Neurosurgical focus最新文献

{"title":"Introduction. Surgical management of infiltrative gliomas.","authors":"Frederick F Lang, Jeffrey S Weinberg, Chibawanye I Ene, Victoria E Clark, Rasheed Zakaria, Wajd N Al-Holou, Linton Evans, Matthew J Shepard, Richard G Everson","doi":"10.3171/2025.5.FOCUS24675","DOIUrl":"https://doi.org/10.3171/2025.5.FOCUS24675","url":null,"abstract":"","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 2","pages":"E1"},"PeriodicalIF":3.0,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144765095","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Noninvasive mapping of visuospatial attention by navigated repetitive transcranial magnetic stimulation in patients with parietal lobe tumors.","authors":"Vicki M Butenschoen, Franziska Hausler, Axel Schröder, Bernhard Meyer, Sandro M Krieg","doi":"10.3171/2025.5.FOCUS25255","DOIUrl":"https://doi.org/10.3171/2025.5.FOCUS25255","url":null,"abstract":"<p><strong>Objective: </strong>Visuospatial neglect corresponds to a burdening cognitive deficit with reduced space attention and disturbed stimuli detection of the contralateral side. Unilateral strokes, tumor lesions, or intracerebral hemorrhage may cause it. Identifying specific areas responsible for the onset of visuospatial neglect has proven difficult. The authors hereby aimed to detect neglect-positive areas in patients with parietal gliomas undergoing tumor resection through navigated repetitive transcranial magnetic stimulation (nrTMS).</p><p><strong>Methods: </strong>The authors performed a monocentric prospective study that included patients with suspected parietal lobe gliomas. After obtaining patient consent, time-locked nrTMS was performed for neglect testing using the Landmark Task and grayscale test on 52 predefined cortical spots before and after surgery. Errors were categorized as leftward/rightward errors or deviations and no response errors. Additionally, patients performed two paper-and-pencil tests to evaluate clinical neglect.</p><p><strong>Results: </strong>Nineteen patients were enrolled in the study. Ten patients had a glioma, 8 had brain metastases, and 1 patient had a meningioma. Error rates and leftward deviations were significantly higher in the right hemisphere. The supramarginal and angular gyrus and the superior parietal and occipital areas seemed especially important. After surgery, errors increased substantially in the right hemisphere. The Landmark Task and grayscale test showed high sensitivity (100%) and a negative predictive value (100%).</p><p><strong>Conclusions: </strong>Visuospatial neglect can be evoked reliably by nrTMS in patients with parietal tumors. The study showed promising results for the intraoperative use of nrTMS visuospatial neglect mapping.</p>","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 2","pages":"E6"},"PeriodicalIF":3.0,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144765099","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Transsylvian versus transcortical approach to insular glioma: analysis of the extent of resection and postoperative neurological complications in propensity score-matched comparative patient cohorts.","authors":"Kuntal Kanti Das, Sudhakar Madheshiya, Deepak Khatri, Prabhakar Mishra, Kamlesh Singh Bhaisora, Arun Kumar Srivastava, Awadhesh Kumar Jaiswal","doi":"10.3171/2025.5.FOCUS25339","DOIUrl":"https://doi.org/10.3171/2025.5.FOCUS25339","url":null,"abstract":"<p><strong>Objective: </strong>This study aimed to analyze the comparative tumor resection rates and complication profiles of the transsylvian (TS) and transcortical (TC) approaches to the insular glioma (IG) and emphasize the concept of onco-microneurosurgery as a key to surgical success in these difficult areas.</p><p><strong>Methods: </strong>A retrospective analysis of a single surgeon's prospectively maintained data of surgically resected, newly diagnosed IGs in adult patients (≥ 18 years old) was conducted. Propensity score matching was performed with a tolerance limit of 0.05 for comparison of the TS and TC cohorts. The extent of resection (EOR) was categorized with 90% resection as a cutoff. Neurological complications persisting beyond 3 months were considered permanent complications. These two variables were combined to derive a Composite Postoperative Outcome Index (CPOI) and graded as 0, 1a, 1b, or 2.</p><p><strong>Results: </strong>Fifty-two patients (male-to-female ratio of 2.25:1) were studied, with 26 patients in each group. Radical tumor resection (≥ 90%) was obtained in 77% patients (n = 40), with transient and permanent neurological complication rates of 46.2% (n = 24) and 15.4% (n = 8), respectively. A significantly higher rate of maximal safe resection (CPOI grade 0) was obtained using a TS approach for the entire TS cohort (p = 0.008), as well as subgroups of non-giant segmental IGs (p = 0.011) and those with specific Berger-Sanai zone II involvement (p = 0.01). The TC approach was found to be significantly safer in giant IGs when a subtotal resection was performed (p = 0.03). Permanent neurological complications with ≥ 90% EOR (CPOI grade 1b) were significantly higher in the TC group (p = 0.009), including non-giant segmental IGs (p = 0.001) and those specifically involving Berger-Sanai zone II (p = 0.01) of the insula. Long-term functional status and disease progression were similar in both groups.