Barriers to neurosurgery for medical students: a national study focused on the intersectionality of gender and race.

OBJECTIVE: Despite 51.2% of medical school graduates being female, only 29.8% of neurosurgery residency applicants are female. Furthermore, only 12.6% of neurosurgery applicants identify as underrepresented in medicine (URM). Evaluating the entry barriers for female and URM students is crucial in promoting the equity and diversity of the neurosurgical workforce. The objective of this study was to evaluate barriers to neurosurgery for medical students while considering the interaction between gender and race. METHODS: A Qualtrics survey was distributed widely to US medical students, assessing 14 factors of hesitancy toward neurosurgery. Likert scale responses, representing statement agreeability, converted to numeric values on a 7-point scale were analyzed by Mann-Whitney U-test and ANOVA comparisons with Bonferroni correction. RESULTS: Of 540 respondents, 68.7% were female and 22.6% were URM. There were 22.6% male non-URM, 7.4% male URM, 53.5% female non-URM, and 15.2% female URM respondents. The predominant reasons for hesitancy toward neurosurgery included work/life integration, length of training, competitiveness of residency position, and perceived malignancy of the field. Females were more hesitant toward neurosurgery due to maternity/paternity needs (p = 0.005), the absence of seeing people like them in the field (p < 0.001), and opportunities to pursue health equity work (p < 0.001). Females were more likely to have difficulties finding a mentor in neurosurgery who represented their identities (p = 0.017). URM students were more hesitant toward neurosurgery due to not seeing people like them in the field (p < 0.001). Subanalysis revealed that when students were stratified by both gender and URM status, there were more reasons for hesitancy toward neurosurgery that had significant differences between groups (male URM, male non-URM, female URM, and female non-URM students), suggesting the importance of intersectionality in this analysis. CONCLUSIONS: The authors highlight the implications of gender and racial diversity in the neurosurgical workforce on medical student interest and recruitment. Their findings suggest the importance of actively working to address these barriers, including 1) maternity/paternity policy reevaluation, standardization, and dissemination; and 2) actively providing resources for the creation of mentorship relationships for both women and URM students in an effort to create a workforce that aligns with the changing demographics of medical graduates to continue to improve diversity in neurosurgery.

Reliable Prediction of Discharge Disposition Following Cervical Spine Surgery With Ensemble Machine Learning and Validation on a National Cohort.

STUDY DESIGN: A retrospective cohort study. OBJECTIVE: The purpose of this study is to develop a machine learning algorithm to predict nonhome discharge after cervical spine surgery that is validated and usable on a national scale to ensure generalizability and elucidate candidate drivers for prediction. SUMMARY OF BACKGROUND DATA: Excessive length of hospital stay can be attributed to delays in postoperative referrals to intermediate care rehabilitation centers or skilled nursing facilities. Accurate preoperative prediction of patients who may require access to these resources can facilitate a more efficient referral and discharge process, thereby reducing hospital and patient costs in addition to minimizing the risk of hospital-acquired complications. METHODS: Electronic medical records were retrospectively reviewed from a single-center data warehouse (SCDW) to identify patients undergoing cervical spine surgeries between 2008 and 2019 for machine learning algorithm development and internal validation. The National Inpatient Sample (NIS) database was queried to identify cervical spine fusion surgeries between 2009 and 2017 for external validation of algorithm performance. Gradient-boosted trees were constructed to predict nonhome discharge across patient cohorts. The area under the receiver operating characteristic curve (AUROC) was used to measure model performance. SHAP values were used to identify nonlinear risk factors for nonhome discharge and to interpret algorithm predictions. RESULTS: A total of 3523 cases of cervical spine fusion surgeries were included from the SCDW data set, and 311,582 cases were isolated from NIS. The model demonstrated robust prediction of nonhome discharge across all cohorts, achieving an area under the receiver operating characteristic curve of 0.87 (SD=0.01) on both the SCDW and nationwide NIS test sets. Anterior approach only, age, elective admission status, Medicare insurance status, and total Elixhauser Comorbidity Index score were the most important predictors of discharge destination. CONCLUSIONS: Machine learning algorithms reliably predict nonhome discharge across single-center and national cohorts and identify preoperative features of importance following cervical spine fusion surgery.

