Faculty Bibliography

BACKGROUND AND OBJECTIVES/OBJECTIVE:Myxopapillary ependymomas (MPE) and intradural lumbosacral schwannomas may be challenging to distinguish based on presenting characteristics and preoperative imaging. Accurate differentiation is crucial, as MPEs carry a risk of cerebrospinal fluid dissemination and warrant earlier intervention, a more tailored surgical strategy, consideration for adjuvant radiation, and frequent surveillance. Here, we describe our institutional experience with these tumors and develop a radiomics-based machine learning model to help distinguish them on preoperative imaging. METHODS:Institutional surgical records from 2011 to 2025 were queried and clinical data were extracted for the retrospective cohort analysis. Tumors were manually segmented in ITK-Snap from T1 postcontrast images, and radiomics features were extracted using the PyRadiomics package. An ensemble of random forest, k-nearest neighbors, and naive Bayes classifiers was trained on a subset of radiomics features using nested cross-validation. RESULTS:< .001) in MPEs, likely due to longitudinal tumor growth along the filum. Excluding scoliotic patients did not significantly alter discrimination, suggesting robustness to vertebral column malalignment that may coexist with intradural tumors. CONCLUSION/CONCLUSIONS:A radiomics-based machine learning model demonstrated excellent discriminative ability between MPE and lumbosacral schwannoma, achieving high accuracy and robustness to vertebral alignment variations. These results suggest that radiomics-based models may be developed into a useful tool for preoperative planning and patient counseling.

Vertebral metastatic disease results from many types of cancer and can have a devastating impact on patient mobility, psychological health, quality of life, and ultimately overall patient survival. However, the development of radiotherapy and surgical techniques has rapidly surged in conjunction with ongoing advances in basic science and translational studies. In this review, we discuss the paradigm shift in our understanding of the epidemiology and treatment algorithms for spinal oncology, ranging from preoperative optimization strategies, radiation and surgical techniques, the utilization of molecular markers and targeted therapeutics in medical oncology, and prognostication tools that underscore a new multidisciplinary approach to spinal oncology care.

BACKGROUND AND OBJECTIVES/OBJECTIVE:Intramedullary spinal cord tumor (IMSCT) resection carries a high risk of postoperative neurological deficit because of neural tract manipulation and myelotomy. Although short-term and long-term neurological recovery represent key treatment outcomes, current prognostication methods are lacking and would benefit from further complex analysis. METHODS:From March 2009 to August 2025, all adult IMSCT resections at our institution were reviewed. Demographic, oncologic, and perioperative data were extracted from electronic medical records. This included preoperative and follow-up neurological examination data in the form of American Spinal Injury Association Impairment Scale (AIS) grading, Modified McCormick Scale (MMCS), and ambulatory status. Independent ambulation served as the primary outcome for 4 machine learning models. Each model was sequentially evaluated using area under the receiver operating characteristic curve (AUROC). RESULTS:Fifty-four patients underwent 55 surgeries for IMSCT resection. Encapsulated lesions predominated IMSCT pathology, with grade II ependymoma comprising 28 (50.9%) resections, 5 hemangioblastomas (9.1%), and 5 cavernous hemangiomas (9.1%). Gross total resection was achieved in 36 cases (65.5%), with encapsulated tumors more readily achieving gross total resection vs unencapsulated (84.6% vs 18.8%, P < .01). By 4 weeks, conversion of MMCS, but not AIS grade, significantly correlated with concurrent ambulatory conversion (P < .01 vs P = .15). At 6 months, both AIS grade conversion (P < .01) and MMCS conversion (P < .01) significantly correlated with ambulatory conversion. For predicting ambulation at latest follow-up from 4 weeks postoperatively, the comprehensive granular model achieved an AUROC of 0.833, outperforming the AIS grade (0.583), American Spinal Injury Association Motor Score (0.667), and MMCS (0.667) models. By the 6-month follow-up, the comprehensive granular model achieved strong discrimination (AUROC 1.00). CONCLUSION/CONCLUSIONS:Follow-up IMSCT data demonstrate a postoperative lability that stabilizes by 6 months into a reliably modeled outcome. By enhancing the granularity of recovery data, accurate independent ambulation modeling may improve counseling for patients with IMSCT.

