Thromboelastometry-Based Prophylaxis for Venous Thromboembolism in the Acute Period Following Isolated Severe Traumatic Brain Injury.

Objective: Traumatic brain injury (TBI) is an independent risk factor for venous thromboembolism (VTE). This study aimed to determine the optimal timing for initiating pharmacological thromboprophylaxis for VTE in patients with isolated severe TBI using rotational thromboelastometry (ROTEM). Methods: This single-center observational study enrolled 115 patients aged 18-59 years with isolated severe TBI within the first 48 hours after injury. Results: Using ROTEM data, we identified hypercoagulation due to an increase in clot density (MCF EXTEM >72), which was attributed to fibrinogen (MCF FIBTEM >25). From day 4, hypercoagulation occurred in 14.8% of the patients. By day 7, these changes were observed in 85.2% of patients. According to brain computed tomography findings, patients who received early VTE chemoprophylaxis on days 3-4 after severe TBI did not experience progression of hemorrhagic foci. Conclusion: Our results emphasize the clinical significance of thromboelastometry in patients with isolated severe traumatic TBI. Anticoagulant prophylaxis started on 3-4 days after severe TBI was relatively safe, and most patients did not experience hemorrhagic foci progression. The data acquired in this study may enable the optimization of VTE chemoprophylactic approaches, thereby reducing the associated risks to patients.

Reconstruction of orbital walls after resection of cranioorbital meningiomas: a systematic review and meta-analysis of individual patient data.

Following meningioma removal, there are numerous methods available for reconstructing the orbital wall. This systematic review seeks to summarize the published data on the surgical treatment of cranioorbital meningiomas, and to analyze the effectiveness and safety of various techniques and materials used for the reconstruction of bony orbital walls. We conducted a search of the two databases and included original articles with a series of 10 or more cases. Descriptive statistics and meta-analysis of individual patient date were performed. The analysis included a total of 858 patients from 29 sources. No reconstruction of the orbital walls was performed in 525 patients (61.2%), while 333 observations (38.8%) involved resection followed by reconstruction. A relative improvement in eye position was achieved in 94.4% of cases with a 95% CI of (88.92%; 97.25%). However, normalization of eye position, regardless of reconstruction technique, was only present in 6.22% of cases with a 95% CI of (1.24%; 25.9%). The best results were observed with the use of autologous bone implants (64%, 95% CI [33.35%; 86.33%]) and titanium implants (55.78%, 95% CI [2.86%; 98.18%]). In cases of endoscopic resection and microsurgical resection without reconstruction, symmetrical eye position accounted for only 1.94% (95% CI [0%; 96.71%]) and 2.35% (95% CI [0.13%; 31.23%]), respectively. The frequency of normalization of eye position differed significantly (p < 0.01) among the subgroups. A total of 49 postoperative complications were registered, with wound infection (1.52%, 95% CI [0.86%; 2.65%]) and wound cerebrospinal fluid leak (1.32%, 95% CI [0.6%; 2.91%]) being the most frequent. No significant differences were found in the rates of complications among the different subgroups. One of the primary objectives of cranioorbital meningioma surgery is to correct the position of the eye. Simultaneous reconstruction of the bony orbital leads to better cosmetic outcomes. Postoperative complications did not depend on the reconstructive technique or the materials.

Intraoperative Neurophysiologic Monitoring in Predicting Dysphagia After Brainstem and Fourth Ventricle Surgery.

OBJECTIVE: Dysphagia represents the main complication of posterior fossa neurosurgery. Adequate diagnosis of this complication is warranted to prevent untimely extubation with subsequent aspiration. Intraoperative neurophysiologic monitoring (IONM) modalities may be used for this purpose. However, it is not known which IONM modality may be significant for diagnosis. This study aimed to define the most significant IONM modality for dysphagia prognostication after posterior fossa neurosurgery. METHODS: The analysis included 46 patients (34 with tumors of the fourth ventricle and 12 with brainstem localization) who underwent surgical excision of the tumor. Neurologic symptoms before and after neurosurgery were noted and magnetic resonance imaging with the subsequent volume estimation of the removed mass was performed, followed by an IONM findings analysis (mapping of the nucleus of the caudal cranial nerves [CN] and corticobulbar motor-evoked potentials [CoMEP]). RESULTS: Aggravation of dysphagia was noted in 24% of the patients, more often in patients with tumor localization in the fourth ventricle (26%) than in those with brainstem mass lesions (16%). Mapping of the caudal cranial nerve nuclei did not correlate with the dysfunction of these structures. CoMEP was significantly associated with the neurologic state of the CN. The decrease in CoMEP is a significant prognostic factor for postoperative bulbar symptoms appearance or aggravation. CONCLUSIONS: Mapping the CN is an important identification tool. The CoMEP modality should be used intraoperatively to determine the functional state of the CN and predict postoperative dysphagia.