</p><p><strong>Conclusions: </strong>These results suggest the continued role of the TS approach in IG resection in the contemporary era. Irrespective of the approach, the key variable appears to be a meticulous microsurgical technique, supplemented by the available adjuncts, in the preservation of perforator arteries and subcortical circuitry. Thus, an optimally designed, individual institution-tailored hybrid onco-microneurosurgical approach is the most pragmatic approach to IGs.</p>","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 2","pages":"E8"},"PeriodicalIF":3.0,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144765103","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Management of glioma-associated epilepsy in the molecular era: a review of the literature and an institutional experience.","authors":"Poojan D Shukla, Minh P Nguyen, Anthony T Lee, Edward F Chang, Jacob S Young","doi":"10.3171/2025.5.FOCUS25334","DOIUrl":"https://doi.org/10.3171/2025.5.FOCUS25334","url":null,"abstract":"<p><p>Seizures following resection for glioma can significantly impact patient quality of life. Resection and antiseizure medications have been first-line treatments for glioma-associated epilepsy since the survival benefit of maximizing the extent of resection was established. Given recent advances in tumor molecular profiling and neuron-glioma circuit interactions, should the management of tumor-associated epilepsy change? Here the authors present a literature review of the current state of the surgical and medical management of postoperative seizures in patients with glioma, summarize key findings from investigations of the molecular processes governing tumor-associated seizures, and provide a retrospective review correlating tumor mutational profiles obtained from next generation sequencing with seizure history in patients from a single institution. This paradigm of comparing clinical seizure outcomes and tumor genetics may broaden the understanding of glioma-associated epilepsy and prognostic factors, potentially leading to new therapeutic strategies.</p>","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 2","pages":"E11"},"PeriodicalIF":3.0,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144765097","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MRI-based habitat radiomics for preoperatively predicting IDH status in gliomas.","authors":"Wan-Yi Zheng, Ri-Hui Yang, Zhi-Fang Wan, Hui Li, Chao Ma, Qun-Hui Ouyang, Si Li, Ke-Jian Wang, Gui-Hua Jiang, Ping Liu","doi":"10.3171/2025.5.FOCUS25135","DOIUrl":"https://doi.org/10.3171/2025.5.FOCUS25135","url":null,"abstract":"<p><strong>Objective: </strong>The intratumoral heterogeneous vascular permeability and cell density of gliomas are associated with IDH mutation status. Therefore, the authors aimed to construct vascular-cellular habitats based on MRI to investigate their correlation with IDH status.</p><p><strong>Methods: </strong>In this retrospective analysis, 165 patients with pathologically confirmed glioma who underwent preoperative contrast-enhanced T1-weighted imaging and diffusion-weighted imaging (DWI) at three hospitals were included. Four spatial habitats (subregions) based on contrast-enhanced T1-weighted and DWI-derived apparent diffusion coefficient (ADC) images were defined using K-means voxel-wise clustering. The sensitive habitat of IDH mutation was identified and radiomic features were extracted and screened from the whole tumor and the four habitats. Logistic regression classifiers were used to construct predictive models for IDH mutation.</p><p><strong>Results: </strong>Of the included patients, 109 (mean age 50.78 years) were assigned to the training set and 56 (mean age 48.21 years) to the external test set. The high contrast enhancement (CE)-high ADC subregion was determined as the sensitive habitat. The four habitats model achieved an area under the receiver operating characteristic curve (AUC) of 0.716 (95% CI 0.553-0.879) in the external test set, indicating better performance than that of the traditional whole tumor model (AUC 0.619, 95% CI 0.446-0.792). Model performance was further improved when focusing on the sensitive habitat, for which the external test set AUC was 0.817 (95% CI 0.676-0.958).