Neurosurgery最新文献

Background and objectives: The Accreditation Council for Graduate Medical Education Neurosurgery Milestones were implemented to advance competency-based training in neurosurgery; however, research on milestones in neurosurgery has been more limited, and there has been no comprehensive study on the milestone ratings and the comparability of Milestones 1.0 and 2.0. The goal of this study was to describe the levels and trends of competency ratings across Milestones 1.0 and 2.0 for neurosurgical residents in the United States over the decade of implementation. Milestones 1.0 and 2.0 cover both nontechnical competencies and critical technical skills.

Methods: We conducted a retrospective analysis of milestone assessments from 2013 to 2023 across 124 US neurosurgery residency programs, encompassing both Milestones 1.0 (2013-2018) and Milestones 2.0 (2018-2023). We used descriptive statistics to examine the distributions of milestone ratings and identify performance trends in milestone ratings across postgraduate year (PGY) levels.

Results: The highest average rating across any milestone and all PGY levels in Milestones 1.0 was professionalism (mean = 3.17, SD 1.02) whereas the average resident rating for Patient Care was the lowest (mean = 2.89, SD = 1.03). However, a shift occurred for Milestones 2.0 where Interpersonal and Communication Skills became the highest for early PGY levels and Medical Knowledge for senior residents. Patient Care remained the lowest scoring competency under both systems. Subcompetencies such as Critical Care and Brain Tumor management consistently showed high scores, whereas areas such as Surgical Treatment of Epilepsy, Pain and Peripheral Nerves, and Pediatric Neurosurgery demonstrated lower scores.

Conclusion: A balance between nontechnical competencies and critical technical skills is necessary to ensure comprehensive neurosurgical training. Established benchmarks can enhance the utility of milestone data and support the development of well-rounded, competent neurosurgeons.

Background and objectives: Identifying predictors of shunt failure in acute care settings is challenging. Previous studies largely rely on traditional statistical analyses to evaluate risk factors associated with shunt failure. However, these methods have not been adapted for emergency department (ED) presentations, where machine learning (ML) techniques may offer enhanced predictive power. Thus, we explored the application of ML and artificial intelligence in predicting necessity for shunt revision in pediatric ED visits.

Methods: A retrospective analysis was conducted on 1167 pediatric neurosurgical consults for suspected shunt malfunction from 2017 to 2022. We collected 24 historical, clinical, and radiographic variables. Various ML techniques were used, including support vector machine, K-nearest neighbors, random forest, an artificial neural network (ANN), and a proprietary large language model named ShuntGPT (SGPT), which classifies the need for shunt revision.

Results: Among 1167 ED consults for shunt malfunction, 285 patients (24.4%) underwent revision. Multivariate analysis identified increased ventricle size, abnormal shunt series, lethargy, altered mental status, and bradycardia as predictors of need for revision. The best-performing ML models included ANN (accuracy 84%, area under the curve [AUC] 0.88, 71% sensitivity, 88% specificity) and SGPT (accuracy 87%, precision 0.80, recall 0.64, AUC 0.927). Traditional classifiers (support vector machine, K-nearest neighbors, random forest) achieved AUCs of 0.81 to 0.86 with varying trade-offs in sensitivity and precision. With imaging results removed from training data, the performance of all models suffered, however, SGPT retained a high level of discrimination (AUC 0.84). SGPT continued to show high accuracy (83%) on a separate validation cohort.

Conclusion: Advanced ML models generally outperformed traditional statistical analyses, albeit with concerns about overfitting due to the extensive variable set. SGPT, in particular, showed superior performance, likely due to its capacity to interpret nuanced text. This model represents a promising tool to enhance decision-making in pediatric acute care settings regarding shunt malfunctions.

Background and objectives: Access to specialty surgical care is growing in many low-income countries, but it remains unclear how hospital workforces can leverage technology to manage large numbers of increasingly complex patients. Large language models (LLMs) may be helpful for this type of clinical decision support, but their real-world performance and safety remain uncertain. The objective of this study was to evaluate feasibility, usability, and potential benefits and risks of an LLM-based assistant for postoperative neurosurgical care in the Gambia.

Methods: A prospective, single-arm implementation study was conducted at the Edward Francis Small Teaching Hospital. A convenience sample of 4 medical officers (MOs) and 5 nurses assigned to the neurosurgical service participated. A prompted GPT-4o Turbo was deployed on OpenAI Pro accounts to support performance. Usability, helpfulness, and safety were the primary outcomes. Cost-effectiveness was a secondary outcome.

Results: Participants completed 75 LLM-assisted interactions on 9 postoperative neurosurgery patients. Usability metrics indicated a moderately high cognitive workload, marginally acceptable usability of the LLM system, and high perceived ease of use. Management plan quality improved in 45 of 75 mock rounds interactions (60%), with a mean improvement of 8.5% (P < .001) on mock rounds scoring rubrics. The improvement was greater for MOs (21.0% change) than nurses (6.5% change). In hypothetical case dilemmas, MO plan accuracy improved by 22.7% (P = .001), and critical errors declined from 33.3% to 0%. Fourteen care changes for 9 patients were attributed to LLM suggestions, including 6 that potentially prevented major morbidity. No unsafe outputs were detected. Exploratory cost analysis suggested potential savings from clinical care changes exceeded the labor costs involved in LLM use.

