首页|期刊导航|外科理论与实践|人工智能多组学助力甲状腺癌诊疗的应用与展望

人工智能多组学助力甲状腺癌诊疗的应用与展望OA

Application and prospects of AI-assisted multi-omics for diagnosis and treatment of thyroid cancer

中文摘要英文摘要

甲状腺癌作为最常见的内分泌系统恶性肿瘤,其精准诊疗正经历由人工智能(AI)与多组学技术共同驱动的深刻变革.面对超声判读主观性强、细胞学诊断不确定、滤泡状甲状腺癌术前鉴别困难及复发风险评估缺乏精准指标等长期难题,该领域已从依赖单一信息维度与人工经验判断的传统模式,逐步迈向数据驱动、多维整合的精准诊疗新阶段.首先,AI技术凭借深度学习与计算机视觉优势,显著提升了超声影像智能判读的准确性与效率,推动结节良恶性鉴别从经验主导转向AI辅助决策.其次,基因组学与蛋白质组学等多组学技术系统揭示了甲状腺结节相关的关键驱动基因和分子标志物,为不确定性结节的精准分类提供了重要补充信息.同时,整合基因组学、蛋白质组学等多层次信息的AI模型在诊断效能与复发风险预测方面展现出独特优势,为个体化诊疗决策与全程病人监测管理提供了新思路.总之,甲状腺结节AI驱动多组学研究已从单一技术探索逐步走向多维融合、面向临床转化的创新阶段.

Thyroid cancer,the most common malignancy of the endocrine system,is undergoing a profound transformation in precision diagnosis and treatment driven jointly by artificial intelligence(AI)and multi-omics technologies.Confronted with longstanding challenges,including the strong subjectivity of ultrasound interpretation,diagnostic uncertainty in cytology,difficulty in preoperative differentiation of follicular neoplasms,and the lack of precise indicators for recurrence risk assessment,this field is gradually shifting from a traditional paradigm reliant on a single information dimension and empirical judgment toward a data-driven,multidimensional integrative approach to precision diagnosis and treatment.First,leveraging the advantages of deep learning and computer vision,AI technologies have significantly improved the accuracy and efficiency of intelligent ultrasound image interpretation,shifting nodule malignancy differentiation from experience-dominated practice toward AI-assisted decision-making.Second,multi-omics technologies such as genomics and proteomics have systematically revealed key driver genes and molecular biomarkers associated with thyroid nodules,providing important supplementary information for the precise classification of indeterminate nodules.Meanwhile,AI models integrating multilevel information from genomics,proteomics,and other omics layers have demonstrated distinct advantages in diagnostic performance and recurrence risk prediction,offering new insights for individualized treatment decision-making and continuous patient monitoring.Overall,research on AI-driven multi-omics approaches in thyroid nodules has progressed from isolated technical exploration toward multidimensional integration aimed at clinical translation.

蔡雪;王瑛睿;郭天南

西湖大学医学院、未来产业生物医学学院、医学蛋白质组学全国重点实验室,西湖大学附属杭州市第一人民医院||西湖智能蛋白质组学中心,西湖生命科学与生物医学实验室||西湖大学生命科学学院,未来产业研究中心,浙江 杭州 310012西湖大学医学院、未来产业生物医学学院、医学蛋白质组学全国重点实验室,西湖大学附属杭州市第一人民医院||西湖智能蛋白质组学中心,西湖生命科学与生物医学实验室||西湖大学生命科学学院,未来产业研究中心,浙江 杭州 310012西湖大学医学院、未来产业生物医学学院、医学蛋白质组学全国重点实验室,西湖大学附属杭州市第一人民医院||西湖智能蛋白质组学中心,西湖生命科学与生物医学实验室||西湖大学生命科学学院,未来产业研究中心,浙江 杭州 310012

医药卫生

人工智能多组学甲状腺结节精准诊疗

Artificial intelligence(AI)Multi-omicsThyroid nodulePrecision diagnosis and treatment

《外科理论与实践》 2026 (4)

300-305,6

10.16139/j.1007-9610.2026.04.04

评论