Volume 42 · Issue 05 · 2026

Hemifacial spasm refers to repeated paroxysmal and involuntary convulsions of one facial expression muscle (orbicularis oculi, expression muscle, orbicularis oralis). Its onset mostly occurs in the lower eyelid, gradually expanding to the cheeks to all expression muscles on one side of the face, and in severe cases, it can involve the neck muscles; Symptoms are generally aggravated when fatigue, emotional agitation or tension, and in severe cases, difficulty in opening eyes and skewed mouth corners may occur. Very few patients have bilateral disease, but generally do not have bilateral simultaneous onset^[1,2]。 In some patients, facial twitching is accompanied by pulsatile murmur in the ear^[3,4,5]。 Treatment modalities for hemifacial spasm include drug therapy, botulinum toxin (botulinum toxin for short) injection, and microvascular decompression (MVD). The Cranial Nerve Professional Group of Peripheral Nerve Professional Committee of Chinese Medical Doctors Association compiled the Guidelines for the Treatment of Hemifacial Spasm (2026 Edition) (hereinafter referred to as the Guidelines) by summarizing the latest research progress at home and abroad and combining clinical practice, hoping to further standardize the diagnosis and treatment of hemifacial spasm, enhance the treatment effect, and finally improve the quality of life of patients.

Glioma is the most common primary central nervous system tumor in the cranium. The annual incidence rate of glioma in China is 5/100,000~8/100,000, and the 5-year mortality rate ranks third among all tumors, second only to pancreatic cancer and lung cancer^[1]。 Among the primary brain malignant tumors, glioblastoma (GBM, World Health Organization Grade 4) has the highest incidence, accounting for about 50%, and the annual incidence rate is about 3.2/100,000; Followed by diffuse astrocytoma with an annual incidence rate of 0.5/100,000^[2]。

To compare the efficacy of aggressive medical therapy versus stenting in patients with symptomatic intracranial atherosclerotic stenosis (sICAS) with hypoperfusion.

One hundred and twenty-fifth patients receiving aggressive medical therapy (medical group) and 257 patients undergoing stenting plus medical therapy (stent group) at the Cerebrovascular Department of Interventional Center, Zhengzhou University People′s Hospital from January 2020 to December 2022 were retrospectively enrolled. Primary endpoints included transient ischemic attack (TIA), stroke, all-cause death within 30 days, and new cerebral infarction in the target vascular territory from 30 days to 1 year. Secondary endpoints included ischemic stroke, TIA, hemorrhagic stroke, all-cause death, and cardiovascular events. The technical success, in-stent restenosis at 6 months, stenosis rate of the target vessel, and modified Rankin scale (mRS) score at the last follow-up were also assessed. Baseline characteristics and outcomes were compared before and after 1 ∶1 propensity score matching (PSM).

In the stent group, the postoperative stenosis rate [median(Q₁, Q₃)] was 15% (10%, 20%), with a technical success rate of 98.1% (252/257). At 6 months, in-stent restenosis occurred in 18.1% (25/138) of patients, and symptomatic restenosis in 4.3% (6/138). The stenosis rate at follow-up was 25%(20%, 35%). Before PSM, the primary endpoint rate was higher in the medical group [13.6% (17/125) vs. 7.0% (18/257), OR=2.090, 95%CI: 1.037-4.212, P=0.036], especially from 30 days to 1 year [11.2% (14/125) vs. 3.9% (10/257), OR=3.115, 95%CI: 1.342-7.230, P=0.006]. After PSM (106 cases per group), the stent group still had a lower primary endpoint rate [4.7% (5/106) vs. 13.2% (14/106), OR=3.074, 95%CI: 1.066-8.868, P=0.030] and lower incidence of stroke in the target territory [3.8% (4/106) vs. 11.3% (12/106), OR=3.255, 95%CI: 1.015-10.444, P=0.038]. There were no significant differences in secondary endpoints or mRS scores at the last follow-up between groups (all P>0.05).

Among sICAS patients with hypoperfusion, stenting may provide better clinical benefits than aggressive medical therapy alone, especially in reducing the risk of new cerebral infarction in the target vascular territory from 30 days to 1 year.

To analyze the clinical outcomes of endovascular recanalization in patients with symptomatic chronic internal carotid artery occlusion (CICAO), and to explore the influencing factors for successful recanalization.