BACKGROUND AND OBJECTIVES/OBJECTIVE:General purpose vision-language models (VLMs) demonstrate impressive capabilities, but their opaque training on uncurated internet data poses critical limitations for high-stakes decision making, such as in neurosurgery. We present CNS-Obsidian, a neurosurgical VLM trained on peer-reviewed neurosurgical literature, and demonstrate its clinical utility compared with GPT-4o in a real-world setting. METHODS:We compiled 23 984 articles from Neurosurgery Publications journals, yielding 78 853 figures and captions. Using GPT-4o and Claude Sonnet-3.5, we converted these image-text pairs into 263 064 training samples across 3 formats: instruction fine-tuning, multiple-choice questions, and differential diagnosis. We trained CNS-Obsidian, a fine-tune of the 34-billion parameter Large Language and Visual Assistant-Next model. In a blinded, randomized deployment trial at NYU Langone Health (August 30-November 30, 2024), neurosurgeons were assigned to use either CNS-Obsidian or a Health Insurance Portability and Accountability Act-compliant GPT-4o end point as a diagnostic copilot after patient consultations. Primary outcomes were diagnostic helpfulness and accuracy, assessed through user ratings and presence of the correct diagnosis within the VLM-provided differential, respectively. RESULTS:CNS-Obsidian matched GPT-4o on synthetic questions (76.13% vs 77.54%, P = .235), but only achieved 46.81% accuracy on human-generated questions vs GPT-4o's 65.70% (P < 10-15). In the randomized trial, 70 consultations were evaluated (32 CNS-Obsidian, 38 GPT-4o) from 959 total consults (7.3% utilization). CNS-Obsidian received positive ratings in 40.62% of cases vs 57.89% for GPT-4o (P = .230). Both models included correct diagnosis in approximately 60% of cases (59.38% vs 65.79%, P = .626). CONCLUSION/CONCLUSIONS:Domain-specific VLMs trained on curated scientific literature can approach frontier model performance in specialized medical domains despite being orders of magnitude smaller and less expensive to train. This establishes a transparent framework for scientific communities to build specialized artificial intelligence models. However, low clinical utilization suggests chatbot interfaces may not align with specialist workflows, indicating need for alternative artificial intelligence integration strategies.

BACKGROUND:Glioblastoma intramedullary spinal cord metastasis (GISCM) is a rare sequela of high-grade astrocytoma and glioblastoma multiforme (GBM). Discrete intramedullary spinal cord metastases are less common than spinal leptomeningeal spread and may follow a more indolent course. Once identified as GISCM, palliative maximal safe resection of the tumor may be considered to alleviate neurological symptoms. Reports describing the surgical management of these rare lesions, including the use of emerging technologies that may aid in maximal safe resection, are sparse. A further understanding is also required regarding the course of disease and factors contributing to mortality in GISCM. METHODS:We reviewed the intraoperative management and clinical course of three patients treated for GISCM at our institution between 2015 and 2024. We additionally conducted a PRISMA-guided systematic literature review of PubMed Central, MEDLINE, and Bookshelf databases through May 26th, 2025, including original patient reports of GISCM from cranial astrocytoma or GBM. The disease course, management strategies, and causes of mortality in previously reported cases were analyzed. RESULTS:Our institutional cohort had a mean time to spinal metastasis of 26.2 months from diagnosis of cranial disease (range 17.5-40.5 months), with a mean survival of 9.2 months following maximal safe resection of extramedullary components (range 7-12 months). In two cases, intraoperative Stimulated Raman Histology (SRH) was employed to facilitate the rapid identification of metastatic GBM, thereby influencing surgical strategy. In one case, 5-aminolevulinic acid (5-ALA) was used to differentiate between tumor and spinal cord parenchyma, facilitating maximal safe debulking without neurological injury. Literature review identified 38 prior reported cases of GISCM, with a median time to spinal diagnosis of 11.0 months and a median survival of 3.5 months thereafter. The cause of death in the review cohort often involved multiple factors, and when analyzed for contributing factors to death, 38.7% involved cranial progression, 38.7% involved progression of spinal disease, and 29.0% involved medical complications. Gait ataxia at presentation was associated with shorter survival in review patients, potentially reflecting advanced disease with extramedullary cord compression. CONCLUSION/CONCLUSIONS:GISCM represents an entity distinct from leptomeningeal disease and may be managed in conjunction with recurrent cranial disease. Surgical debulking is a technically feasible strategy that can be safely facilitated using tools employed in the management of intracranial GBM, facilitating maximal safe resection without compromising survival.

