通讯作者: 翟宁, 519236005@qq.com

DOI：10.12201/bmr.202512.00030

Advances and prospects of artificial intelligence in precision phenotyping of heart failure

• 摘要：心力衰竭(heart failure,HF)是一种异质性极高的临床综合征，传统左心室射血分数(left ventricular ejection fraction,LVEF)分类法已难以满足精准医疗需求。近年，人工智能(artificial intelligence,AI)技术发展，为心衰的精准分型开辟了新途径。AI技术通过无监督学习发掘数据隐藏模式，或通过监督学习验证表型及预测风险，能自动分析临床电子病历(electronic health records,EHR)、心脏影像学和多组学等多模态数据。此举有助于识别新亚型、优化风险分层、匹配个性化治疗，并通过增强模型的可解释性解决临床信任难题。尽管面临数据局限、算法“黑箱”特性和临床转化困境等挑战，AI分型仍正推动HF管理向生物学机制驱动的精准医疗转变。

  关键词： 心力衰竭; 人工智能; 机器学习; 表型分型; 异质性; 精准医疗; 风险分层

   

  Abstract: Heart failure (HF) is a highly heterogeneous clinical syndrome, and the traditional classification based on left ventricular ejection fraction (LVEF) is increasingly unable to meet the demands of precision medicine. In recent years, the rapid development of artificial intelligence (AI) technology has opened new avenues for the precision phenotyping of HF. By leveraging unsupervised learning to uncover hidden data patterns or supervised learning to validate phenotypes and predict risks, AI technologies can automatically analyze multi-modal data, including electronic health records (EHR), cardiac imaging, and multi-omics. These approaches assist clinicians in identifying novel disease subtypes that are difficult to distinguish using traditional methods, optimizing risk stratification for high-risk patients, matching personalized treatments to specific phenotypes, and addressing clinical trust issues by enhancing model interpretability, while also revealing new pathophysiological mechanisms. Although AI phenotyping currently faces challenges such as data limitations, the black box nature of algorithms, and difficulties in clinical translation, it is driving the transformation of HF management toward a precision medicine model driven by biological mechanisms.

  Key words: Heart Failure; Artificial Intelligence; Machine Learning; Phenotyping; Heterogeneity; Precision Medicine; Risk Stratification

  提交时间：2025-12-10

  版权声明：作者本人独立拥有该论文的版权，预印本系统仅拥有论文的永久保存权利。任何人未经允许不得重复使用。
• html

• 图表

• 赵浩楠, 蒋萌萌, 刘蒙伟, 杨李鹏, 薛乾隆. 人工智能在心肺复苏动态反馈及调节中的应用研究进展. 2025. doi: 10.12201/bmr.202510.00042

  周航宇, 李杰. 射血分数保留的心力衰竭合并慢性肾脏疾病的研究进展. 2025. doi: 10.12201/bmr.202504.00077

  蔡欣欣, 吴夏阳. SGLT2i治疗心力衰竭的作用机制及研究进展. 2025. doi: 10.12201/bmr.202507.00052

  李韬, 冯贺霞. 人工智能在医疗健康领域中的创新应用、风险挑战与治理对策. 2025. doi: 10.12201/bmr.202501.00067

  张柄烩, 王珑. 人工智能在针灸腧穴定位与处方优化中的应用进展. 2025. doi: 10.12201/bmr.202508.00053

  郑琰莉, 韩福海, 李舒玉, 苏文星. 人工智能大模型在医疗领域的应用现状与前景展望. 2023. doi: 10.12201/bmr.202312.00027

  .河北大学附属医院消化内科 河北 保定, 杰. 基于深度学习的人工智能技术在胃肠间质瘤中的应用进展. 2024. doi: 10.12201/bmr.202411.00057

  胡博越, 史继红. 人工智能技术在医学图书馆应用的发展趋势与展望. 2025. doi: 10.12201/bmr.202505.00024

  芮晨, 陈玥琪, 李金斌, 张胜发. 人工智能在基层2型糖尿病健康管理中的应用进展与趋势. 2025. doi: 10.12201/bmr.202506.00072

  董琨, 杨芬, 杨扬. 人工智能在全科医师培训中的应用与思考. 2023. doi: 10.12201/bmr.202305.00011

• 序号	提交日期	编号	操作
  1	2025-11-03	10.12201/bmr.202512.00030V1	下载

• 公开评论  匿名评论  仅发给作者 提交评论