A total of 106 consecutive patients with symptomatic CICAO who underwent endovascular recanalization in the Department of Neurosurgery, Beijing Hospital from May 2012 to November 2024 were retrospectively enrolled. Clinical data, imaging data and treatment outcomes were collected. Successful endovascular recanalization was defined as achieving the Thrombolysis in Cerebral Infarction (TICI) grade 3 post operation, and unsuccessful otherwise. Univariate analysis and multivariate logistic regression analyses were performed to explore the influencing factors for successful endovascular recanalization.

Of the 106 patients, 76 (71.7%) achieved successful recanalization and 30 (28.3%) failed. Perioperative complications occurred in 8 patients (7.5%), including 4 cases (3.8%) of new postoperative cerebral infarction, 2 cases (1.9%) of intracranial hemorrhage, and 2 cases (1.9%) of visual field defect; 2 patients with postoperative intracranial hemorrhage died. Univariate analysis demonstrated that patients in the successful recanalization group were younger than those in the failed recanalization group (P=0.020), and the proportions of patients with an occlusion-to-recanalization time of <6 months and those with a blunt or tapered stump at the origin of the internal carotid artery occlusion segment were both higher than those in the failed recanalization group (all P<0.05). There were no significant differences between the two groups in other factors, including sex, body mass index, lesion side, essential hypertension, diabetes mellitus, smoking history, alcohol consumption history, anesthesia method, length of occluded segment, calcification at the origin of the occluded internal carotid artery segment, ophthalmic artery reflux, and concomitant contralateral carotid artery stenosis or occlusion (all P>0.05). Multivariate logistic regression analysis revealed that an interval from occlusion to endovascular recanalization of <6 months (OR=16.50, 95%CI: 4.20-64.74, P<0.001), presence of a blunt stump (OR=13.04, 95%CI: 2.25-75.45, P=0.004) or tapered stump (OR=11.36, 95%CI: 2.05-63.06, P=0.005) were associated with a higher rate of successful recanalization in patients with symptomatic CICAO.

An interval from occlusion to endovascular recanalization of less than 6 months and the presence of a blunt or tapered stump predict higher rates of successful endovascular recanalization in patients with symptomatic CICAO.

To explore the influencing factors of symptomatic intracranial hemorrhage (sICH) after mechanical thrombectomy in patients with acute large-vessel occlusion type ischemic stroke (AIS-LVO) and to construct a nomogram model based on the findings.

The clinical data of 181 patients with AIS-LVO accompanied by large core infarction who were admitted to the Neurointervention Department of Nanyang Central Hospital from June 2022 to June 2025 were retrospectively collected. All patients received endovascular treatment (EVT). The patients were divided into the sICH group (38 cases) and the non-sICH group (143 cases) based on whether sICH occurred after EVT. Through univariate analyse and multivariate logistic regression analyse, the factors influencing symptomatic intracranial hemorrhage after mechanical thrombectomy were screened, and a nomogram risk model was constructed accordingly. The discrimination area of the model was evaluated by drawing the receiver operating characteristic curve and the area under the curve (AUC). The accuracy of the model was evaluated by the Bootstrap method and the Hosmer-Lemeshow goodness-of-fit test. The predictive efficacy of the model was evaluated by the clinical decision curve.

The overall incidence of sICH in the patients was 21.0% (38/181). The multivariate logistic regression analyses showed that preoperative intravenous thrombolysis (IVT) (OR=2.255, 95%CI: 1.016-5.004), prolonged time from onset to puncture (OR=1.002, 95%CI: 1.000-1.003), increased NIHSS score at admission (OR=1.125, 95%CI: 1.054-1.202), and failure of vascular recanalization (OR=8.849, 95%CI: 1.374-57.011) were all independent risk factors for sICH after EVT (all P<0.05). Based on these factors, a nomogram model was constructed, with an AUC of 0.768 (95%CI: 0.686-0.849). The Hosmer-Lemeshow goodness-of-fit test indicated that the model had high accuracy (P=0.636). The decision curve analysis showed that the model had good clinical practical value.

Preoperative IVT, prolonged time from onset to puncture, increased NIHSS score at admission and failure of vessel recanalization are independent risk factors for sICH after EVT in patients with AIS-LVO accompanied by large core infarction. The nomogram model established based on these factors has good predictive efficacy and clinical application value.