</p><p><strong>Conclusions: </strong>MRI habitat analysis based on contrast-enhanced T1-weighted and DWI sequences had high prediction capabilities for glioma IDH mutation status, which could be used to refine individualized treatment regimens for patients with glioma.</p>","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 2","pages":"E4"},"PeriodicalIF":3.0,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144765098","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The effect of tumor growth kinetics on survival and time toxicity after resection of recurrent glioblastoma.","authors":"Adeline L Fecker, Cooper Stateler, Molly Joyce, Sidharth Sengupta, Hanna E Minns, Joseph G Nugent, Jordan L Smith, Seunggu Jude Han, Matthew D Wood, Ahmed M Raslan, Ramon F Barajas, Stephen G Bowden","doi":"10.3171/2025.5.FOCUS25331","DOIUrl":"10.3171/2025.5.FOCUS25331","url":null,"abstract":"<p><strong>Objective: </strong>Reoperation for recurrent glioblastoma (GBM) remains controversial but is commonly offered to maximize patient quality of life (QOL) and functional status. While there is mixed evidence that waiting for surgery in primary GBM does not affect survival, this has not been studied in the context of recurrent disease, which is more aggressive. The aim of this study was to assess how recurrent tumor kinetics (i.e., tumor growth relative to the time to resection) affect survival and QOL, as measured by time toxicity, after repeat resection in patients with recurrent GBM.</p><p><strong>Methods: </strong>A prospectively collected database was queried for patients with first-time recurrent IDH-wildtype GBM, with progression confirmed by the modified Response Assessment in Neuro-Oncology criteria, from 2012 to 2022. Only patients who were recommended surgery as the first-line treatment for disease progression were included. Recursive partitioning analysis was used to automatically detect growth rate and time to repeat resection (TTRR) thresholds that affected repeat resection survival (RRS). Time toxicity was the percentage of days of medical contact for the TTRR and RRS.</p><p><strong>Results: </strong>Seventy-three patients were included in the analysis. The median TTRR of the overall cohort was 27.2 days (range 1-90 days), with a mean TTRR time toxicity of 25.2% (SD 29.2%). The median RRS was 270 days (range 23-1495 days), with a mean RRS time toxicity of 19.3% (SD 17.7%). Patients with tumor growth of 0.08 cm3 per day or faster (faster-growth group) had significantly worse RRS (p = 0.004) and time toxicity (p = 0.016). Patients who underwent repeat resection ≥ 43 days (longer TTRR group) after confirmed progression had significantly worse survival (p = 0.015) and the time toxicity increased after surgery (11.6% to 25.1%, p = 0.014). While there was no significant survival difference between the faster tumor growth and longer TTRR group compared with the faster growth and shorter TTRR group (p = 0.20), the group with faster tumor growth but shorter TTRR had better survival (median 179 days vs 81 days).</p><p><strong>Conclusions: </strong>Faster-growing GBM recurrence was associated with worse survival and time toxicity. Patients with a shorter time to surgery had improved survival compared with those with longer wait times, and this effect partially attenuated the poor risk of fast tumor growth. Therefore, tumor kinetics in recurrent GBM appear to be an important consideration for patient prognosis and QOL after surgery.</p>","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 2","pages":"E3"},"PeriodicalIF":3.0,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144765100","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A multiregional multimodal machine learning model for predicting outcome of surgery for symptomatic hemorrhagic brainstem cavernous malformations.","authors":"Xuchen Dong, Haohuai Gui, Kai Quan, Zongze Li, Ying Xiao, Jiaxi Zhou, Yuchuan Zhao, Dongdong Wang, Mingjian Liu, Haojing Duan, Shaoxuan Yang, Xiaolei Lin, Jun Dong, Lin Wang, Yu Ma, Wei Zhu","doi":"10.3171/2025.4.FOCUS24778","DOIUrl":"https://doi.org/10.3171/2025.4.FOCUS24778","url":null,"abstract":"<p><strong>Objective: </strong>Given that resection of brainstem cavernous malformations (BSCMs) ends hemorrhaging but carries a high risk of neurological deficits, it is necessary to develop and validate a model predicting surgical outcomes. This study aimed to construct a BSCM surgery outcome prediction model based on clinical characteristics and T2-weighted MRI-based radiomics.</p><p><strong>Methods: </strong>Two separate cohorts of patients undergoing BSCM resection were included as discovery and validation sets. Patient characteristics and imaging data were analyzed. An unfavorable outcome was defined as a modified Rankin Scale score > 2 at the 12-month follow-up. Image features were extracted from regions of interest within lesions and adjacent brainstem. A nomogram was constructed using the risk score from the optimal model.