《生命科学》 2025, 37(12): 1624-1633

基于人工智能分析肿瘤基因的个性化治疗研究

王亚军 , 陈　瑛 , 杨得军 , 雷新刚^*

上海慧医钛康智能技术有限公司，上海 200126

摘　要：

肿瘤基因组学数据的快速累积与人工智能技术的进步为个性化癌症治疗提供了前所未有的发展契机。本综述系统梳理了人工智能在肿瘤基因数据分析中的应用现状、关键技术及挑战。文章首先概述了肿瘤基因数据的特征及分析难点，详细阐述了人工智能在基因变异识别、致病性预测、风险分层、治疗推荐等方面的典型应用；随后探讨了大语言模型、多模态深度学习等前沿技术在肿瘤精准医疗中的创新应用，并分析了临床实践中面临的数据隐私、模型可解释性、临床转化障碍等关键挑战；最后，本文提出了加强多学科融合、改进模型可解释性、建立规范验证机制等解决方案，为人工智能驱动的精准肿瘤医学提供了发展路径。

通讯作者：雷新刚 , Email:dustin.lei@taihealth.cn

Research on personalized treatment of tumors based on artificial intelligence analysis of tumor genes

WANG Ya-Jun , CHEN Ying , YANG De-Jun , LEI Xin-Gang^*

Shanghai Taihealth Intelligent Technology Co., Ltd., Shanghai 200126, China

Abstract:

The rapid accumulation of tumor genomics data and the advancement of artificial intelligence (AI) technology have brought revolutionary opportunities for personalized cancer treatment. This review systematically summarizes the current applications of AI in tumor genomics data analysis, key technologies, and challenges. The article first outlines the characteristics and analytical difficulties of tumor genomics data, and then details the typical applications of AI in gene variant identification, pathogenicity prediction, risk stratification, and treatment recommendation. It further explores the innovative applications of large language models and multimodal deep learning in precision oncology and analyzes key challenges in clinical practice, such as data privacy, limited model interpretability, and barriers to clinical translation. Finally, the article proposes solutions to strengthen multidisciplinary integration, improve model interpretability, and establish standardized validation mechanisms, providing a development path for AI-driven precision oncology.

Communication Author：LEI Xin-Gang , Email:dustin.lei@taihealth.cn