To compare and analyze gender differences in the characteristics of intracranial aneurysm (IA).

A retrospective study was conducted on patients diagnosed with IA who were hospitalized at the Department of Neurosurgery, Tongji Hospital, affiliated to Tongji Medical College of Huazhong University of Science and Technology, from January 2023 to December 2023. Gender differences in age, total cholesterol and triglyceride levels, presence of multiple aneurysm, aneurysm location, maximum diameter, and morphology of ruptured and unruptured aneurysm were analyzed.

A total of 866 patients (1 115 IAs) were included, comprising 289 males (368 IAs) and 577 females (747 IAs), with a male-to-female ratio of 1 ∶2. Among them, 193 patients had multiple aneurysm (63 males, 130 females), and 294 patients (294 IAs) had ruptured aneurysm (109 males, 185 females). There were no statistically significant differences in age distribution, proportion of multiple or ruptured aneurysm, IA morphology, or admission triglyceride levels between male and female patients (all P>0.05). However, the total cholesterol level was significantly higher in female patients compared to males (P<0.001). When analyzing all aneurysm collectively, the location distribution and maximum diameter differed significantly between genders (both P<0.001). Male patients had a higher proportion of IA located in the anterior communicating artery and posterior circulation, whereas female patients had a higher proportion in the internal carotid artery and posterior communicating artery. Regarding the size, male and female patients had a higher proportion of aneurysm with a maximum diameter of <5 mm. Among the 294 ruptured IAs, the location distribution and maximum diameter also showed significant gender differences (both P≤0.001). Ruptured IA in males were more frequently located in the anterior communicating artery and had a higher proportion of aneurysm with a maximum diameter greater than 10 mm. Conversely, ruptured IA in females were more common in the posterior communicating artery, with a higher proportion measuring less than 5 mm in maximum diameter.

There seem to be gender differences in the overall incidence, location, aneurysm size, and location and size of ruptured aneurysm in IA.

To explore the assessment value of the hemoglobin, albumin, lymphocytes and platelet count (HALP) score combined with the composite inflammatory indicators for the surgical efficacy in elderly patients with spontaneous intracerebral hemorrhage (sICH).

The data on 340 elderly patients with sICH admitted to the Department of Neurosurgy, Affiliated Hospital of North China University of Science and Technology from June 2022 to January 2025 were consecutively collected. Based on their survival status 30 days after onset, the patients were divided into two groups: the good efficacy group (143 cases) and poor efficacy group (mRS≥3 points) (197 cases). The HALP scores of these two groups at different time points were compared and analyzed. Univariate analyse and multivariate logistic regression analyse were used to identify the risk factors associated with poor prognosis in these patients. A nomogram visualization prediction tool was constructed and internal validation was performed, the predictive efficacy of the HALP score combined with the levels of composite inflammatory indicators for the therapeutic effect in patients with sICH was analyzed using the receiver operating characteristic (ROC) curve.

Compared with the poor efficacy group, the good efficacy group had fewer patients with alcohol history, prior stroke, or anticoagulant use, and lower rates of midline shift. This group had higher peripheral blood albumin, hemoglobin, platelet count, lymphocyte count, monocyte count, high-density lipoprotein, admission Barthel index score, Cally index, and LMR, while C-reactive protein, homocysteine, platelet distribution width, admission NIHSS score, and MHR score were lower (all P<0.05). HALP scores at different time points in sICH patients showed significant differences (Wald χ²_Group=63.92, Wald χ²_Time=1426.83, Wald χ²_Interaction=43.30; all P<0.001). The good prognosis group had significantly higher HALP scores preoperatively and at postoperative 1, 3, 5, and 7 days (all P<0.001). Multivariate logistic regression showed that higher MHR and lower postoperative 3-day HALP score, Cally index, and LMR were independent risk factors for poor 30-day efficacy in sICH (all P<0.05). ROC analysis demonstrated that combining postoperative 3-day HALP score with composite inflammatory indicators achieved the best assessment capability, with an AUC of 0.904, a sensitivity of 80.71%, and a specificity of 86.01%; all of which were superior to the fitting effects of individual markers(AUC_{postoperative 3-day HALP}=0.814, AUC_cally=0.814, AUC_MHR=0.814, AUC_MLR=0.814). A nomogram predicting short-term postoperative efficacy in sICH was developed based on postoperative 3-day HALP and composite inflammatory scores. Internal validation showed close alignment between observed and predicted outcome rates.

