《生命科学》 2026, 38(5): 821-835

肿瘤出芽在恶性肿瘤中的研究进展

王龙宇^1,2 , 胡欣然² , 王 远^1,2,3,*

¹锦州医科大学附属第一医院病理科，锦州 121001 ²锦州医科大学基础医学院病理学教研室，锦州 121001 ³锦州医科大学基础医学院生物人类学研究所，锦州 121001

摘　要：

肿瘤出芽（tumor budding，TB）是恶性肿瘤浸润前沿的微观形态特征，已被证实具有明确的预后预测价值。本文系统综述了T B在不同类型实体瘤中的评估标准、分子机制及临床意义，重点探讨T B与上皮-间质转化（epithelial-mesenchymal transition，EMT）及肿瘤微环境（tumor microenvironment，TME）的关联，及其在个体化治疗决策中的潜在作用。近年来，人工智能（artificial intelligence，AI）技术在TB的自动化识别与空间定量分析中展现出显著优势，为克服传统病理评估中标准不统一和主观性强等局限提供了新的解决方案。未来应通过多中心合作建立跨癌种的统一评估体系，并推动其在临床预后分层与个体化治疗中的转化应用。

通讯作者：王 远 , Email:wangy1@jzmu.edu.cn

Research advances in tumor budding in malignant tumors

WANG Long-Yu^1,2 , HU Xin-Ran² , WANG Yuan^1,2,3,*

¹Department of Pathology, The First Affiliated Hospital of Jinzhou Medical University, Jinzhou 121001, China ²Department of Pathology, Basic Medical College, Jinzhou Medical University, Jinzhou 121001, China ³Institute of Biological Anthropology, School of Basic Medical Sciences, Jinzhou Medical University, Jinzhou 121001, China

Abstract:

Tumor budding (TB), defined as the presence of discrete single tumor cells or small clusters of up to four cells at the invasive front, is a key histopathological hallmark of aggressive tumor behavior and a robust, independent prognostic factor across numerous cancer types. This review synthesizes current knowledge on TB, with a focus on its assessment methodologies, biological underpinnings, and clinical relevance. Standardized evaluation is crucial for its application. The  International Tumor Budding Consensus Conference (ITBCC) criteria provide a foundational three-tier grading system, and  now are widely adopted in colorectal cancer. However, its application to other malignancies such as gastric or lung carcinoma  requires adaptation due to histological heterogeneity. While hematoxylin and eosin (HE) staining is routinely used,  immunohistochemistry (IHC) enhances diagnostic accuracy by facilitating the precise identification of epithelial tumor buds. Biologically, TB is closely associated with the epithelial-mesenchymal transition (EMT), a process driven by tumor  microenvironment (TME) signals including hypoxia and crosstalk with cancer-associated fibroblasts. This triggers key  pathways such as TGF-β and Wnt/β-catenin, resulting in the loss of epithelial markers and acquisition of mesenchymal markers, thereby promoting cellular invasion. Furthermore, TB is linked to an immunosuppressive TME, characterized by a reduction in cytotoxic lymphocytes and an increase in regulatory immune cells and programmed death-ligand 1(PD-L1) expression, suggesting its role in facilitating immune evasion. Clinically, high-grade TB consistently correlates with adverse clinic-pathological features including lymph node metastasis and advanced stage, and predicts worse survival outcomes in cancers such as colorectal, gastric, pancreatic, and breast carcinoma. Despite its prognostic strength, the integration of TB into
routine clinical practice faces several challenges: a lack of uniform standards beyond colorectal cancer,   significant interobserver variability, an urgent need for prospective validation, and the integration of artificial intelligence (AI) technologies  into diagnostic workflows. AI offers promising solutions by enabling automated, quantitative, and reproducible TB assessment. Future priorities should include establishing cancer-specific guidelines through multicenter collaborations, validating the predictive value of TB for therapies such as immunotherapy, and developing AI-driven tools for practical  clinical integration. Ultimately, TB holds significant potential to evolve into a precise biomarker for risk stratification and personalized treatment strategies in oncology.

Communication Author：WANG Yuan , Email:wangy1@jzmu.edu.cn