| 35 | 0 | 24 |
| 下载次数 | 被引频次 | 阅读次数 |
治疗效果异质性（HTE）分析是对临床研究中不同患者群体治疗效果差异性表现的系统探索，有助于推动个体化治疗策略的实施。机器学习方法对复杂交互作用的识别能力为HTE更准确、高效地获得治疗优势人群提供了可行路径。系统梳理了目前临床研究中HTE分析常用的机器学习模型，归纳为惩罚回归模型、因果树模型、贝叶斯模型和元学习模型四种类型；阐述了不同类别机器学习模型的优缺点及在HTE分析中的适用数据类型，探讨了不同机器学习方法在中医药临床研究中的可能应用场景，以期为在中医药临床研究中应用HTE分析提供方法学支持。
Abstract：Heterogeneity treatment effect(HTE) analysis systematically examines variations in treatment responses across different patient populations in clinical research, enabling the advancement of personalized treatment strategies. Machine learning methods, with their ability to identify complex interactions, provide an effective approach for more accurately and efficiently identifying responsive populations in HTE analysis. This article systematically reviewed machine learning models commonly used in HTE analysis within clinical research, categorizing them into four types, i. e. penalized regression models, causal tree models, Bayesian models, and meta-learning models. We also outlined the advantages, disadvantages, and suitable data types for each category within HTE analysis, and explored potential application scenarios for different machine learning methods in clinical research of traditional Chinese medicine(TCM), so as to provide methodological support for implementing HTE analysis in TCM clinical research.
[1]PACANOWSKI M, LIU Q. Precision medicine 2030[J].Clin Pharmacol Ther,2020,107(1):62-64.
[2]虞梦婷.中医药精准医疗理念在消化道肿瘤防治中的应用价值[J].中医药管理杂志，2023,31(23):138-140.
[3]KENT DM, PAULUS JK, VAN KLAVEREN D, et al.The predictive approaches to treatment effect heterogeneity(PATH)statement[J]. Ann Intern Med,2020,172(1):35-45.
[4]LIPKOVICH I, DMITRIENKO A, D'AGOSTINO SRBR.Tutorial in biostatistics:data-driven subgroup identification and analysis in clinical trials[J]. Stat Med,2017,36(1):136-196.
[5]YARNELL CJ, FRALICK M. Heterogeneity of treatment effect:an evolution in subgroup analysis[J]. NEJM Evid,2024,3(4):EVIDe2400054.
[6]李宇飞，张文倩，张颖，等.中医药临床试验亚组分析设计和报告建议[J/OL].中国中西医结合杂志.（2025-06-17）[2025-08-29]. https：//link. cnki. net/urlid/11.2787. R. 20250616. 1656. 004.
[7]INIESTA R, STAHL D, MCGUFFIN P. Machine learning,statistical learning and the future of biological research in psychiatry[J]. Psychol Med,2016,46(12):2455-2465.
[8]SUSUKIDA R, AMIN-ESMAEILI M, BADILLO-GOICOECHEA E, et al. Application of causal forest model to examine treatment effect heterogeneity in substance use disorder psychosocial treatments[J]. Int J Methods Psychiatr Res,2025,34(1):e70011.
[9]GILBERT JB. Modeling item-level heterogeneous treatment effects:a tutorial with the glmer function from the lme4 package in R[J]. Behav Res Methods,2024,56(5):5055-5067.
[10]YANG J, MWANGI AW, KANTOR R, et al. Tree-based subgroup discovery using electronic health record data:heterogeneity of treatment effects for DTG-containing therapies[J]. Biostatistics,2024,25(2):323-335.
[11]BROOKS JM, CHAPMAN CG, CHEN BK, et al. Assessing the properties of patient-specific treatment effect estimates from causal forest algorithms under essential heterogeneity[J]. BMC Med Res Methodol,2024,24(1):66.
[12]SHILLAN D, STERNE JAC, CHAMPNETS A, et al.Use of machine learning to analyse routinely collected intensive care unit data:a systematic review[J]. Crit Care,2019,23(1):284.
[13]HU L, JI J, LI F. Estimating heterogeneous survival treatment effect in observational data using machine learning[J]. Stat Med,2021,40(21):4691-4713.
[14]WILLKE RJ, ZHENG Z, SUBEDI P, et al. From concepts, theory, and evidence of heterogeneity of treatment effects to methodological approaches:a primer[J]. BMC Med Res Methodol,2012,12(1):185.
[15]TIBSHIRANI R. Regression shrinkage and selection via the Lasso[J]. J R Stat Soc Series B Stat Methodol,1996,58(1):267-288.
[16]HOERL AE, KENNARD RW. Ridge regression:biased estimation for nonorthogonal problems[J]. Technometrics,1970,12(1):55-67.
[17]HUI ZOU, TREVOR HASTIE. Regularization and variable selection via the elastic net[J]. J R Statist Soc B,2005,67(2):301-320.