The postoperative 3-day HALP score, Cally index, MHR, and LMR are risk factors for poor prognosis in patients with sICH. The postoperative efficacy nomogram model established by combining these indicators had good assessment efficacy for sICH patients.

To investigate the application value of a predictive model based on anatomical and radiomic features in preoperative assessment of the accuracy of hand-related motor area localization in brain glioma patients.

Retrospectively collected the preoperative task-state functional magnetic resonance imaging (task-fMRI) data of patients with glioma admitted to the Neurosurgery Center, Beijing Tiantan Hospital, Capital Medical University based on blood oxygen level-dependent imaging. The data collected from the Beijing Neurosurgical Institute from January 2016 to June 2018 were used as the training set (125 cases) and the internal validation set (19 cases), while the data collected from the Institute of Biophysics, Chinese Academy of Sciences from July 2018 to June 2019 were used as the external validation set (46 cases). All patients completed the task-fMRI scan within 72 hours before surgery using the same type of magnetic resonance equipment and the same sequence. The shortest distance from gliomas to the hand motor area (denoted as "D" ) was calculated as an anatomical feature. Three region-of-interest (ROI) delineation strategies were applied: the complete hand motor cortex region ( "HK" ), the extended region including superior and inferior adjacent layers ( "HKUL" ), and the entire precentral gyrus region ( "PG" ), from which radiomic features were extracted basing on the above anatomical and radiological features, a random forest algorithm was developed to build a model for predicting the accuracy of hand-related motor function area localization. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) and precision-recall curves, while improvements in performance were assessed through the Delong test, continuous net reclassification improvement (NRI), and integrated discrimination improvement (IDI).

A total of 190 patients was included, with no statistically significant differences in baseline characteristics among the three datasets (all P>0.05), except for brain tumors. In the training set, internal validation set, and external validation set, the D-alone model demonstrated high predictive performance, with AUC values of 0.91, 1.00, and 0.87, respectively. In the training set, the D+ HK model achieved an AUC of 0.92, slightly outperforming the D-alone model (AUC=0.91). In the internal validation set, the D+ HK, D+ HKUL, and D+ PG models all achieved perfect predictive performance with AUCs of 1.00. In the external validation set, the D+ HK, D+ HKUL, and D+ PG models (AUCs=0.89, 0.94, and 0.88, respectively) outperformed the D-alone model (AUC=0.87). Across the three datasets, the F1 scores of the different predictive models were similar, and the Brier scores were comparable. NRI results indicated that the D+ HK and D+ HKUL models significantly improved patient reclassification ability (both P<0.05). Compared with the D-alone model, the D+ HK, D+ HKUL, or D+ PG model did not show significant improvement in IDI. Delong′s test suggested that the combined models did not significantly enhance discrimination, with their primary advantages lying in classification calibration and reclassification capability.

The predictive model combining anatomical and radiomic features can improve the accuracy of preoperative task-fMRI localization of hand-related motor cortex in glioma patients, particularly enhancing classification calibration and reclassification capability. This approach provides valuable decision support for the precise preservation of motor cortex in clinical practice.

To investigate machine learning models based on preoperative MRI for predicting isocitrate dehydrogenase (IDH) mutation in diffuse gliomas and to validate their predictive performance.

A retrospective analysis was performed on MRI data from 224 patients with diffuse gliomas (WHO grades 2-4) who were treated in the Neurosurgery Center, Beijing Tiantan Hospital, Capital Medical University between January 2021 and November 2023. MRI sequences included T1-MPRAGE, T2-FLAIR, and ADC derived from DTI. Patients were randomly divided into a training set (n=157) and a testing set (n=67) at a 7 ∶3 ratio. Tumors were manually segmented on T1-MPRAGE, T2-FLAIR, and ADC images, and radiomic features were extracted. Principal component analysis was applied for dimensionality reduction, followed by feature selection using analysis of variance and recursive feature elimination. In the training set, Logistic regression (LR), LASSO Logistic regression, and support vector machine (SVM) algorithms were used to construct IDH mutation prediction models. Using single-sequence features and combined multi-sequence features, 10-fold cross-validation was performed in both the training and testing sets. The accuracy, sensitivity, and specificity of the model were calculated based on receiver operating characteristic curves, and predictive performance was evaluated using the area under the curve (AUC), with AUC values of 0.5-0.7 indicating low accuracy, >0.7-0.9 indicating moderate accuracy, and >0.9 indicating high accuracy.