Conclusion: LLM use was associated with improved plan quality without observed safety concerns, while also prompting clinically meaningful care changes. Larger, controlled studies are needed to determine generalizability, durability of benefit, and patient-centered outcomes.

Background and objectives: Supratentorial herniation is traditionally described as a mechanical shift of brain structures due to elevated intracranial pressure. However, this model inadequately explains the stepwise physiological decline observed in patients with progressive brainstem dysfunction. We aimed to propose and support a novel conceptual model of herniation based on a vertically oriented compaction gradient and the emergence of a "demarcation line of tissue compaction" as a physiological boundary between viable and irreversibly failing neural tissue.

Methods: A prospective observational study was conducted between 2022 and 2025, involving 287 patients with traumatic supratentorial brain herniation admitted to the neurosurgical departments of 3 urban hospitals. Clinical staging was based on consciousness, pupillary reactivity, brainstem reflexes, motor responses, respiratory pattern, and the Kerdo Vegetative Index. Postmortem morphological analysis was conducted in 129 cases to determine the rostrocaudal extent of brainstem injury.

Results: Five distinct stages of herniation were identified, each correlated with specific clinical features and anatomic damage. A progressive downward pattern of neurological deterioration and structural destruction was observed from the diencephalon to the medulla. The progression was often preceded by a physiological cluster-Hyper-H triad-manifesting as hyperthermia, hypertonia, and hormonal dysregulation. Clinical signs and autonomic shifts consistently mirrored this descent. The findings support the presence of a transitional frontier-termed the demarcation line of compaction-separating salvageable function from collapse.

Conclusion: Supratentorial herniation represents not only spatial displacement but also a structured physiological degradation. The compaction model provides a dynamic framework for understanding herniation progression and may assist in bedside assessment, staging, and early intervention. The Hyper-H triad may serve as an early physiological indicator of impending herniation. Further prospective and imaging-based studies are warranted to validate this approach.

Background and objectives: Lung cancer is the most common primary neoplasm that metastasizes to the brain. Although stereotactic radiosurgery (SRS) using American Society for Radiation Oncology guideline-suggested doses improves local control, it may increase adverse radiation events (AREs). This study evaluates whether reduced SRS doses can achieve comparable local control while minimizing toxicity in patients with brain metastases from non-small cell lung cancer (NSCLC) primary receiving concurrent systemic therapy.

Methods: We conducted a retrospective study of 264 patients with 1390 metastases from lung primary treated with SRS between December 2015 and January 2025. Of these, 82 patients with 433 metastases had NSCLC and received concurrent systemic therapy, defined as immunotherapy or targeted therapy within 30 days of SRS. Propensity score matching identified 38 patients with 93 metastases in the higher-dose (HD) group (20-24 Gy for <2 cm, 18 Gy for ≥2 to <3 cm) and 42 patients with 93 metastases in the reduced-dose (RD) group (<20 Gy for <2 cm, <18 Gy for ≥2 to <3 cm).

Results: Radiographic AREs, including perilesional edema and radiation necrosis, were significantly more common in the HD group than in the RD group (23.7% vs 10.8%, P = .020). Local control was noninferior in the RD group (94.6%) compared with the HD group (90.3%, P = .400). Cumulative progression at 1, 3, and 5 years remained comparable between the RD (13%, 15%, and 15%, respectively) and HD (7.2%, 10%, and 10%, respectively) groups (P = .500).

Conclusion: In the contemporary era, RD SRS delivered concurrently with immunotherapy or targeted therapy may lower AREs without compromising local control to treat NSCLC brain metastases.

Background and objectives: Venous sinus stenting (VSS) has emerged as an alternative treatment for selected idiopathic intracranial hypertension (IIH) cases, but its comparative effectiveness and safety vs ventricular shunting (VS) remain uncertain. Given the heterogeneity of IIH, we hypothesized that outcomes differ by patient phenotype defined by obesity and visual disturbance at diagnosis.

Methods: Adults with IIH who underwent VSS or VS shunting were identified using TriNetX. Patients were stratified into 4 prespecified phenotypes: nonobese-visual-disturbance (NO-V), nonobese-no-visual-disturbance (NO-NV), obese-visual-disturbance (OV), and obese-no-visual-disturbance (ONV). Within each phenotype, 1:1 propensity score matching balanced baseline demographics, comorbidities, and laboratory parameters. Primary outcomes were persistent headache, papilledema, and visual disturbance within 24 months. Secondary outcomes included retreatment, head computed tomography use, and emergency department visits.

