neuroGPT-X: toward a clinic-ready large language model.

OBJECTIVE: The objective was to assess the performance of a context-enriched large language model (LLM) compared with international neurosurgical experts on questions related to the management of vestibular schwannoma. Furthermore, another objective was to develop a chat-based platform incorporating in-text citations, references, and memory to enable accurate, relevant, and reliable information in real time. METHODS: The analysis involved 1) creating a data set through web scraping, 2) developing a chat-based platform called neuroGPT-X, 3) enlisting 8 expert neurosurgeons across international centers to independently create questions (n = 1) and to answer (n = 4) and evaluate responses (n = 3) while blinded, and 4) analyzing the evaluation results on the management of vestibular schwannoma. In the blinded phase, all answers were assessed for accuracy, coherence, relevance, thoroughness, speed, and overall rating. All experts were unblinded and provided their thoughts on the utility and limitations of the tool. In the unblinded phase, all neurosurgeons provided answers to a Likert scale survey and long-answer questions regarding the clinical utility, likelihood of use, and limitations of the tool. The tool was then evaluated on the basis of a set of 103 consensus statements on vestibular schwannoma care from the 8th Quadrennial International Conference on Vestibular Schwannoma. RESULTS: Responses from the naive and context-enriched Generative Pretrained Transformer (GPT) models were consistently rated not significantly different in terms of accuracy, coherence, relevance, thoroughness, and overall performance, and they were often rated significantly higher than expert responses. Both the naive and content-enriched GPT models provided faster responses to the standardized question set than expert neurosurgeon respondents (p < 0.01). The context-enriched GPT model agreed with 98 of the 103 (95%) consensus statements. Of interest, all expert surgeons expressed concerns about the reliability of GPT in accurately addressing the nuances and controversies surrounding the management of vestibular schwannoma. Furthermore, the authors developed neuroGPT-X, a chat-based platform designed to provide point-of-care clinical support and mitigate the limitations of human memory. neuroGPT-X incorporates features such as in-text citations and references to enable accurate, relevant, and reliable information in real time. CONCLUSIONS: The present study, with its subspecialist-level performance in generating written responses to complex neurosurgical problems for which evidence-based consensus for management is lacking, suggests that context-enriched LLMs show promise as a point-of-care medical resource. The authors anticipate that this work will be a springboard for expansion into more medical specialties, incorporating evidence-based clinical information and developing expert-level dialogue surrounding LLMs in healthcare.

Virtual Reality Simulation for the Middle Cranial Fossa Approach: A Validation Study.

BACKGROUND AND OBJECTIVES: Virtual reality (VR) surgical rehearsal is an educational tool that exists in a safe environment. Validation is necessary to establish the educational value of this platform. The middle cranial fossa (MCF) is ideal for simulation because trainees have limited exposure to this approach and it has considerable complication risk. Our objectives were to assess the face, content, and construct validities of an MCF VR simulation, as well as the change in performance across serial simulations. METHODS: Using high-resolution volumetric data sets of human cadavers, the authors generated a high-fidelity visual and haptic rendering of the MCF approach using CardinalSim software. Trainees from Neurosurgery and Otolaryngology-Head and Neck Surgery at two Canadian academic centers performed MCF dissections on this VR platform. Randomization was used to assess the effect of enhanced VR interaction. Likert scales were used to assess the face and content validities. Performance metrics and pre- and postsimulation test scores were evaluated. Construct validity was evaluated by examining the effect of the training level on simulation performance. RESULTS: Twenty trainees were enrolled. Face and content validities were achieved in all domains. Construct validity, however, was not demonstrated. Postsimulation test scores were significantly higher than presimulation test scores ( P < .001 ). Trainees demonstrated statistically significant improvement in the time to complete dissections ( P < .001 ), internal auditory canal skeletonization ( P < .001 ), completeness of the anterior petrosectomy ( P < .001 ), and reduced number of injuries to critical structures ( P = .001 ). CONCLUSION: This MCF VR simulation created using CardinalSim demonstrated face and content validities. Construct validity was not established because no trainee included in the study had previous MCF approach experience, which further emphasizes the importance of simulation. When used as a formative educational adjunct in both Neurosurgery and Otolaryngology-Head and Neck Surgery, this simulation has the potential to enhance understanding of the complex anatomic relationships of critical neurovascular structures.

Missed pituitary microadenoma during endoscopic transsphenoidal surgery for short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing with symptom relief: illustrative case.