A Century of Evolution in Spine Surgery Publications: A Bibliometric Analysis of the Field From 1900 to 2023.

BACKGROUND AND OBJECTIVES: Spine surgery has advanced in concert with our deeper understanding of its elements. Narrowly focused bibliometric analyses have been conducted previously, but never on the entire corpus of the field. Using big data and bibliometrics, we appraised the entire corpus of spine surgery publications to study the evolution of the specialty as a scholarly field since 1900. METHODS: We queried Web of Science for all contents from 13 major publications dedicated to spine surgery. We next queried by topic [topic = (spine OR spinal OR vertebrae OR vertebral OR intervertebral OR disc OR disk)]; these results were filtered to include articles published by 49 other publications that were manually determined to contain pertinent articles. Articles, along with their metadata, were exported. Statistical and bibliometric analyses were performed using the Bibliometrix R package and various Python packages. RESULTS: Eighty-five thousand five hundred articles from 62 journals and 134 707 unique authors were identified. The annual growth rate of publications was 2.78%, with a surge after 1980, concurrent with the growth of specialized journals. International coauthorship, absent before 1970, increased exponentially with the formation of influential spine study groups. Reference publication year spectroscopy allowed us to identify 200 articles that comprise the historical roots of modern spine surgery and each of its subdisciplines. We mapped the emergence of new topics and saw a recent lexical evolution toward outcomes- and patient-centric terms. Female and minority coauthorship has increased since 1990, but remains low, and disparities across major publications persist. CONCLUSION: The field of spine surgery was borne from pioneering individuals who published their findings in a variety of journals. The renaissance of spine surgery has been powered by international collaboration and is increasingly outcomes focused. While spine surgery is gradually becoming more diverse, there is a clear need for further promotion and outreach to under-represented populations.

Health system-scale language models are all-purpose prediction engines.

Physicians make critical time-constrained decisions every day. Clinical predictive models can help physicians and administrators make decisions by forecasting clinical and operational events. Existing structured data-based clinical predictive models have limited use in everyday practice owing to complexity in data processing, as well as model development and deployment1-3. Here we show that unstructured clinical notes from the electronic health record can enable the training of clinical language models, which can be used as all-purpose clinical predictive engines with low-resistance development and deployment. Our approach leverages recent advances in natural language processing4,5 to train a large language model for medical language (NYUTron) and subsequently fine-tune it across a wide range of clinical and operational predictive tasks. We evaluated our approach within our health system for five such tasks: 30-day all-cause readmission prediction, in-hospital mortality prediction, comorbidity index prediction, length of stay prediction, and insurance denial prediction. We show that NYUTron has an area under the curve (AUC) of 78.7-94.9%, with an improvement of 5.36-14.7% in the AUC compared with traditional models. We additionally demonstrate the benefits of pretraining with clinical text, the potential for increasing generalizability to different sites through fine-tuning and the full deployment of our system in a prospective, single-arm trial. These results show the potential for using clinical language models in medicine to read alongside physicians and provide guidance at the point of care.

Early Experience of Surgical Planning for STA-MCA Bypass Using Virtual Reality.

