| With the rapid advancement of artificial intelligence (AI), large language models (LLMs) have gained increasing attention in the medical field. As a critical component of clinical medicine, perioperative management encompasses preoperative assessment, intraoperative monitoring, and postoperative recovery. Effective perioperative care is vital for reducing surgical risks, improving patient satisfaction, and enhancing postoperative outcomes and quality of life. While traditional perioperative approaches have achieved certain success, challenges persist, including inefficient data processing, insufficient personalized treatment, suboptimal patient communication, and inaccurate prognostic evaluations. In recent years, LLMs have demonstrated potential in addressing these challenges through their robust natural language processing and data analysis capabilities. This review explores the latest advancements in LLMs for perioperative applications, including preoperative evaluation, intraoperative monitoring, postoperative care, and clinician-patient communication. It highlights their advantages in enhancing diagnostic efficiency, improving patient-provider interactions, and optimizing personalized care. However, challenges such as data privacy, clinical reliability, and ethical concerns are also discussed. By analyzing existing literature, this review evaluates the practical impact of LLMs on perioperative management and envisions their future role in transforming perioperative care. Further research and clinical validation are advocated to maximize their potential in patient-centered healthcare. |