Data Quality Estimation Via Model Performance: Machine Learning as a Validation Tool.

In our recent study, the attempt to classify neurosurgical operative reports into routinely used expert-derived classes exhibited an F-score not exceeding 0.74. This study aimed to test how improving the classifier (target variable) affected the short text classification with deep learning on real-world data. We redesigned the target variable based on three strict principles when applicable: pathology, localization, and manipulation type. The deep learning significantly improved with the best result of operative report classification into 13 classes (accuracy = 0.995, F1 = 0.990). Reasonable text classification with machine learning should be a two-way process: the model performance must be ensured by the unambiguous textual representation reflected in corresponding target variables. At the same time, the validity of human-generated codification can be inspected via machine learning.

Multinomial Classification of Neurosurgical Operations Using Gradient Boosting and Deep Learning Algorithms.

This study aimed at testing the feasibility of neurosurgical procedures classification into 100+ classes using natural language processing and machine learning. A catboost algorithm and bidirectional recurrent neural network with a gated recurrent unit showed almost the same accuracy of ∼81%, with suggestions of correct class in top 2-3 scored classes up to 98.9%. The classification of neurosurgical procedures via machine learning appears to be a technically solvable task which can be additionally improved considering data enhancement and classes verification.

Predicting the Length of Stay in Neurosurgery with RuGPT-3 Language Model.

In this study, we update the evaluation of the Russian GPT3 model presented in our previous paper in predicting the length of stay (LOS) in neurosurgery. We aimed to assess the performance the Russian GPT-3 (ruGPT-3) language model in LOS prediction using narrative medical records in neurosurgery compared to doctors' and patients' expectations. Doctors appeared to have the most realistic LOS expectations (MAE = 2.54), while the model's predictions (MAE = 3.53) were closest to the patients' (MAE = 3.47) but inferior to them (p = 0.011). A detailed analysis showed a solid quality of ruGPT-3 performance based on narrative clinical texts. Considering our previous findings obtained with recurrent neural networks and FastText vector representation, we estimate the new result as important but probably improveable.

Surgical Treatment of Large and Giant Anterior Cerebral Artery Aneurysms in the Endovascular Era: Analysis of Large Clinical Series.

OBJECTIVE: Analysis of results of surgical treatment of 112 patients with large and giant anterior cerebral artery (ACA) aneurysms. METHODS: This is the largest clinical series of large and giant ACA aneurysms ever reported. Retrospective analysis of medical charts, radiographic studies, and intraoperative videos was performed. Univariate and multivariate analysis of relations among 3 types of outcomes (complete aneurysm occlusion, ischemic complications, and clinical outcomes) and predicting factors was performed. RESULTS: Most aneurysms were communicant (84.8%). Aneurysms of the A1 segment and distal segments (A2-A5) were encountered rarely. Complete aneurysm occlusion (class I) was shown in 95 cases (90.5%). In 4 patients, only exploration and wrapping were performed because of severe atherosclerosis and chronic intraluminal thrombi. Partial occlusion (class III) was shown in 2 patients, and neck residual (class II) in 4 patients. Cerebral ischemia developed in 29 patients postoperatively. Occlusion or injury of the A1 segment and anterior communicating artery perforators and recurrent branch of Heubner were the most common reason for ischemic complications. At follow-up, 97 patients (86.6%) had favorable outcomes. Ten patients (8.9%) had unfavorable outcomes because of postoperative complications. CONCLUSIONS: Microsurgery provides effective and relatively safe occlusion of complex ACA aneurysms. Direct clipping can be applied in most cases independent of their precise location. In certain cases, alternative methods can be safely used. According to multivariate analysis, giant aneurysm size, aneurysm-related mass effect, involvement of large arterial branches in the aneurysm wall, and intraoperative complications were significant predicting factors for ischemic complications.

A Comparison of Word Embeddings to Study Complications in Neurosurgery.

Our study aimed to compare the capability of different word embeddings to capture the semantic similarity of clinical concepts related to complications in neurosurgery at the level of medical experts. Eighty-four sets of word embeddings (based on Word2vec, GloVe, FastText, PMI, and BERT algorithms) were benchmarked in a clustering task. FastText model showed the best close to the medical expertise capability to group medical terms by their meaning (adjusted Rand index = 0.682). Word embedding models can accurately reflect clinical concepts' semantic and linguistic similarities, promising their robust usage in medical domain-specific NLP tasks.

Length of Stay Prediction in Neurosurgery with Russian GPT-3 Language Model Compared to Human Expectations.