A total of 1 684 radiomic features were extracted from each of the three MRI sequences. Across the LR, LR-LASSO, and SVM, the combination of T1-MPRAGE, T2-FLAIR, and ADC features consistently yielded superior performance compared with single-sequence or two-sequence combinations. In the training set, the AUC reached 0.998 (95%CI: 0.994-1.000), 0.998 (95%CI: 0.994-1.000), and 0.996 (95%CI: 0.990-1.000) for the LR, LR-LASSO, and SVM, respectively. In the testing set, the corresponding AUC was 0.938 (95%CI: 0.870-1.000), 0.938 (95%CI: 0.867-1.000), and 0.939 (95%CI: 0.867-1.000) respectively, indicating high predictive accuracy, with the SVM model demonstrating the best overall performance.

Machine learning models based on combined T1-MPRAGE, T2-FLAIR, and ADC imaging features could effectively predict IDH mutation in diffuse gliomas.

To construct MRI-based radiomics models for predicting the characteristics of large pituitary neuroendocrine tumors (PitNETs) and to validate their clinical application efficacy.

Clinical and imaging data of 234 patients with large PitNETs (maximum tumor diameter ≥10 mm) diagnosed in the Department of Neurosurgery at the Second Affiliated Hospital of Soochow University (132 cases), Changshu Second People′s Hospital (54 cases) and Wuzhong People′s Hospital of Suzhou (48 cases) from January 2020 to December 2024 were retrospectively analyzed. Patients were randomly divided into a training set (163 cases) and a validation set (71 cases) at a 7 ∶3 ratio. According to intraoperative observations of tumor blood supply, texture, and cavernous sinus invasion, patients were divided into corresponding groups: 82 cases in the rich blood supply group and 152 cases in the average blood supply group (based on blood supply); 134 cases in the tough texture group and 100 cases in the average texture group (based on texture characteristics); 72 cases in the invasion group and 162 cases in the non-invasion group (based on cavernous sinus invasion). Clinical and MRI features were screened by univariate analysis and multivariate Logistic regression analyses. Radiomics features were processed with Z-score normalization, followed by initial screening using t-test (P<0.05) and dimensionality reduction using least absolute shrinkage and selection operator regression with ten-fold cross validation. Three machine learning algorithms, namely multivariate logistic regression (LR), K-nearest neighbors (KNN), and Gaussian naive Bayes (GNB), were used to construct imaging feature prediction models. Receiver operating characteristic (ROC) curves were plotted, and the area under the curve (AUC) was used to evaluate the predictive performance of the models (AUC>0.9 indicated excellent performance, while AUC>0.8 to 0.9 indicated good performance, and Delong test was used to compare the differences in predictive efficacy among the models. Decision curve analysis (DCA) was also employed to assess the clinical utility of the models.

Univariate analysis and multivariate Logistic regression analyses showed that increased maximum tumor diameter was an independent predictor of rich blood supply in PitNETs (OR=1.118, 95%CI: 1.070-1.167, P=0.014). Irregular tumor shape (OR=0.367, 95%CI: 0.184-0.733, P=0.005) and presence of cystic degeneration (OR=0.197, 95%CI: 0.098-0.397, P<0.001) were independent predictors of average texture in PitNETs. Increased maximum tumor diameter was an independent predictor of cavernous sinus invasion (OR=1.343, 95%CI: 1.222-1.476, P<0.001). After radiomics feature selection, 6 features were selected from the combined sequence of T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), and contrast-enhanced imaging (CEI). For cavernous sinus invasion prediction, 3 features were selected from the combined sequences of T2WI and CEI. For blood supply prediction, the LR, KNN, and GNB models all demonstrated good or excellent predictive efficacy (training set AUC: 0.932, 0.933, 0.885 respectively; validation set AUC: 0.870, 0.870, 0.866 respectively). For texture characteristics prediction, the three models also exhibited good or excellent predictive efficacy (training set AUC: 0.927, 0.933, 0.909 respectively; validation set AUC: 0.879, 0.875, 0.876 respectively). For cavernous sinus invasion prediction, the KNN model showed signs of overfitting, while the LR and GNB models both displayed excellent predictive efficacy (training set AUC: 0.935, 0.931 respectively; validation set AUC: 0.909, 0.905 respectively). DCA results indicated that all radiomics models provided good clinical net benefits.