Results: After matching (NO-V = 69, NO-NV = 186, OV = 305, ONV = 440 per arm), cohorts were well balanced. In NO-NV, VS was associated with higher rates of headache (risk difference [RD]: 0.11, P = .049), visual disturbances (RD: 0.11, P = .010), and emergency department visits (RD: 0.15, P = .006). NO-V showed no primary outcome differences, although vestibular disorders were more common with shunts (RD: 0.15, P = .008). In OV, VS shunting was linked to higher rates of visual deficits (RD: 0.11, P = .045). Conversely, in ONV, shunting was associated with lower rates of papilledema (RD: -0.6, P = .024) and pulsatile tinnitus (RD: -0.5, P = .004) but increased headaches (RD: 0.12, P < .001). Across phenotypes, shunted patients had greater head computed tomography use and retreatment rates, while VSS patients more often required ongoing medical therapy.

Conclusion: The comparative effectiveness of VSS and VS in IIH varies by phenotype. VSS appears preferable in nonobese patients without visual symptoms, whereas VS may better control papilledema in obese patients without visual disturbance. Both treatments show similar efficacy in patients with baseline visual impairment. These findings underscore the need for phenotype-guided, prospective trials.

Background and objectives: Meningioma consistency is linked to important clinical and surgical outcomes, and preoperative knowledge of consistency may have significant implications for operative planning and decision making. This study aimed to determine whether intraoperative meningioma consistency is associated with preoperative T2-weighted imaging (T2WI) and diffusion-weighted imaging.

Methods: Tumor consistency was prospectively assigned in 188 meningioma patients at the time of surgery over a twelve-year period (2012-2024) using a 5-point scale with validated interuser reliability. Signal intensity was quantified as the intensity ratio between region of interest in the meningioma and the middle cerebellar peduncle using a previously validated methodology (tumor/cerebellar peduncle T2WI intensity [TCTI] ratios). Mean and maximum signal intensity were measured using T2 and diffusion-weighted MRI sequences.

Results: When using mean T2WI measurements, TCTI ratios were lower for grade 4 + 5 tumors (1.41 ± 0.31, 95% CI: -0.418 to -0.182, P < .001) compared with consistency grade 3 tumors (1.71 ± 0.32, reference group) and grade 1 + 2 tumors (1.98 ± 0.36, 95% CI: -0.453 to -0.074, P < .004). A similar trend was observed using max region of interest measurements. No association was found between consistency and diffusion-weighted imaging. TCTI ratios using mean (odds ratio: 0.031, 95% CI: 0.009 to 0.101, P < .001) and max (odds ratio: 0.027, 95% CI: 0.007 to 0.097, P < .001) measurements were independent predictors of consistency when adjusting for covariates. In an receiver operating characteristic analysis comparing T2 mean ratios for grade 1 + 2 to 4 + 5 tumors, the area under the curve was 0.868 and precision-recall area under the curve was 0.960. A TCTI ratio cutoff of 1.536 was 75.0% sensitive and 100% specific in differentiating grade 1 + 2 and grade 4 + 5 tumors using T2WI alone.

Conclusion: In meningiomas, the TCTI ratio is associated with intraoperative tumor consistency, with higher signal intensity ratios predicting softer meningioma consistency. Application in this manner may augment surgical planning and improve preoperative patient counseling.

Background and objectives: Clinical risk factors for seizure presentation in meningioma patients have been reported, but molecular correlates of seizures in meningioma remain unexplored.

Methods: We assessed a total of 144 primary meningiomas for clinical data and performed bulk RNA sequencing (n = 141), whole-exome sequencing (n = 35), targeted DNA sequencing (n = 67), and DNA methylation analysis (n = 62).

Results: Clinically, seizure presentation was associated with higher rates of emergency room presentation (P = .005), altered consciousness (P = .014), and fewer headaches (P = .015). Radiographically, seizure presentation correlated with higher rates of cerebral edema (P = .005), nonhomogenous enhancement (P = .001), and intratumoral calcification (P = .034). Pathologically, seizure presenting tumors had higher molecular immunology borstel 1 index (P = .001) and cellular atypia in World Health Organization grade 1 tumors (P = .023) but were not associated with World Health Organization grade 2 status (P = .059). There was no difference in rates of canonical mutations between groups. Chromosome 14q loss was enriched in the seizure group (P = .004). Seizure presentation was associated with the Molecular Group C (MenG C) (P = .023). Transcriptomic analysis revealed downregulation of GABAergic signaling, synaptic pathways, and neurotransmitter receptors in meningiomas causing seizure. Dysregulated genes included both MenG C-specific (GABRG1, GRIA1, DKK2) and seizure-specific/MenG-independent (GABRA3, GABRQ, GRIK3, NPY1R) genes.

Conclusion: Here, we show that seizure causing meningiomas demonstrate distinct clinical, radiographic, chromosomal, and transcriptional features. Novel associations with seizure presentation in meningioma include MenG C status, chromosome 14q loss, and seizure-specific dysregulation of neurotransmitter receptor genes.