</p><p><strong>Results: </strong>The discovery and validation sets comprised 218 and 49 patients, respectively (mean age 40 ± 14 years, 127 females); 63 patients in the discovery set and 35 in the validation set had an unfavorable outcome. The eXtreme Gradient Boosting imaging model with selected radiomics features achieved the best performance (area under the receiver operating characteristic curve [AUC] 0.82). Patients were stratified into high- and low-risk groups based on risk scores computed from this model (optimal cutoff 0.37). The final integrative multimodal prognostic model attained an AUC of 0.90, surpassing both the imaging and clinical models alone.</p><p><strong>Conclusions: </strong>Inclusion of BSCM and brainstem subregion imaging data in machine learning models yielded significant predictive capability for unfavorable postoperative outcomes. The integration of specific clinical features enhanced prediction accuracy.</p>","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 1","pages":"E7"},"PeriodicalIF":3.3,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144541555","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep learning-based clinical decision support system for intracerebral hemorrhage: an imaging-based AI-driven framework for automated hematoma segmentation and trajectory planning.","authors":"Zhichao Gan, Xinghua Xu, Fangye Li, Ron Kikinis, Jiashu Zhang, Xiaolei Chen","doi":"10.3171/2025.5.FOCUS25246","DOIUrl":"https://doi.org/10.3171/2025.5.FOCUS25246","url":null,"abstract":"<p><strong>Objective: </strong>Intracerebral hemorrhage (ICH) remains a critical neurosurgical emergency with high mortality and long-term disability. Despite advancements in minimally invasive techniques, procedural precision remains limited by hematoma complexity and resource disparities, particularly in underserved regions where 68% of global ICH cases occur. Therefore, the authors aimed to introduce a deep learning-based decision support and planning system to democratize surgical planning and reduce operator dependence.</p><p><strong>Methods: </strong>A retrospective cohort of 347 patients (31,024 CT slices) from a single hospital (March 2016-June 2024) was analyzed. The framework integrated nnU-Net-based hematoma and skull segmentation, CT reorientation via ocular landmarks (mean angular correction 20.4° [SD 8.7°]), safety zone delineation with dual anatomical corridors, and trajectory optimization prioritizing maximum hematoma traversal and critical structure avoidance. A validated scoring system was implemented for risk stratification.</p><p><strong>Results: </strong>With the artificial intelligence (AI)-driven system, the automated segmentation accuracy reached clinical-grade performance (Dice similarity coefficient 0.90 [SD 0.14] for hematoma and 0.99 [SD 0.035] for skull), with strong interrater reliability (intraclass correlation coefficient 0.91). For trajectory planning of supratentorial hematomas, the system achieved a low-risk trajectory in 80.8% (252/312) and a moderate-risk trajectory in 15.4% (48/312) of patients, while replanning was required due to high-risk designations in 3.8% of patients (12/312).</p><p><strong>Conclusions: </strong>This AI-driven system demonstrated robust efficacy for supratentorial ICH, addressing 60% of prevalent hemorrhage subtypes. While limitations remain in infratentorial hematomas, this novel automated hematoma segmentation and surgical planning system could be helpful in assisting less-experienced neurosurgeons with limited resources in primary healthcare settings.</p>","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 1","pages":"E5"},"PeriodicalIF":3.3,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144541558","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine learning approaches for predicting prolonged hospital length of stay after lumbar fusion surgery in patients aged 75 years and older: a retrospective cohort study based on comprehensive geriatric assessment.","authors":"Qijun Wang, Shuaikang Wang, Peng Wang, Shibao Lu","doi":"10.3171/2025.4.FOCUS24614","DOIUrl":"10.3171/2025.4.FOCUS24614","url":null,"abstract":"<p><strong>Objective: </strong>Postoperative recovery following lumbar fusion surgery in patients aged 75 years and older often requires a prolonged length of stay (PLOS) in the hospital. Accurately predicting the risk of PLOS and assessing its risk factors for preoperative optimization are crucial to guide clinical decision-making. The aim of this study was to select the risk factors for PLOS and develop a machine learning (ML) model to estimate the likelihood of PLOS based on comprehensive geriatric assessment (CGA) domains in older patients undergoing lumbar fusion surgery.