[Applications and prospects of artificial intelligence for decision-making in spinal surgery]

The application of artificial intelligence in decision support for spinal surgery has developed rapidly and now widely covers key steps such as automatic segmentation and classification of spinal imaging, determination of surgical indications, pedicle screw trajectory planning, and prediction of postoperative prognosis and complication risks. However, most current studies remain limited to single-center retrospective analyses, and the external validation, interpretability, and clinical implementation of these models are still inadequate. In the future, it will be necessary to leverage multi-center, multimodal, high-quality clinical data resources to strengthen the evaluation of model robustness and clinical effectiveness, improve assessment frameworks and reporting standards, and promote the deep integration of artificial intelligence with intraoperative navigation systems, robotic platforms, and wearable devices, thereby gradually transforming experience-based decisions into data-driven decision-making and prediction, so as to better assist spine surgeons in making safe, precise, and clinically implementable decisions.