MRI-based radiomics models constructed with machine learning can effectively predict the blood supply, texture, and cavernous sinus invasion of large PitNETs, demonstrating clinical application value.

To explore the expression characteristics of inflammatory cells and related markers in the outer membrane of chronic subdural hematoma (CSDH) at various stages and their potential roles.

Specimens were obtained from 18 patients with CSDH admitted to Neurosurgery Center, Beijing Tiantan Hospital, Capital Medical University between April and October 2021. All patients underwent burr hole drainage surgery, and the outer membrane of CSDH was intraoperatively harvested. In this study, histological staging was performed based on hematoxylin and eosin (HE) staining results according to the Nagahori classification, the degree of fibrosis was evaluated by Masson staining, and the expression of CD3 (T-cell marker), CD20 (B-cell marker), CD68 (macrophage marker), Galectin3 (neutrophil marker), and transforming growth factor β (TGF-β) and CD31 was detected by immunohistochemistry. The inflammatory cell characteristics, fibrotic changes and vascular remodeling of the outer membrane of chronic subdural hematoma were analyzed. Spearman correlation analysis was used to examine the correlations among various detection indicators.

Of the 18 patients, 4 cases were classified as stage Ⅰ, 4 cases as stage Ⅱ, 5 cases as stage Ⅲ, and 5 cases as stage Ⅳ according to CSDH outer membrane staging. Masson staining showed that collagen deposition in the outer membrane gradually increased from stage Ⅰ to Ⅳ(P<0.001). Immunohistochemical staining results showed that the expression of CD3, CD20, CD68, Galectin3, and TGF-β peaked at stage Ⅱ, whereas CD31 expression was the highest at stage Ⅲ. In addition, the expression level of TGFβ in the outer membrane was positively correlated with the expression levels of CD3, CD20 and CD68 (r values: 0.830, 0.708 and 0.943 respectively, all P<0.01), suggesting that TGF-β may be primarily produced by macrophages.

Inflammatory cells exhibit dynamic changes during the onset and progression of CSDH. Macrophages are relatively abundant during the inflammatory phage at stage Ⅱ. Neovascularization is most prominent at stage Ⅲ. The outer membrane undergoes gradual fibrosis from stage Ⅰ to Ⅳ. Macrophages and TGF-β may play pivotal role in the pathogenesis and progression of CSDH.

To investigate the expression profile of secreted frizzled-related protein 4 (SFRP4) in gliomas and its potential mechanisms in the tumor immune microenvironment.

Public data from the Chinese Glioma Genome Atlas (CGGA) cohort 693 (n=693) were extracted, and the gene expression data were standardized. Subsequently, the expression of SFRP4 in gliomas with different World Health Organization (WHO) grades, isocitrate dehydrogenase (IDH) mutation status, and 1p/19q co-deletion status was analyzed. Based on the ranking of SFRP4 gene expression levels, the top 231 samples were defined as the high-expression group and the last 231 samples as the low-expression group for Kaplan-Meier survival analysis. Based on the single-cell RNA sequencing data from CGGA 6148 (n=6 148), standardization, clustering, and cell annotation were performed. Tumor cells were classified according to the Neftel four subtypes, and the expression status of SFRP4 in each cluster was analyzed. Trajectory analysis was conducted using Monocle2 (2.26.0) to explore the temporal expression distribution of SFRP4. Spatial transcriptomic data (GSE194329) (n=1) were used to visualize and qualitatively evaluate the colocalization of SFRP4 with markers of tumors, blood vessels, immunity, and neurons. The built-in signaling pathway functional enrichment analysis in CellChat (version 1.6.1) was performed to explore the cell-cell communication network of SFRP4-highly expressing tumor cells.