BACKGROUND: The superficial temporal artery (STA)-to-middle cerebral artery (MCA) bypass requires precise preoperative planning, and 3-dimensional virtual reality (VR) models have recently been used to optimize planning of STA-MCA bypass. In the present report, we have described our experience with VR-based preoperative planning of STA-MCA bypass. METHODS: Patients from August 2020 to February 2022 were analyzed. For the VR group, using 3-dimensional models from the patients' preoperative computed tomography angiograms, VR was used to locate the donor vessels, potential recipient, and anastomosis sites and plan the craniotomy, which were referenced throughout surgery. Computed tomography angiograms or digital subtraction angiograms were used to plan the craniotomy for the control group. The procedure time, bypass patency, craniotomy size, and postoperative complication rates were assessed. RESULTS: The VR group included 17 patients (13 women; age, 49 ± 14 years) with Moyamoya disease (76.5%) and/or ischemic stroke (29.4%). The control group included 13 patients (8 women; age, 49 ± 12 years) with Moyamoya disease (92.3%) and/or ischemic stroke (7.3%). For all 30 patients, the preoperatively planned donor and recipient branches were effectively translated intraoperatively. No significant difference were found in the procedure time or craniotomy size between the 2 groups. Bypass patency was 94.1% for the VR group (16 of 17) and 84.6% for the control group (11 of 13). No permanent neurological deficits occurred in either group. CONCLUSIONS: Our early experience has shown that VR can serve as a useful, interactive preoperative planning tool by enhancing visualization of the spatial relationship between the STA and MCA without compromising the surgical results.

Responder Analysis of Pain Relief After Surgery for the Treatment of Spinal Metastatic Tumors.

BACKGROUND: Central tendency analysis studies demonstrate that surgery provides pain relief in spinal metastatic tumors. However, they preclude patient-specific probability of treatment outcome. OBJECTIVE: To use responder analysis to study the variability of pain improvement. METHODS: In this single-center, retrospective analysis, 174 patients were studied. Logistic regression modeling was used to associate preoperative characteristics with rating the Brief Pain Inventory (BPI) worst pain item 0 to 4. Linear regression modeling was used to associate preoperative characteristics with minimal clinically important improvement (MCI) in physical functioning defined by a 1-point decrease in the BPI Interference Construct score from preoperative baseline to 6 months postoperatively. RESULTS: Patient-level analysis revealed that 60% of patients experienced an improvement in pain. At least half experienced a decrease in pain resulting in MCI in physical functioning. Cutpoint analysis revealed that 48% were responders. Increasing scores on the preoperative pain intensity BPI items, the MD Anderson Symptom Inventory (MDASI) Core Symptom Severity Construct, the MDASI Spine Tumor-Specific Construct, the presence of preoperative neurologic deficits, and postoperative complications were associated with lower probability of treatment success while increasing severity in all BPI pain items, and MDASI constructs were associated with increased probability of MCI in physical function. Significant mortality and loss to follow-up intrinsic to this patient population limit the strength of these data. CONCLUSION: Although patients with milder preoperative symptoms are likely to achieve better pain relief after surgery, patients with worse preoperative symptom also benefit from surgery with adequate pain relief with an improvement in physical function.

Pragmatic Prediction of Excessive Length of Stay After Cervical Spine Surgery With Machine Learning and Validation on a National Scale.

BACKGROUND: Extended postoperative hospital stays are associated with numerous clinical risks and increased economic cost. Accurate preoperative prediction of extended length of stay (LOS) can facilitate targeted interventions to mitigate clinical harm and resource utilization. OBJECTIVE: To develop a machine learning algorithm aimed at predicting extended LOS after cervical spine surgery on a national level and elucidate drivers of prediction. METHODS: Electronic medical records from a large, urban academic medical center were retrospectively examined to identify patients who underwent cervical spine fusion surgeries between 2008 and 2019 for machine learning algorithm development and in-sample validation. The National Inpatient Sample database was queried to identify cervical spine fusion surgeries between 2009 and 2017 for out-of-sample validation of algorithm performance. Gradient-boosted trees predicted LOS and efficacy was assessed using the area under the receiver operating characteristic curve (AUROC). Shapley values were calculated to characterize preoperative risk factors for extended LOS and explain algorithm predictions. RESULTS: Gradient-boosted trees accurately predicted extended LOS across cohorts, achieving an AUROC of 0.87 (SD = 0.01) on the single-center validation set and an AUROC of 0.84 (SD = 0.00) on the nationwide National Inpatient Sample data set. Anterior approach only, elective admission status, age, and total number of Elixhauser comorbidities were important predictors that affected the likelihood of prolonged LOS. CONCLUSION: Machine learning algorithms accurately predict extended LOS across single-center and national patient cohorts and characterize key preoperative drivers of increased LOS after cervical spine surgery.

