the two systems [13], and Vitrea has even shown higher accuracy than RAPID for penumbra volume estimation [15].

In our study, strong correlations between Chronos and Vitrea were found both in the full patient cohort and after exclusion of cases with inadequate segmentation. After excluding such cases, correlation coefficients improved to 0.98 for core and 0.93 for critically hypoperfused tissue, indicating that segmentation quality has a substantial impact on the accuracy of quantitative measurements.

An important finding is that adequate segmentation was achieved in 84.6% of cases for each system independently. This suggests that both platforms have comparable capabilities for automated identification of brain infarction; however, segmentation was simultaneously adequate in only 76.9% of studies, indicating some differences in their approaches. In most instances, inadequate segmentation was related to erroneous identification of core or critically hypoperfused regions outside the vascular territory of the occluded large artery, likely reflecting both the quality of perfusion data (including motion and bone-related artifacts) and the performance of the segmentation algorithms at the final stages of perfusion map processing.

Using the Wilcoxon signed-rank test for paired samples after exclusion of studies with inadequate segmentation, no statistically significant differences were found between ischemic core and critically hypoperfused volumes obtained with Chronos and Vitrea. This confirms that the domestic ArchiMed PRO Chronos system provides quantitatively comparable results to the widely used Vitrea platform in the analysis of CTP data. It is also noteworthy that the standard thresholds rCBF < 30% and Tmax >6 s used in Chronos are aligned with RAPID criteria and those employed in the majority of clinical studies of CTP in extended time windows for reperfusion therapy in ischemic stroke [16]; the present results therefore support the feasibility of implementing Chronos in stroke centers within the Russian Federation.

These findings are consistent with a growing body of evidence indicating that, despite fundamental differences in deconvolution algorithms and segmentation thresholds across vendors, automated CTP processing systems can provide comparable estimates of ischemic core and critically hypoperfused tissue when threshold values are appropriately standardized and validated, thereby enabling similar performance in selecting candidates for reperfusion therapy. In a large multicenter study by N. Kim et al. (n = 327) [17] comparing two independent automated CTP platforms, strong agreement was demonstrated for ischemic core volume (concordance correlation coefficient p = 0.942) and critically hypoperfused volume (p = 0.835) within the first 24 hours from symptom onset. Similarly, Yedavalli et al. [18] compared Olea and RAPID for core volume estimation in a multicenter cohort and showed that, with comparable thresholds, the systems yield similar results, although penumbra volumes exhibited greater variability [11].

It should be emphasized that achieving equivalent results across software platforms depends on harmonizing scan parameters and performing detailed validation under local conditions, since in addition to different deconvolution algorithms vendors often implement distinct methods for selecting the arterial input function (AIF) and venous output function (VOF), as well as different motion-correction strategies [19]. A recent multicenter study by Alwood et al. [20], which analyzed 362 patients, reported statistically significant differences in ischemic core volume estimates between the Viz.ai and Rapid systems (median 25.9 cm³ vs. 18.2 cm³, p< 0.001); however, this did not translate into significant differences in overall identification of thrombectomy candidates according to DEFUSE-3 criteria (p = 0.68).

Furthermore, the study by K.J. Chung et al. [14] demonstrated the feasibility of systematic calibration of ischemic core thresholds across different deconvolution algorithms using a digital perfusion phantom, which may facilitate harmonization of core volumes and core–penumbra profiles generated by various software solutions. This approach underscores the importance of recognizing that different platforms can be aligned through threshold adjustment to improve inter-software agreement. It also highlights the need for rigorous standardization and validation of any new or newly implemented CTP analysis system under local conditions and scanning protocols before clinical deployment, particularly in extended time windows for reperfusion therapy, where accurate quantitative assessment of ischemic core and penumbra is critical for selecting patients with the greatest potential benefit from reperfusion treatment.

Limitations

This study was conducted in a relatively small sample (n=52) from a single stroke center. The comparison was performed as a within-subject evaluation of two software packages on the same dataset, without forming independent clinical groups; therefore, a rigorous assessment of the impact of potential confounders (such as differences in clinical profile, time from symptom onset, or treatment strategy) on segmentation results is not possible within this design. At the same time, this approach minimizes between-group differences and makes algorithm-related differences in data processing the main source of variability, although it does not completely eliminate the possible influence of primary perfusion data quality (motion artefacts, bolus characteristics, scanning parameters). In addition, the analysis was limited to perfusion parameters and automatically derived core and critically hypoperfused tissue volumes, without comparison to reference morphologic outcomes (e.g., final infarct volume on follow-up MRI or CT), which precludes direct conclusions about the diagnostic accuracy of the two systems and allows assessment only of inter-software agreement in quantitative estimates. The threshold values for rCBF/rCBV and temporal parameters (Tmax, TTP, rMTT) are built-in for Chronos and Vitrea and are based on published data on perfusion cut-offs.