BACKGROUND: Short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing (SUNCT) is a rare headache disorder that has been associated with pituitary adenomas. Resection has been posited to be curative. OBSERVATIONS: A 60-year-old female presented with a 10-year history of SUNCT, which had been medically refractory. Sellar magnetic resonance imaging (MRI) showed a 2 × 2 mm nodule in the right anterolateral aspect of the pituitary. Endoscopic endonasal transsphenoidal resection of the pituitary microadenoma with neuronavigation was performed. The patient felt immediate relief from the headaches. Postoperative MRI showed persistence of the pituitary microadenoma and the resection tract to be inferomedial to the lesion. The right middle and partial superior turbinectomy site was close to the sphenopalatine foramen (SPF). The patient was discharged on postoperative day 1 and remained headache-free without any medications at the 4-month follow-up. LESSONS: Resection of pituitary lesions associated with SUNCT may not necessarily be the cause of SUNCT resolution. Manipulation of the middle and superior turbinate close to the SPF may lead to a pterygopalatine ganglion block. This may be the mechanism of cure for SUNCT in patients with related pituitary lesions who undergo endonasal resection.

The user experience design of a novel microscope within SurgiSim, a virtual reality surgical simulator.

PURPOSE: Virtual reality (VR) simulation has the potential to advance surgical education, procedural planning, and intraoperative guidance. "SurgiSim" is a VR platform developed for the rehearsal of complex procedures using patient-specific anatomy, high-fidelity stereoscopic graphics, and haptic feedback. SurgiSim is the first VR simulator to include a virtual operating room microscope. We describe the process of designing and refining the VR microscope user experience (UX) and user interaction (UI) to optimize surgical rehearsal and education. METHODS: Human-centered VR design principles were applied in the design of the SurgiSim microscope to optimize the user's sense of presence. Throughout the UX's development, the team of developers met regularly with surgeons to gather end-user feedback. Supplemental testing was performed on four participants. RESULTS: Through observation and participant feedback, we made iterative design upgrades to the SurgiSim platform. We identified the following key characteristics of the VR microscope UI: overall appearance, hand controller interface, and microscope movement. CONCLUSION: Our design process identified challenges arising from the disparity between VR and physical environments that pertain to microscope education and deployment. These roadblocks were addressed using creative solutions. Future studies will investigate the efficacy of VR surgical microscope training on real-world microscope skills as assessed by validated performance metrics.

Vertebrae segmentation in reduced radiation CT imaging for augmented reality applications.

PURPOSE: There is growing evidence for the use of augmented reality (AR) navigation in spinal surgery to increase surgical accuracy and improve clinical outcomes. Recent research has employed AR techniques to create accurate auto-segmentations, the basis of patient registration, using reduced radiation dose intraoperative computed tomography images. In this study, we aimed to determine if spinal surgery AR applications can employ reduced radiation dose preoperative computed tomography (pCT) images. METHODS: We methodically decreased the imaging dose, with the addition of Gaussian noise, that was introduced into pCT images to determine the image quality threshold that was required for auto-segmentation. The Gaussian distribution's standard deviation determined noise level, such that a scalar multiplier (L: [0.00, 0.45], with steps of 0.03) simulated lower doses as L increased. We then enhanced the images with denoising algorithms to evaluate the effect on the segmentation. RESULTS: The pCT radiation dose was decreased to below the current lowest clinical threshold and the resulting images produced segmentations that were appropriate for input into AR applications. This held true at simulated dose L = 0.06 (estimated 144 mAs) but not at L = 0.09 (estimated 136 mAs). The application of denoising algorithms to the images resulted in increased artifacts and decreased bone density. CONCLUSIONS: The pCT image quality that is required for AR auto-segmentation is lower than that which is currently employed in spinal surgery. We recommend a reduced radiation dose protocol of approximately 140 mAs. This has the potential to reduce the radiation experienced by patients in comparison to procedures without AR support. Future research is required to identify the specific, clinically relevant radiation dose thresholds required for surgical navigation.

Data analytics interrogates robotic surgical performance using a microsurgery-specific haptic device.

OBJECTIVES: With the increase in robot-assisted cases, recording the quantifiable dexterity of surgeons is essential for proficiency evaluations. The present study employs sensor-based kinematics and recorded surgeon experience for evaluating a new haptic device. METHODS: Thirty surgeons performed a task simulating micromanipulation with neuroArmPLUSHD and two commercially available hand-controllers. The surgical performance was evaluated based on subjective measures obtained from survey and objective features derived from the sensors. Statistical analyses were performed to assess the hand-controllers and regression analysis was used to identify the key features and develop a machine learning model for surgical skill assessment. FINDINGS: MANCOVA tests on objective features demonstrated significance (α = 0.05) for time (p = 0.02), errors (p = 0.01), distance (p = 0.03), clutch incidents (p = 0.03), and forces (p = 0.00). The majority of metrics were in favor of neuroArmPLUSHD. The surgeons found it smoother, more comfortable, less tiring, and easier to maneuver with more realistic force feedback. The ensemble machine learning model trained with 5-fold cross-validation showed an accuracy (SD) of 0.78 (0.15) in surgeon skill classification. CONCLUSIONS: This study validates the importance of incorporating a superior haptic device in telerobotic surgery for standardization of surgical education and patient care.

The impact of machine learning on patient care: A systematic review.