Robust Prediction of Non-home Discharge After Thoracolumbar Spine Surgery With Ensemble Machine Learning and Validation on a Nationwide Cohort.

BACKGROUND: Delays in postoperative referrals to rehabilitation or skilled nursing facilities contribute toward extended hospital stays. Facilitating more efficient referrals through accurate preoperative prediction algorithms has the potential to reduce unnecessary economic burden and minimize risk of hospital-acquired complications. We develop a robust machine learning algorithm to predict non-home discharge after thoracolumbar spine surgery that generalizes to unseen populations and identifies markers for prediction. METHODS: Retrospective electronic health records were obtained from our single-center data warehouse (SCDW) to identify patients undergoing thoracolumbar spine surgeries between 2008 and 2019 for algorithm development and internal validation. The National Inpatient Sample (NIS) database was queried to identify thoracolumbar surgeries between 2009 and 2017 for out-of-sample validation. Ensemble decision trees were constructed for prediction and area under the receiver operating characteristic curve (AUROC) was used to assess performance. Shapley additive explanations values were derived to identify drivers of non-home discharge for interpretation of algorithm predictions. RESULTS: A total of 5224 cases of thoracolumbar spine surgeries were isolated from the SCDW and 492,312 cases were identified from NIS. The model achieved an AUROC of 0.81 (standard deviation [SD] = 0.01) on the SCDW test set and 0.77 (SD = 0.01) on the nationwide NIS data set, thereby demonstrating robust prediction of non-home discharge across all diverse patient cohorts. Age, total Elixhauser comorbidities, Medicare insurance, weighted Elixhauser score, and female sex were among the most important predictors of non-home discharge. CONCLUSIONS: Machine learning algorithms reliably predict non-home discharge after thoracolumbar spine surgery across single-center and national cohorts and identify preoperative features of importance that elucidate algorithm decision-making.

Management and surgical outcomes of dystrophic scoliosis in neurofibromatosis type 1: a systematic review.

OBJECTIVE: Neurofibromatosis type 1 (NF1) dystrophic scoliosis is an early-onset, rapidly progressive multiplanar deformity. There are few studies on the surgical management of this patient population. Specifically, perioperative morbidity, instrument-related complications, and quality-of-life outcomes associated with surgical management have not been systematically evaluated. In this study, the authors aimed to perform a systematic review on the natural history, management options, and surgical outcomes in patients who underwent NF1 dystrophic scoliosis surgery. METHODS: A PubMed search for articles with "neurofibromatosis" and either "dystrophic" or "scoliosis" in the title or abstract was performed. Articles with 10 or more patients undergoing surgery for NF1 dystrophic scoliosis were included. Data regarding indications, treatment details, morbidity, and outcomes were summarized and analyzed with descriptive statistics. RESULTS: A total of 310 articles were identified, 48 of which were selected for full-text review; 30 studies describing 761 patients met the inclusion criteria. The mean age ranged from 7 to 22 years, and 99.7% of patients were younger than 18 years. The mean preoperative coronal Cobb angle was 75.2°, and the average correction achieved was 40.3°. The mean clinical follow-up in each study was at least 2 years (range 2.2-19 years). All patients underwent surgery with the intent of deformity correction. The scoliosis regions addressed were thoracic curves (69.6%) and thoracolumbar (11.1%) and lumbar (14.3%) regions. The authors reported on a variety of approaches: posterior-only, combined anterior-posterior, and growth-friendly surgery. For fixation techniques, 42.5% of patients were treated with hybrid constructs, 51.5% with pedicle screw-only constructs, and 6.0% with hook-based constructs. Only 0.9% of patients underwent a vertebral column resection. The nonneurological complication rate was 14.0%, primarily dural tears and wound infections. The immediate postoperative neurological deficit rate was 2.1%, and the permanent neurological deficit rate was 1.2%. Ultimately, 21.5% required revision surgery, most commonly for implant-related complications. Loss of correction in both the sagittal and coronal planes commonly occurred at follow-up. Five papers supplied validated patient-reported outcome measures, showing improvement in the mental health, self-image, and activity domains. CONCLUSIONS: Data on the surgical outcomes of dystrophic scoliosis correction are heterogeneous and sparse. The perioperative complication rate appears to be high, although reported rates of neurological deficits appear to be lower than clinically observed and may be underreported. The incidence of implant-related failures requiring revision surgery is high. There is a great need for multicenter prospective studies of this complex type of deformity.