Study DesignLiterature review with clinical recommendations.ObjectiveProviding a clear and concise overview based on the of key literature and consensus expert opinion on spinal fusion following stabilization for spine metastases and offer actionable recommendations on when to fuse and not fuse in this patient population.MethodsKey articles from the published literature on spinal metastases treated with stabilization followed by fusion were reviewed, and clinical recommendations were formulated. The recommendations are categorized as either strong or conditional based on an assessment of methodological quality and expert opinion. This assessment considers factors such as experience, risks, burdens, costs, patient values, and circumstances.ResultsFour articles were selected by practicing spinal oncology surgeons and each was evaluated for its methodological strength and its scientific evidence.ConclusionFusion rarely influences clinical outcomes in metastatic spine surgery. Treatment should prioritize mechanical stability, pain control, functional preservation, and timely continuation of oncologic therapy rather than pursuing bony arthrodesis. Fusion should be considered exclusively in select long-surviving patients, however routine attempts to enhance fusion or delay adjuvant therapy are not justified.[Formula: see text].

OBJECTIVE:Chordomas are rare, locally aggressive primary neoplasms. Resection with negative margins is the primary recommended therapeutic approach, while adjuvant radiotherapy and chemotherapy can also play a role in their treatment in certain situations, including lesions with positive margins or those that are poorly differentiated or dedifferentiated. Cervical spine chordomas pose significant surgical challenges given their proximity to critical anatomical structures and the mechanical constraints of the cervical spine. In the current case series, authors aimed to explore the clinical and patient-reported outcomes (PROs) of the surgical treatment of cervical chordomas in a large multicenter cohort. METHODS:This multicenter case series analysis utilized data from the prospectively collected Primary Tumor Research and Outcomes Network (PTRON) registry, from its inception (May 16, 2016) to data extraction (February 29, 2024). The study population was restricted to patients with histologically confirmed cervical chordomas involving levels C0-7, who underwent surgical treatment at one of the participating centers, and for whom both the initially planned and postoperatively pathologically confirmed surgical margins were documented. Patient demographics, tumor characteristics, surgical and adjuvant treatments, local recurrence-free survival (LRFS), overall survival (OS), and perioperative adverse events were retrieved. PROs included the Spine Oncology Study Group Outcomes Questionnaire version 2.0 (SOSGOQ2.0), EQ-5D, and SF-36 version 2.0 (SF-36v2). RESULTS:Thirty-eight patients were identified, 12 of whom underwent true en bloc resection (EBR), 18 of whom underwent deliberate intralesional resection, and 8 of whom underwent EBR after intralesional surgery or in whom EBR failed. True EBR led to better LRFS (92% vs 83% vs 63%, respectively) and OS (83% vs 39% vs 50%, respectively). Surgical adverse events within 1 year were more frequent with true EBR (100% vs 39% vs 75%, respectively). EQ-5D, SOSGOQ2.0, and SF-36v2 showed improvement with true EBR, whereas the trends for PROs from the other groups were more variable. CONCLUSIONS:This multicenter case series analysis provides critical insights into the clinical outcomes and PROs in the largest cohort of surgically treated cervical spine chordomas described to date. It underscores the importance and challenges of wide resection for oncological control. It establishes the associated morbidity and provides an overview of PROs following surgery. These findings contribute valuable evidence to inform shared decision-making and optimize patient care.

Study DesignNarrative Review.ObjectivesTo summarize the scientific contributions generated from the AO Spine Knowledge Forum Tumor (AOSKFT) databases, focusing on primary spine tumors, and highlight key findings, research trends, and future directions.MethodsData from the Primary Tumor Retrospective (PT-Retro) and Primary Tumor Research Outcome Network (PTRON) registries were analyzed. The nineteen studies included were peer-reviewed manuscripts focused on primary spine tumors, excluding abstracts, book chapters, systematic reviews, and metastatic studies.ResultsThe PT-Retro registry compiled data from 1495 patients across 18 primary tumor histologies, offering insights into recurrence, survival, and treatment paradigms. Key findings emphasize the importance of Enneking-appropriate (EA) resection in improving survival and reducing recurrence in tumors such as chordoma, chondrosarcoma, and osteosarcoma. Genetic markers, including hTERT promoter mutations and rs2305089 SNP, were linked to prognosis in specific histologies. Benign tumors, such as giant cell tumors and aneurysmal bone cysts, demonstrated variable outcomes with different surgical approaches and selective arterial embolization.ConclusionsThe AOSKFT registries have significantly advanced knowledge in primary spine tumor management, emphasizing preoperative staging, surgical margins, and multidisciplinary approaches. International, multicentric registries are essential for studying rare diseases like primary spine tumors, enabling robust data collection, improved statistical power, and broader applicability of findings across diverse clinical settings. Ongoing prospective data collection through PTRON will further refine evidence-based care for these rare and challenging conditions.