BACKGROUND: Despite the expanding use of machine learning (ML) in fields such as finance and marketing, its application in the daily practice of clinical medicine is almost non-existent. In this systematic review, we describe the various areas within clinical medicine that have applied the use of ML to improve patient care. METHODS: A systematic review was performed in accordance with the PRISMA guidelines using Medline(R), EBM Reviews, Embase, Psych Info, and Cochrane Databases, focusing on human studies that used ML to directly address a clinical problem. Included studies were published from January 1, 2000 to May 1, 2018 and provided metrics on the performance of the utilized ML tool. RESULTS: A total of 1909 unique publications were reviewed, with 378 retrospective articles and 8 prospective articles meeting inclusion criteria. Retrospective publications were found to be increasing in frequency, with 61 % of articles published within the last 4 years. Prospective articles comprised only 2 % of the articles meeting our inclusion criteria. These studies utilized a prospective cohort design with an average sample size of 531. CONCLUSION: The majority of literature describing the use of ML in clinical medicine is retrospective in nature and often outlines proof-of-concept approaches to impact patient care. We postulate that identifying and overcoming key translational barriers, including real-time access to clinical data, data security, physician approval of "black box" generated results, and performance evaluation will allow for a fundamental shift in medical practice, where specialized tools will aid the healthcare team in providing better patient care.

Management of Venous Sinus Thrombosis.

Cerebral venous sinus thrombosis (CVST) is a rare subtype of cerebrovascular disease representing 0.5% of strokes. The signs and symptoms of CVST are often nonspecific, and variable in duration, with the common results being delayed diagnosis and treatment. Increased awareness in the medical community and advancements in imaging modalities have produced faster diagnosis with improved patient outcomes. The preferred initial treatment is with a low molecular weight heparin. After the acute stage of CVST, treatment with a vitamin K antagonist (oral anticoagulant therapy) is recommended. Current evidence suggests that in the future, factor Xa inhibitor drugs may be used for long-term therapy.

Conflict of interest policies and disclosure requirements in neurosurgical journals.

OBJECTIVE: An increasing amount of funding in neurosurgery research comes from industry, which may create a conflict of interest (COI) and the potential to bias results. The reporting and handling of COIs have become difficult, particularly as explicit policies themselves and definitions thereof continue to vary between medical journals. In this study, the authors sought to evaluate the prevalence and comprehensiveness of COI policies among leading neurosurgical journals. METHODS: The authors conducted a cross-sectional study of publicly available online disclosure policies in the 20 highest-ranking neurosurgical journals, as determined by Google Scholar Metrics, in July 2016. RESULTS: Overall, 89.5% of the highest-impact neurosurgical journals included COI policy statements. Ten (53%) journals requested declaration of nonfinancial conflicts, while 2 journals specifically set a time period for COIs. Sixteen journals required declaration from the corresponding author, 13 from all authors, 6 from reviewers, and 5 from editors. Four journals were included in the International Committee of Medical Journal Editors (ICMJE) list of publications that follow the Uniform Requirements for Manuscripts Submitted to Biomedical Journals (currently known as Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals). Five journal policies included COI declaration verification, management, or enforcement. The neurosurgery journals with more comprehensive COI policies were significantly more likely to have higher h5-indices (p = 0.003) and higher impact factors (p = 0.01). CONCLUSIONS: In 2016, the majority of, but not all, high-impact neurosurgical journals had publically available COI disclosure policies. Policy inclusiveness and comprehensiveness varied substantially across neurosurgical journals, but COI comprehensiveness was associated with other established markers of individual journals' favorability and influence, such as impact factor and h5-index.

Industry Financial Relationships in Neurosurgery in 2015: Analysis of the Sunshine Act Open Payments Database.

OBJECTIVE: The 2013 Physician Payments Sunshine Act mandates that all U.S. drug and device manufacturers disclose payments to physicians. All payments are made available annually in the Open Payments Database (OPD). Our aim was to determine prevalence, magnitude, and nature of these payments to physicians performing neurologic surgery in 2015 and to discuss the role that financial conflicts of interest play in neurosurgery. METHODS: All records of industry financial relationships with physicians identified by the neurological surgery taxonomy code in 2015 were accessed via the OPD. Data were analyzed in terms of type and amounts of payments, companies making payments, and comparison with previous studies. RESULTS: In 2015, 83,690 payments (totaling $99,048,607) were made to 7613 physicians by 330 companies. Of these, 0.01% were >$1 million, and 73.2% were <$100. The mean payment ($13,010) was substantially greater than the median ($114). Royalties and licensing accounted for the largest monetary value of payments (74.2%) but only 1.7% of the total number. Food and beverage payments were the most commonly reported transaction (75%) but accounted for only 2.5% of total reported monetary value. Neurologic surgery had the second highest average total payment per physician of any specialty. CONCLUSIONS: The neurological surgery specialty receives substantial annual payments from industry in the United States. The overall value is driven by a small number of payments of high monetary value. The OPD provides a unique opportunity for increased transparency in industry-physician relationships facilitating disclosure of financial conflicts of interest.