[18]ATHEY S, IMBENS G. Recursive partitioning for heterogeneous causal effects[J]. Proc Natl Acad Sci U S A,2016,113(27):7353-7360.
[19]WAGER S, ATHEY S. Estimation and inference of heterogeneous treatment effects using random forests[J]. J Am Stat Assoc,2018,113(523):1228-1242.
[20]PEARL J. Causality:Models, Reasoning, and Inference[M]. Cambridge:Cambridge University Press, 2009:14.
[21]CHIPMAN HA, GEORGE EI, MCCULLOCH RE. BART:Bayesian additive regression trees[J]. ANN APPL STAT,2010. doi:10. 1214/09-AOAS285.
[22]HAHN PR , MURRAY JS , CARVALHO CM. Bayesian regression tree models for causal inference:regularization,confounding,and heterogeneous effects(with discussion)[J]. Bayesian Anal,2020,15(3):965-1056.
[23]HOSPEDALES T, ANTONIOU A, MICAELLI P, et al.Meta-learning in neural networks:a survey[J]. IEEE Trans Pattern Anal Mach Intell,2022,44(9):5149-5169.
[24]SALDITT M, ECKES T, NESTLER S. A tutorial introduction to heterogeneous treatment effect estimation with meta-learners[J]. Adm Policy Ment Health,2024,51(5):650-673.
[25]李牧之，王思村，党海霞，等.以患者为中心研究模式与适应性设计在中医药临床研究中的实践[J].世界临床药物，2024,45(12):1213-1218.
[26]祝世讷.中医学原理探究[M].北京：中国中医药出版社，2019:174.
[27]王天芳.中医辨证论治学基础[M].北京：中国中医药出版社，2016:2.
[28]许博文，张潇潇，李杰，等.基于优势人群的肿瘤中医药防治疗效提升策略[J].中医杂志，2022,63(4):318-322.
[29]DU Y, CHEN H, VARADHAN R. Lasso estimation of hierarchical interactions for analyzing heterogeneity of treatment effect[J]. Stat Med,2021,40(25):5417-5433.
[30]SUN C, ZHANG M, QIAO Q, et al. Integrating intrinsic radiosensitivity and immune status for predicting benefits of radiotherapy in head and neck squamous cell carcinoma[J]. Med Sci Monit,2021,27(1):e932126.
[31]IRMAK?N B, HAVLA J, MANSMANN U. Multivariable prognostic prediction of efficacy and safety outcomes and response to fingolimod in people with relapsingremitting multiple sclerosis[J]. Mult Scler Relat Disord,2025,93(1):106247.
[32]ALI H, SHAHZAD M, SARFRAZ S, et al. Application and impact of Lasso regression in gastroenterology:a systematic review[J]. Indian J Gastroenterol,2023,42(6):780-790.
[33]VENKATASUBRAMANIAM A, MATEEN BA, SHIELDS BM,et al. Comparison of causal forest and regressionbased approaches to evaluate treatment effect heterogeneity:an application for type 2 diabetes precision medicine[J]. BMC Med Inform Decis Mak,2023,23(1):110.
[34]TAGUCHI A, KATO K, FURUSAWA A, et al. Heterogeneous treatment effect of dose-dense paclitaxel plus carboplatin therapy for advanced ovarian cancer[J]. Int J Cancer,2024,155(6):1068-1077.
[35]SHIBA K, INOUE K. Harnessing causal forests for epidemiologic research:key considerations[J]. Am J Epidemiol,2024,193(6):813-818.
[36]JAWADEKAR N, KEZIOS K, ODDEN MC, et al. Practical guide to honest causal forests for identifying heterogeneous treatment effects[J]. Am J Epidemiol,2023,192(7):1155-1165.
[37]CARON A, BAIO G, MANOLOPOULOU I. Shrinkage bayesian causal forests for heterogeneous treatment effects estimation[J]. J Comput Graph Stat,2022,31(4),1202-1214.
[38]BLETTE BS, GRANHOLM A, LI F, et al. Causal bayesian machine learning to assess treatment effect heterogeneity by dexamethasone dose for patients with COVID-19 and severe hypoxemia[J]. Sci Rep,2023,13(1):6570.
基本信息:
DOI：10.13288/j.11-2166/r.2025.21.004
中图分类号:R24
引用信息:
[1]李牧之,王思村,和晓珺,等.机器学习方法在治疗效果异质性分析中的应用及中医药领域适用性探讨[J].中医杂志,2025,66(21):2199-2203.DOI:10.13288/j.11-2166/r.2025.21.004.
基金信息:
国家重点研发计划（2025YFC3507905）; 中国中医科学院基本科研业务费优秀青年科技人才（创新类）培养专项（ZZ13-YQ-076）; 国家自然科学基金（82105052,82474682）