Salvage Superficial Temporal Artery to Middle Cerebral Artery Direct Bypass Using an Interposition Graft for Failed Encephaloduroarteriosynangiosis in Moyamoya Disease.

BACKGROUND: Moyamoya disease may present with either hemorrhagic or ischemic strokes. Surgical bypass has previously demonstrated superiority when compared to natural history and medical treatment alone. The best bypass option (direct vs. indirect), however, remains controversial in regard to adult ischemic symptomatic moyamoya disease. Multiple studies have demonstrated clinical as well as angiographic effectiveness of direct bypass in adult hemorrhagic moyamoya disease. In particular, there are limited data regarding strategies in the setting of failed indirect bypass with recurrent hemorrhagic strokes. Here, we describe a salvage procedure. METHODS: We describe a case of a 52-year-old man who presented with hemorrhagic moyamoya disease and failed previous bilateral encephaloduroarteriosynangiosis (EDAS) procedures at an outside institution. On a 3-year follow-up diagnostic cerebral angiogram, no synangiosis was noted on the right side and only minimal synangiosis was present on the left. The left hemisphere was significant for a left parietal hypoperfusion state. We performed a salvage left proximal superficial temporal artery to distal parietal M4 middle cerebral artery bypass using the descending branch of the lateral circumflex artery as an interposition graft with preservation of the existing EDAS sites. RESULTS: The patient underwent the procedure successfully and recovered well with resolution of headaches and no further strokes or hemorrhages on the 1-year follow-up magnetic resonance imaging of the brain. CONCLUSIONS: This case presents the use of a salvage direct bypass technique for recurrent symptomatic hemorrhagic moyamoya disease after failed EDAS. The strategy, approach, and technical nuances of this unique case have implications for revascularization options.

Cortical stimulation parameters for functional mapping.

PURPOSE: There is significant variation in how patients respond to cortical electrical stimulation. It has been hypothesized that individual demographic and pathologic factors, such as age, sex, disease duration, and MRI findings, may explain this discrepancy. The purpose of our study is to identify specific patient characteristics and their effect on cortical stimulation, and discover the extent of variation in behavioral responses that exists among patients with epilepsy. METHOD: We retrospectively analyzed data from 92 patients with medically intractable epilepsy who had extra-operative cortical electrical stimulation. Mapping records were evaluated and information gathered about demographic data, as well as the thresholds of stimulation for motor, sensory, speech, and other responses; typical seizure behavior; and the induction of afterdischarges. RESULTS: Ninety-two patient cortical stimulation mapping reports were analyzed. The average of the minimum thresholds for motor response was 4.15mA±2.67. The average of the minimum thresholds for sensory response was 3.50mA±2.15. The average of the minimum thresholds for speech response was 4.48mA±2.42. The average of the minimum thresholds for afterdischarge was 4.33mA±2.37. Most striking were the degree of variability and wide range of thresholds seen between patients and within the different regions of the same patient. CONCLUSION: Wide ranges of thresholds exist for the different responses between patients and within different regions of the same patient. With multivariate analysis in these series, no clinical or demographic factors predicted physiological response or afterdischarge threshold levels.

Parkinson-causing α-synuclein missense mutations shift native tetramers to monomers as a mechanism for disease initiation.

