Medical large language models and systems in the clinical application of spinal diseases: Current status, challenges, and future prospects

Large Language Models (LLMs), represented by the Generative Pretrained Transformer (GPT), are profoundly transforming the healthcare sector. Spine medicine, a discipline heavily reliant on complex imaging data, detailed clinical records, and evidence-based medical practice, serves as an ideal testing ground for exploring and applying these advanced artificial intelligence technologies. It holds the promise of optimizing clinical workflows, enhancing the quality of diagnosis and treatment decisions, and improving patient communication. We systematically searched PubMed and Embase from January 2023 to September 2025 for studies investigating LLMs in spinal diseases. Original research articles published in English with a Journal Impact Factor (JIF) ≥ 3.0 were included. Reviews, case reports, animal studies, and non-orthopedic topics were excluded. Data from eligible studies were extracted and narratively synthesized. This review aims to systematically and comprehensively examine the current state of clinical applications of medical large models and related intelligent systems in the field of spinal diseases. The focus is on analyzing their core technical pathways, specific clinical application scenarios, and their medical value, and performance evaluation results, thereby identifying current opportunities and key challenges. Furthermore, it anticipates future developments, from leveraging general-purpose models to constructing specialized models based on high-quality, large-scale, multimodal spine-specific datasets. The translational potential of this article : The translational potential of this article lies in its provision of a comprehensive roadmap and practical framework for implementing artificial intelligence in spinal surgery. It systematically synthesizes core application scenarios for large language models—including clinical documentation assistance and preoperative planning—while explicitly addressing four critical challenges requiring resolution for successful clinical integration: regulatory compliance, data privacy protection, algorithmic bias mitigation, and workflow integration. It establishes an actionable foundation for collaborative efforts among clinicians, developers, and policymakers to deploy safe, effective, and compliant AI tools in spinal care.