CONCLUSION

This study was designed to address the clinical validation of the Russian automated CTP processing system ArchiMed PRO Chronos by comparing it with the widely used Vitrea Canon Medical Systems platform in patients with acute ischemic stroke and large vessel occlusion. A high level of inter-software agreement was demonstrated for ischemic core and critically hypoperfused tissue volumes, with comparable rates of adequate automated segmentation, supporting the feasibility of using Chronos for quantitative CTP analysis in stroke centers. Even at this stage, the findings support integration of the domestic system into clinical workflows for selecting candidates for reperfusion therapy and contribute to the development of perfusion imaging analysis pathways that are independent of foreign software solutions in the setting of acute ischemic stroke.

ADDITIONAL INFORMATION

Author contributions: I.L. Gubskiy, definition of the concept, data analysis, revision and editing of the manuscript, research management; M.M. Beregov, data analysis, working with data, revision and editing of the manuscript; N.E. Staroverov, I.A. Larionov, software, working with data; K.Yu. Kazachkov, A.P. Stepanchenko, V.A. Nechaev, research support, validation; N.A. Marskaya, data management, validation; N.A. Shamalov, research management, validation. Thereby, all authors provided approval of the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics approval: The study was approved by the Local Ethics Committee of the Pirogov Russian National Research Medical University, Ministry of Health of the Russian Federation, Moscow, Russia (protocol No. 256 dated 15 December 2025).

Funding sources: This work was partially supported within the framework of the research project «Treatment Effect 2025–2027», registration card No. 125052706451-2.

Disclosure of interests: I.L. Gubskiy, M.M. Beregov, N.E. Staroverov, and I.A. Larionov are authors of the computer software “APC ArchiMed PRO Brain Perfusion” and “APC ArchiMed PRO Chronos” (RU2022664654, RU2022664655). N.E. Staroverov and I.A. Larionov are employees of the developer company of the Chronos system (ArtVision LLC).

Statement of originality: The results obtained are being published in open access for the first time.

Data availability statement: The authors provide restricted access to the study data upon reasonable request; no information containing patients’ personal data is shared.

Generative AI: Generative AI tools were used during the preparation of this manuscript. For searching and preliminary analysis of the scientific literature on automated CT perfusion analysis, the following tools were employed: (1) Perplexity AI (web service https://www.perplexity.ai, developed by Perplexity AI, Inc., San Francisco, USA), used in January–February 2026; (2) Scopus AI, a generative AI assistant integrated into the Scopus database (developed by Elsevier, The Netherlands, accessed via the Scopus platform, https://www.scopus.com), also in January–February 2026. All texts, references, and interpretations suggested by AI tools during the literature search were critically reviewed and edited by the authors, who take full responsibility for the final content of the manuscript.

About the authors

Ilya L. Gubskiy

The Russian National Research Medical University named after N.I. Pirogov

Author for correspondence.
Email: gubskiy.ilya@gmail.com
ORCID iD: 0000-0003-1726-6801
SPIN-code: 9181-3091

MD, PhD

Russian Federation, Moscow

Mikhail M. Beregov

Federal Center of Brain Research and Neurotechnologies

Email: mik.beregov@gmail.com
ORCID iD: 0000-0003-1899-8131
SPIN-code: 2559-0307
Russian Federation, Moscow

Nikolai E. Staroverov

Saint Petersburg Electrotechnical University LETI

Email: nik0205st@mail.ru
ORCID iD: 0000-0002-4404-5222
SPIN-code: 3147-2108

PhD

Russian Federation, Saint Petersburg

Ivan A. Larionov

Saint Petersburg Electrotechnical University LETI

Email: ivan.al.larionov@gmail.com
ORCID iD: 0000-0001-9620-9471
SPIN-code: 8210-9840

PhD

Russian Federation, Saint Petersburg

Kirill Yu. Kazachkov

City Clinical Hospital named after S.S. Yudin

Email: Kir-82@mail.ru
ORCID iD: 0009-0009-0142-1177
SPIN-code: 1257-8900
Russian Federation, Moscow

Andrey P. Stepanchenko

City Clinical Hospital named after S.S. Yudin

Email: aps65@mail.ru
ORCID iD: 0000-0001-5655-2929
SPIN-code: 7369-4096

MD, PhD

Russian Federation, Moscow

Valentin A. Nechaev

City Clinical Hospital named after S.S. Yudin

Email: dfkz2005@gmail.com
ORCID iD: 0000-0002-6716-5593
SPIN-code: 2527-0130

MD, PhD

Russian Federation, Moscow

Nataliya A. Marskaya

Federal Center of Brain Research and Neurotechnologies

Email: marskayana@gmail.com
ORCID iD: 0000-0002-0789-4823
SPIN-code: 5578-2649
Russian Federation, Moscow

Nikolay A. Shamalov

Federal Center of Brain Research and Neurotechnologies

Email: shamalovn@gmail.com
ORCID iD: 0000-0001-6250-0762
SPIN-code: 2865-9817

MD, PhD

Russian Federation, Moscow

References

Supplementary files