ß-Sheet-rich α-synuclein (αS) aggregates characterize Parkinson's disease (PD). αS was long believed to be a natively unfolded monomer, but recent work suggests it also occurs in α-helix-rich tetramers. Crosslinking traps principally tetrameric αS in intact normal neurons, but not after cell lysis, suggesting a dynamic equilibrium. Here we show that freshly biopsied normal human brain contains abundant αS tetramers. The PD-causing mutation A53T decreases tetramers in mouse brain. Neurons derived from an A53T patient have decreased tetramers. Neurons expressing E46K do also, and adding 1-2 E46K-like mutations into the canonical αS repeat motifs (KTKEGV) further reduces tetramers, decreases αS solubility and induces neurotoxicity and round inclusions. The other three fPD missense mutations likewise decrease tetramer:monomer ratios. The destabilization of physiological tetramers by PD-causing missense mutations and the neurotoxicity and inclusions induced by markedly decreasing tetramers suggest that decreased α-helical tetramers and increased unfolded monomers initiate pathogenesis. Tetramer-stabilizing compounds should prevent this.

Purification of α-synuclein from human brain reveals an instability of endogenous multimers as the protein approaches purity.

Despite two decades of research, the structure-function relationships of endogenous, physiological forms of α-synuclein (αSyn) are not well understood. Most in vitro studies of this Parkinson's disease-related protein have focused on recombinant αSyn that is unfolded and monomeric, assuming that this represents its state in the normal human brain. Recently, we have provided evidence that αSyn exists in considerable part in neurons, erythrocytes, and other cells as a metastable multimer that principally sizes as a tetramer. In contrast to recombinant αSyn, physiological tetramers purified from human erythrocytes have substantial α-helical content and resist pathological aggregation into ß-sheet rich fibers. Here, we report the first method to fully purify soluble αSyn from the most relevant source, human brain. We describe protocols that purify αSyn to homogeneity from nondiseased human cortex using ammonium sulfate precipitation, gel filtration, and ion exchange, hydrophobic interaction, and affinity chromatographies. Cross-linking of the starting material and the partially purified chromatographic fractions revealed abundant αSyn multimers, including apparent tetramers, but these were destabilized in large part to monomers during the final purification step. The method also fully purified the homologue ß-synuclein, with a similar outcome. Circular dichroism spectroscopy showed that purified, brain-derived αSyn can display more helical content than the recombinant protein, but this result varied. Collectively, our data suggest that purifying αSyn to homogeneity destabilizes native, α-helix-rich multimers that exist in intact and partially purified brain samples. This finding suggests existence of a stabilizing cofactor (e.g., a small lipid) present inside neurons that is lost during final purification.

N-alpha-acetylation of α-synuclein increases its helical folding propensity, GM1 binding specificity and resistance to aggregation.

A switch in the conformational properties of α-synuclein (αS) is hypothesized to be a key step in the pathogenic mechanism of Parkinson's disease (PD). Whereas the beta-sheet-rich state of αS has long been associated with its pathological aggregation in PD, a partially alpha-helical state was found to be related to physiological lipid binding; this suggests a potential role of the alpha-helical state in controlling synaptic vesicle cycling and resistance to ß-sheet rich aggregation. N-terminal acetylation is the predominant post-translational modification of mammalian αS. Using circular dichroism, isothermal titration calorimetry, and fluorescence spectroscopy, we have analyzed the effects of N-terminal acetylation on the propensity of recombinant human αS to form the two conformational states in interaction with lipid membranes. Small unilamellar vesicles of negatively charged lipids served as model membranes. Consistent with previous NMR studies using phosphatidylserine, we found that membrane-induced α-helical folding was enhanced by N-terminal acetylation and that greater exothermic heat could be measured upon vesicle binding of the modified protein. Interestingly, the folding and lipid binding enhancements with phosphatidylserine in vitro were weak when compared to that of αS with GM1, a lipid enriched in presynaptic membranes. The resultant increase in helical folding propensity of N-acetylated αS enhanced its resistance to aggregation. Our findings demonstrate the significance of the extreme N-terminus for folding nucleation, for relative GM1 specificity of αS-membrane interaction, and for a protective function of N-terminal-acetylation against αS aggregation mediated by GM1.