Patients, relatives, doctors, and healthcare providers anticipate the evidence-based length of stay (LOS) prediction in neurosurgery. This study aimed to assess the quality of LOS prediction with the GPT3 language model upon the narrative medical records in neurosurgery comparing to doctors' and patients' expectations. We found no significant difference (p = 0.109) between doctors', patients', and model's predictions with neurosurgeons tending to be more accurate in prognosis. The modern neural network language models demonstrate feasibility in LOS prediction.

Best Research Practice Implementation: The Experience of the N.N. Burdenko National Medical Research Center of Neurosurgery.

Implementing the best research principles initiates an important shift in clinical research culture, improving efficiency and the level of evidence obtained. In this article, we share our own view on the best research practice and our experience introducing it into the scientific activities of the N.N. Burdenko National Medical Research Center of Neurosurgery (Moscow, Russian Federation). While being adherent to the principles described in the article, the percentage of publications in the international scientific journals in our Center has increased from 7% to 27%, with an overall gain in the number of articles by 2 times since 2014. We believe it is important that medical informatics professionals equally to medical experts involved in clinical research are familiar with the best research principles.

Hippocampal Dosimetry and Mnemonic Function Changes After Stereotactic Irradiation of Cavernous Sinus Meningiomas.

INTRODUCTION: It is believed that hippocampal exposure plays a major role in the development of memory disorders after cranial irradiation. This effect is evident in whole-brain irradiation and is less certain in local irradiation of intracranial targets. The present study aims to clarify the dosimetric features and dynamics of memory functions after local irradiation of the hippocampus when treating cavernous sinus meningiomas. METHODS: The study included 28 patients (24 females and 4 males) with cavernous sinus meningiomas diagnosed according to typical clinical and radiological findings. The mean age was 52 years (30-65 years). Stereotactic radiotherapy in standard fractionation regimen (54 Gy total dose) was the primary treatment in all patients. Patients underwent memory testing (ability to reproduce and recognize) using a previously developed and validated methodology at standard time points: before the start of radiotherapy, at the end of the course, and 6 and 12 months after treatment. Hippocampal dose, dynamics of memory function, and their possible relationship were evaluated. RESULTS: In total, 28 cavernous sinus meningiomas (15 left-sided and 13 right-sided) were treated. The mean target volume was 24.0 ccm (8.2 ccm to 53.8 ccm). Twelve months after radiotherapy, there was an increase in the median total number of recognition errors from 6.5 [4;11] to 9.5 [5;12], p=0.025, the median number of "old-similar" errors from 2 [1;3.25] to 3 [1.75;5], p=0.021, and the median number of "similar-old" errors from 3 [1;5] to 5.5 [3;7], p<0.001. The number of reproduction errors did not increase. A moderate correlation (p = 0.03, correlation coefficient = 0.41) was found between the dose to 10% of the ipsilateral hippocampus and the total number of reproduction errors at the end of the course. No other significant correlations were found at the end of radiotherapy and six and 12 months after it. CONCLUSION: Thus, even partial lateralized exposure of the hippocampus during irradiation of the cavernous sinus meningiomas affects its function in the form of specific pattern separation type disturbances, which are detected as early as 12 months after the impact. The hippocampus in this treatment should be considered as a critical structure whose sensitivity to irradiation requires additional assessment.

A Single-Center Retrospective Descriptive Cohort Study of 211 Pediatric Patients: Cerebrospinal Fluid Leakage After Fourth Ventricle Tumor Resection.

BACKGROUND: Cerebrospinal fluid (CSF) leak remains a significant complication of posterior fossa tumor surgery. The goal of this study was to evaluate the wound CSF leakage rate in pediatric patients who underwent fourth ventricle tumor resection in a single-site setting and to explore the association of CSF leakage with the length of stay in the intensive care unit (ICU) and with dural sealant application. METHODS: Two hundred and eleven patients who underwent fourth ventricle tumor surgery between 2008 and 2016 were included in this study. Patient data were evaluated retrospectively. RESULTS: Postoperative wound CSF leakage was observed in 6 patients (2.8%). One hundred and seventy-six patients (83.4%) stayed in the ICU for ≤1 day, and CSF leakage developed in 4 of these patients (2.3%). Thirty-five patients (16.6%) stayed in the ICU for >1 day, and CSF leakage was observed in 2 of these patients (5.7%). The observed difference was not statistically significant (P = 0.260). There were no statistically significant differences in the CSF leakage rates in groups with and without dural sealing before (n = 2 [3%] and n = 4 [2.8%], respectively; P = 0.99) and after (n = 2 [3.0%] and n = 4 [6.0%], respectively; P = 0.68) application of the propensity score matching algorithm. CONCLUSIONS: The low CSF leakage rate (2.8%) identified in this study was associated with a certain algorithm of patient management. There was no strong evidence that the CSF leakage rate was associated with the length of stay in the ICU or with dural sealant application.