In the CGGA 693 dataset, the expression level of SFRP4 in glioblastoma was higher than that in low-grade gliomas (P<0.001). The expression levels of SFRP4 in patients with IDH wild-type and non-1p/19q co-deletion were higher than those in patients with IDH mutant and 1p/19q co-deletion, respectively (both P<0.001). The overall survival was significantly shorter in the SFRP4 high-expression group than in the low-expression group (P<0.001). Single-cell transcriptomic analysis of CGGA 6148 revealed that SFRP4 was predominantly expressed in tumor cells. Among the Neftel four subtypes, SFRP4 was expressed at the highest level in neural progenitor cell-like glioma cells. In the classic classification of the Cancer Genome Atlas (TCGA), SFRP4 was highly expressed in mesenchymal-like gliomas. Spatial transcriptomic analysis showed that SFRP4 was significantly spatially colocalized with malignant markers of gliomas (such as PTPRZ1 and CX3CR1), and SFRP4-highly expressed tumor cells were distributed around blood vessels and neurons. In immune cells, SFRP4 was significantly colocalized with markers of macrophages and monocytes, but not spatially colocalized with T cell distribution. Trajectory analysis revealed that with tumor initiation and progression, the expression of SFRP4 exhibited a trend of early increase, mid-term decrease, and late increase. Moreover, the expression trend of SFRP4 was inversely correlated with that of the cell proliferation marker CCND1. Cell communication analysis demonstrated that the network centered on SFRP4-high tumor cells showed multi-directional interactions with immune cells including macrophages, oligodendrocytes, and T cells. Pathway enrichment analysis revealed significant enrichment of the complement and TGF-β pathways. Gene ontology analysis indicated that immune response-related pathways were upregulated in SFRP4-high tumor cells.

SFRP4 is upregulated in highly malignant subsets of gliomas, and its high expression is associated with poor survival prognosis. SFRP4-high tumor cells may be involved in the remodeling of the tumor microenvironment and the regulation of immune responses through the complement and TGF-β pathways, suggesting a potential association between SFRP4 and tumor immune escape.

To investigate the protective role of the deubiquitinating enzyme ovarian tumor domain-containing protease 3 (OTUD3) in cerebral ischemia-reperfusion injury (CIRI) in mice and the underlying mechanisms.

A total of 210 adult C57BL/6J mice (102 males, 108 pregnant) and 96 OTUD3 knock out (KO) C57BL/6J mice (72 males, 24 pregnant) were used. Male C57BL/6J mice were assigned to middle cerebral artery occlusion-reperfusion (MCAO/R) groups at 3, 6, 12, and 24 h and a control group (n=6 per group), wild-type (WT) control group and WT+ MCAO/R 24 h group. Male KO mice were divided into KO-OTUD3 control group and KO-OTUD3+ MCAO/R 24 h group. The MCAO/R model was established using the suture occlusion method. Primary cortical neurons were cultured from mouse embryos at embryonic day 17, and oxygen-glucose deprivation/reperfusion (OGD/R) models were established with groups at 3, 6, 12, and 24 h. Cell studies comprised WT neuron control, WT neuron+ OGD/R 24 h, KO-OTUD3 control, KO-OTUD3+ OGD/R 24 h, Vector+ OGD/R 24 h, and OTUD3 overexpression plasmid (pCMV-OTUD3) + OGD/R groups. Protein expression levels of OTUD3 and glutathione peroxidase 4 (GPX4) were detected by Western blotting (WB). Blood-brain barrier integrity was assessed by Evans blue staining, infarct volume by 2, 3, 5-triphenyltetrazolium chloride (TTC) staining, cytotoxicity by lactate dehydrogenase (LDH) release assay, cell viability by methylthiazolyl tetrazolium (MTT) assay, mitochondrial morphology by transmission electron microscopy, reactive oxygen species (ROS) levels by ROS staining, and reduced glutathione (GSH) levels using a GSH assay kit.