</p><p><strong>Methods: </strong>An observational cohort of 242 patients aged ≥ 75 years (median age 80 years) undergoing lumbar fusion surgery at a single center from March 2019 to December 2021 was retrospectively reviewed. Predictor variables consisted of clinical characteristics, CGA variables, and intraoperative variables. The primary outcome was PLOS, defined as a hospital LOS above the 75th percentile in the overall study population. Patients were randomly divided into two groups (7:3) for model training and validation. Ensemble ML algorithms were used to select the significant variables associated with PLOS, and 9 ML models were used to develop predictive models. The Shapley Additive Explanations (SHAP) method was used for model interpretation and feature importance ranking.</p><p><strong>Results: </strong>Three ensemble ML algorithms selected 9 CGA and clinical variables as influential factors of PLOS. The random forest (RF) model had the best predictive performance among the models evaluated, with an area under the receiver operating characteristic curve of 0.822 (95% CI 0.727-0.917) and F1-score of 0.571. SHAP values indicated that the duration of surgery, the number of fusion levels, and age were the most important predictors, while the Fried frailty phenotype was the most important CGA variable for PLOS. The RF model and its SHAP interpretations were deployed online for clinical utility.</p><p><strong>Conclusions: </strong>This ML model could facilitate individual risk prediction and risk factor identification for PLOS in older patients undergoing lumbar fusion surgery, with the potential to improve preoperative optimization.</p>","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 1","pages":"E16"},"PeriodicalIF":3.3,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144541492","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel deep learning system for automated diagnosis and grading of lumbar spinal stenosis based on spine MRI: model development and validation.","authors":"Tianyi Wang, Aobo Wang, Yiling Zhang, Xingyu Liu, Ning Fan, Shuo Yuan, Peng Du, Qichao Wu, Ruiyuan Chen, Yu Xi, Zhao Gu, Qi Fei, Lei Zang","doi":"10.3171/2025.4.FOCUS24670","DOIUrl":"https://doi.org/10.3171/2025.4.FOCUS24670","url":null,"abstract":"<p><strong>Objective: </strong>The study aimed to develop a single-stage deep learning (DL) screening system for automated binary and multiclass grading of lumbar central stenosis (LCS), lateral recess stenosis (LRS), and lumbar foraminal stenosis (LFS).</p><p><strong>Methods: </strong>Consecutive inpatients who underwent lumbar MRI at our center were retrospectively reviewed for the internal dataset. Axial and sagittal lumbar MRI scans were collected. Based on a new MRI diagnostic criterion, all MRI studies were labeled by two spine specialists and calibrated by a third spine specialist to serve as reference standard. Furthermore, two spine clinicians labeled all MRI studies independently to compare interobserver reliability with the DL model. Samples were assigned into training, validation, and test sets at a proportion of 8:1:1. Additional patients from another center were enrolled as the external test dataset. A modified single-stage YOLOv5 network was designed for simultaneous detection of regions of interest (ROIs) and grading of LCS, LRS, and LFS. Quantitative evaluation metrics of exactitude and reliability for the model were computed.</p><p><strong>Results: </strong>In total, 420 and 50 patients were enrolled in the internal and external datasets. High recalls of 97.4%-99.8% were achieved for ROI detection of lumbar spinal stenosis (LSS). The system revealed multigrade area under curve (AUC) values of 0.93-0.97 in the internal test set and 0.85-0.94 in the external test set for LCS, LRS, and LFS. In binary grading, the DL model achieved high sensitivities of 0.97 for LCS, 0.98 for LRS, and 0.96 for LFS, slightly better than those achieved by spine clinicians in the internal test set. In the external test set, the binary sensitivities were 0.98 for LCS, 0.96 for LRS, and 0.95 for LFS. For reliability assessment, the kappa coefficients between the DL model and reference standard were 0.92, 0.88, and 0.91 for LCS, LRS, and LFS, respectively, slightly higher than those evaluated by nonexpert spine clinicians.</p><p><strong>Conclusions: </strong>The authors designed a novel DL system that demonstrated promising performance, especially in sensitivity, for automated diagnosis and grading of different types of lumbar spinal stenosis using spine MRI. The reliability of the system was better than that of spine surgeons. The authors' system may serve as a triage tool for LSS to reduce misdiagnosis and optimize routine processes in clinical work.</p>","PeriodicalId":19187,"journal":{"name":"Neurosurgical focus","volume":"59 1","pages":"E6"},"PeriodicalIF":3.3,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144541556","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}