Interpretable machine learning predicts cardiac resynchronization therapy responses from personalized biochemical and biomechanical features.
Interpretable machine learning predicts cardiac resynchronization therapy responses from personalized biochemical and biomechanical features.
复制标题
DOI：
10.1186/s12911-022-02015-0
复制
发表时间：
2022-10-31
影响因子：
3.5
通讯作者：
Richardson, William J.
中科院分区：
文献类型：
--
作者：
Haque, Anamul;Stubbs, Doug;Hubig, Nina C.;Spinale, Francis G.;Richardson, William J.
关键词：
心脏复苏治疗（CRT）是一种广泛用于左心室（LV）衰竭患者的基于器械的治疗。不幸的是，许多患者并没有从CRT中获益，因此在CRT实施之前识别这组无应答者具有潜在价值。过去的研究表明，预测CRT反应将需要不同的变量，包括人口统计学，生物标志物和LV功能数据。因此，本研究的目的是将不同的变量类型整合到机器学习算法中，以预测个体患者对CRT的反应。我们使用之前从SMART-AV CRT临床试验（n = 794例患者）中获得的数据构建了一种集成分类算法。我们对80%的患者（n = 635）进行了五重分层交叉验证，以在0个月（开始CRT前）收集的变量训练模型，其余20%的患者（n = 159）用作模型验证的保留测试集。为了提高模型的可解释性，我们使用SHapley加法解释（SHAP）分析量化了特征重要性值，并使用本地可解释模型不可知解释（LIME）来解释特定于患者的预测。我们的分类算法纳入了26个患者人口统计学和病史变量、12个生物标志物变量和18个LV功能变量，在71%的患者中正确预测了CRT反应。额外的患者分层以确定具有最高或最低应答可能性的亚组，结果显示96%的准确性，在最高和最低应答组的23例患者中有22例正确预测。通过计算整合CRT干预前可用的一般患者特征、合并症、治疗史、循环生物标志物和LV功能数据，可以改善对个体患者反应的预测。在线版本包含补充材料，可通过10.1186/s12911-022-02015-0获取。
Cardiac Resynchronization Therapy (CRT) is a widely used, device-based therapy for patients with left ventricle (LV) failure. Unfortunately, many patients do not benefit from CRT, so there is potential value in identifying this group of non-responders before CRT implementation. Past studies suggest that predicting CRT response will require diverse variables, including demographic, biomarker, and LV function data. Accordingly, the objective of this study was to integrate diverse variable types into a machine learning algorithm for predicting individual patient responses to CRT. We built an ensemble classification algorithm using previously acquired data from the SMART-AV CRT clinical trial (n = 794 patients). We used five-fold stratified cross-validation on 80% of the patients (n = 635) to train the model with variables collected at 0 months (before initiating CRT), and the remaining 20% of the patients (n = 159) were used as a hold-out test set for model validation. To improve model interpretability, we quantified feature importance values using SHapley Additive exPlanations (SHAP) analysis and used Local Interpretable Model-agnostic Explanations (LIME) to explain patient-specific predictions. Our classification algorithm incorporated 26 patient demographic and medical history variables, 12 biomarker variables, and 18 LV functional variables, which yielded correct prediction of CRT response in 71% of patients. Additional patient stratification to identify the subgroups with the highest or lowest likelihood of response showed 96% accuracy with 22 correct predictions out of 23 patients in the highest and lowest responder groups. Computationally integrating general patient characteristics, comorbidities, therapy history, circulating biomarkers, and LV function data available before CRT intervention can improve the prediction of individual patient responses. The online version contains supplementary material available at 10.1186/s12911-022-02015-0.
登录
查看更多内容
DOI：
10.1111/j.1540-8159.2006.00486.x
发表时间：
2006-12-01
影响因子：
1.8
作者：
Achilli, Augusto;Peraldo, Carlo;Puglisi, Andrea
通讯作者：
Puglisi, Andrea
影响因子：
3.7
作者：
Hu, Szu-Yeu;Santus, Enrico;Lindvall, Charlotta
通讯作者：
Lindvall, Charlotta
影响因子：
3.5
作者：
Uddin, Shahadat;Khan, Arif;Moni, Mohammad Ali
通讯作者：
Moni, Mohammad Ali
影响因子：
3.8
作者：
Osmanska, Joanna;Hawkins, Nathaniel M.;Virani, Sean A.
通讯作者：
Virani, Sean A.
影响因子：
37.8
作者：
Ellenbogen, Kenneth A.;Gold, Michael R.;Stein, Kenneth M.
通讯作者：
Stein, Kenneth M.