WB showed that OTUD3 protein expression in brain tissue or neurons was decreased in the MCAO/R and OGD/R groups at 3, 6, 12, and 24 h after reperfusion compared with the control group, with a more pronounced decrease as reperfusion time extended (all P<0.01). Evans blue and TTC staining revealed that compared with the WT+ MCAO/R 24 h group, the KO-OTUD3+ MCAO/R 24 h group exhibited a greater increase in Evans blue content in brain tissue and a significantly larger cerebral infarct volume (both P<0.05). WB, MTT and LDH assays showed that the pCMV-OTUD3+ OGD/R 24 h group had significantly higher OTUD3 expression, significantly higher neuronal viability, and significantly lower LDH release than the vector+ OGD/R 24 h group (all P<0.05). Transmission electron microscopy revealed that neurons in the ischemic penumbra of the WT+ MCAO/R 24 h and KO-OTUD3+ MCAO/R 24 h groups exhibited reduced mitochondrial volume and increased membrane rupture, showing typical ferroptosis-associated mitochondrial morphological changes. The KO-OTUD3+ MCAO/R 24 h group showed a more pronounced decrease in mitochondrial length and an increase in the number of ruptured mitochondria than the WT+ MCAO/R 24 h group (both P<0.05). ROS levels were significantly higher in the KO-OTUD3+ MCAO/R 24 h group than in the WT+ MCAO/R 24 h group (P<0.01). WB and GSH showed that GPX4 and GSH levels were decreased more markedly in the KO-OTUD3+ MCAO/R 24 h group than in the WT+ MCAO/R 24 h group, and similarly, GPX4 and GSH levels were decreased more markedly in KO-OTUD3 neurons subjected to OGD/R 24 h than in WT neurons subjected to OGD/R 24 h (all P<0.05).

OTUD3 exerts protective effects against CIRI in mouse brain tissue and neurons, and the mechanism may involve OTUD3-mediated reduction of ferroptosis.

The clinical manifestations of Rathke cyst stroke and pituitary neuroendocrine tumor stroke are highly similar, and the preoperative misdiagnosis rate is about 50%. In imaging, the characteristic manifestations of Rathke cyst stroke include fluid-fluid plane, delayed progressive enhancement of cyst wall, "black hole sign" of magnetic sensitivity-weighted imaging sequence, and the lesion is strictly located at the midline, and the pituitary stem is mostly intact. Treatment should follow the principle of individualization according to the severity of symptoms. Patients with severe neuroophthalmic damage need surgical intervention, while those without severe symptoms can choose conservative treatment to avoid hypopituitary function caused by unnecessary surgery. In December 2022 and August 2024, one patient with Rathke cyst stroke was diagnosed and treated in the Department of Neurosurgery of Xiangyang First People's Hospital affiliated to Hubei Medical College. Both patients were misdiagnosed as pituitary neuroendocrine tumor stroke by preoperative MRI, and Rathke cyst with hemorrhage was confirmed by postoperative pathological examination.

Glioma is the most common primary central nervous system tumor in clinic, and surgery is still the most important treatment at present; Accurate identification of tumor boundaries is the core key of surgical treatment. In recent years, breakthroughs have been made in the auxiliary technology of tumor boundary recognition in glioma surgery, including the application of intraoperative neuronavigation technology, intraoperative ultrasound technology (iUS), intraoperative magnetic resonance imaging technology (ioMRI), intraoperative fluorescence technology, cell endoscopy and Raman spectroscopy technology, which have significantly improved the accuracy of surgery. Intraoperative neuronavigation technology can achieve accurate targeted guidance; iUS is economical and convenient, and supports real-time dynamic scanning during surgery; ioMRI provides high-resolution real-time anatomical navigation; Fluorescence technology enables fluorescence boundary visualization through tumor metabolism specificity; Cell endoscopy realizes in-situ real-time imaging of tissue cell structure during operation; Raman spectroscopy technology accurately provides molecular level information with the advantages of label-free, non-invasive and high sensitivity; Multimodal technology fusion combined with artificial intelligence assisted diagnosis is the future development direction. Based on this, this paper reviews the clinical application progress and practical value of various technologies, and discusses their advantages and disadvantages and future development direction.

Facial nerve paralysis caused by facial nerve injury is a common neurological dysfunction in clinic, which seriously affects the quality of life of patients. Because of the complex anatomical structure of facial nerve and the low nerve regeneration efficiency, complications such as nerve fiber re-rupture and muscle atrophy are prone to occur after facial nerve repair. Therefore, the repair of facial nerve injury caused by various reasons is still an urgent problem to be solved in clinic. How to deal with the above problems is the research hotspot of facial nerve injury repair. This paper reviews the main methods and indications of surgical repair of facial nerve injury in recent years, as well as related research in the field of biomedical engineering, in order to provide new ideas for clinical treatment and basic research.