Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Jan 8, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

986

Evaluating Large Language Models for Decision Support in Minimally Invasive Spine Surgery Triage and Procedural

Ahmet Kartal1, Noel F Manalil1, Chiungwen D Cheng1

  • 1Department of Neurological Surgery, Och Spine at NewYork-Presbyterian Hospital, New York, NY, USA.

Global Spine Journal
|December 22, 2025
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Jack Fowle: Combining Values, Experience, and Teamwork to Improve Risk Analysis.

Risk analysis : an official publication of the Society for Risk Analysis·2026
Same author

Identifying an X-Ray Threshold for Cage Subsidence After Single-Level Minimally Invasive Transforaminal Lumbar Interbody Fusion: A Diagnostic Threshold Study Using Intraoperative CT as the Reference Standard.

Journal of clinical medicine·2026
Same author

The "one-and-a-half" minimally invasive transforaminal lumbar interbody fusion: a single-center retrospective case series.

Journal of neurosurgery. Spine·2026
Same author

Wayne Landis: Evolution of Ecological Risk Assessment.

Risk analysis : an official publication of the Society for Risk Analysis·2026
Same author

Benchmarking Multimodal Vision Frontier Models With Lumbar Spine MRIs for Grading Lumbar Spinal Stenosis.

Global spine journal·2026
Same author

Vanadium salt inhibits osteoclast formation and preserves bone integrity in a rat model of type-1 diabetes associated bone loss.

Biometals : an international journal on the role of metal ions in biology, biochemistry, and medicine·2026

Generative artificial intelligence (AI) shows promise in differentiating surgical vs. non-surgical triage for minimally invasive spine surgery (MISS). However, expert oversight remains crucial for procedure selection as AI models mature.

Area of Science:

  • Neurosurgery
  • Artificial Intelligence
  • Medical Informatics

Background:

  • Generative artificial intelligence (AI), particularly large language models (LLMs), is increasingly influencing medical decision-making.
  • Applications of LLMs in minimally invasive spine surgery (MISS) are emerging but require rigorous evaluation.

Purpose of the Study:

  • To assess the ability of OpenAI's ChatGPT-5 Pro and Google's Gemini 2.5 Pro to categorize management of published MISS cases.
  • To measure the agreement of LLM-generated categories with expert classifications at procedural and binary triage levels.

Main Methods:

  • A cross-sectional study utilizing 90 clinical vignettes derived from published MISS case reports.
  • LLMs were prompted to assign management categories, with agreement assessed using Jensen-Shannon divergence, Stuart-Maxwell tests, Cohen's kappa, and McNemar's test.
Keywords:
decision supportlarge language modelsminimally invasive spine surgerynatural language processingsurgical triage

More Related Videos

Optimizing Minimally Invasive Spine Surgery: A Fully 3D CT O-Arm Navigated Workflow in MIS TLIF
08:34

Optimizing Minimally Invasive Spine Surgery: A Fully 3D CT O-Arm Navigated Workflow in MIS TLIF

Published on: October 17, 2025

357

Related Experiment Videos

Last Updated: Jan 8, 2026

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

986
Optimizing Minimally Invasive Spine Surgery: A Fully 3D CT O-Arm Navigated Workflow in MIS TLIF
08:34

Optimizing Minimally Invasive Spine Surgery: A Fully 3D CT O-Arm Navigated Workflow in MIS TLIF

Published on: October 17, 2025

357

Main Results:

  • Both LLMs demonstrated small divergence from expert reference categories (JSD: 0.115 for ChatGPT-5 Pro, 0.112 for Gemini 2.5 Pro).
  • Case-level agreement was slight for ChatGPT-5 Pro (κ=0.146) and fair for Gemini 2.5 Pro (κ=0.245).
  • Agreement significantly improved when collapsing categories to surgical vs. non-surgical triage (κ=0.415-0.587 vs. reference; κ=0.692 between models).

Conclusions:

  • LLMs show potential in distinguishing between surgical and non-surgical triage recommendations in MISS.
  • Expert clinical judgment remains essential for definitive procedure selection in MISS until AI systems achieve greater maturity.
  • These findings provide a foundational understanding for integrating LLMs into surgical triage workflows, highlighting both the potential and limitations of AI in precision spine care.