BACKGROUND AND OBJECTIVES/OBJECTIVE:The Patient-Generated Subjective Global Assessment (PG-SGA) is a standardized tool for assessing malnutrition in patients with cancer. The primary aim of this study was to assess the impact of preoperative nutritional status as measured by PG-SGA on survival in patients requiring surgical intervention and/or radiotherapy for spinal metastases. METHODS:Patients with spinal metastases who underwent surgery and/or radiation therapy for symptomatic spinal metastases were enrolled in the AO Spine Metastatic Tumor Research and Outcomes Network, a prospective international multicenter research registry, between September 2017 and August 2022. Using the PG-SGA, nutritional status was classified into 3 categories: A, well nourished; B, moderately malnourished; and C, severely malnourished. RESULTS:A total of 589 patients met the inclusion criteria; 362 were classified as well nourished (61%), 159 were moderately malnourished (27%), and 68 were severely malnourished (12%). The median survival was 491 days, 328 days, and 117 days for well-nourished, moderately malnourished, and severely malnourished patients, respectively. In the multivariate analyses, severe malnourishment (HR 2.5 95% CI 1.4-4.3, P < .01) and an ECOG performance status of 3 or 4 (HR 2.7 95% CI 1.2-6.0) remained associated with significantly worse survival. CONCLUSION/CONCLUSIONS:Malnutrition as measured by the PG-SGA demonstrated to be significantly and independently associated with postoperative survival. The PG-SGA is a simple and useful tool to identify spinal metastases patients at risk of early postoperative mortality, and inclusion in the preoperative evaluation of these patients should be considered.

BACKGROUND AND OBJECTIVES/OBJECTIVE:Bone metastases, affecting more than 4.8% of patients with cancer annually, and particularly spinal metastases require urgent intervention to prevent neurological complications. However, the current process of manually reviewing radiological reports leads to potential delays in specialist referrals. We hypothesized that natural language processing (NLP) review of routine radiology reports could automate the referral process for timely multidisciplinary care of spinal metastases. METHODS:We assessed 3 NLP models-a rule-based regular expression (RegEx) model, GPT-4, and a specialized Bidirectional Encoder Representations from Transformers (BERT) model (NYUTron)-for automated detection and referral of bone metastases. Study inclusion criteria targeted patients with active cancer diagnoses who underwent advanced imaging (computed tomography, MRI, or positron emission tomography) without previous specialist referral. We defined 2 separate tasks: task of identifying clinically significant bone metastatic terms (lexical detection), and identifying cases needing a specialist follow-up (clinical referral). Models were developed using 3754 hand-labeled advanced imaging studies in 2 phases: phase 1 focused on spine metastases, and phase 2 generalized to bone metastases. Standard McRae's line performance metrics were evaluated and compared across all stages and tasks. RESULTS:In the lexical detection, a simple RegEx achieved the highest performance (sensitivity 98.4%, specificity 97.6%, F1 = 0.965), followed by NYUTron (sensitivity 96.8%, specificity 89.9%, and F1 = 0.787). For the clinical referral task, RegEx also demonstrated superior performance (sensitivity 92.3%, specificity 87.5%, and F1 = 0.936), followed by a fine-tuned NYUTron model (sensitivity 90.0%, specificity 66.7%, and F1 = 0.750). CONCLUSION/CONCLUSIONS:An NLP-based automated referral system can accurately identify patients with bone metastases requiring specialist evaluation. A simple RegEx model excels in syntax-based identification and expert-informed rule generation for efficient referral patient recommendation in comparison with advanced NLP models. This system could significantly reduce missed follow-ups and enhance timely intervention for patients with bone metastases.