Lacson Ronilda C, Baker Bowen, Suresh Harini, Andriole Katherine, Szolovits Peter, Lacson Eduardo
Brigham and Women's Hospital, Boston MA, USA.
Harvard Medical School, Boston, MA, USA.
Clin Kidney J. 2018 Jul 3;12(2):206-212. doi: 10.1093/ckj/sfy049. eCollection 2019 Apr.
We re-analyzed data from the Systolic Blood Pressure Intervention Trial (SPRINT) trial to identify features of systolic blood pressure (SBP) variability that portend poor cardiovascular outcomes using a nonlinear machine-learning algorithm.
We included all patients who completed 1 year of the study without reaching any primary endpoint during the first year, specifically: myocardial infarction, other acute coronary syndromes, stroke, heart failure or death from a cardiovascular event ( = 8799; 94%). In addition to clinical variables, features representing longitudinal SBP trends and variability were determined and combined in a random forest algorithm, optimized using cross-validation, using 70% of patients in the training set. Area under the curve (AUC) was measured using a 30% testing set. Finally, feature importance was determined by minimizing node impurity averaging over all trees in the forest for a specific feature.
A total of 365 patients (4.1%) reached the combined primary outcome over 37 months of follow-up. The random forest classifier had an AUC of 0.71 on the testing set. The 10 most significant features selected in order of importance by the automated algorithm included the urine albumin/creatinine (CR) ratio, estimated glomerular filtration rate, age, serum CR, history of subclinical cardiovascular disease (CVD), cholesterol, a variable representing SBP signals using wavelet transformation, high-density lipoprotein, the 90th percentile of SBP and triglyceride level.
We successfully demonstrated use of random forest algorithm to define best prognostic longitudinal SBP representations. In addition to known risk factors for CVD, transformed variables for time series SBP measurements were found to be important in predicting poor cardiovascular outcomes and require further evaluation.
我们重新分析了收缩压干预试验（SPRINT）的数据，以使用非线性机器学习算法识别预示心血管不良结局的收缩压（SBP）变异性特征。
我们纳入了所有在第一年完成研究且未达到任何主要终点的患者，具体为：心肌梗死、其他急性冠状动脉综合征、中风、心力衰竭或心血管事件导致的死亡（n = 8799；94%）。除临床变量外，确定了代表纵向SBP趋势和变异性的特征，并将其组合到随机森林算法中，使用训练集中70%的患者进行交叉验证优化。使用30%的测试集测量曲线下面积（AUC）。最后，通过最小化森林中所有树的特定特征的节点杂质平均值来确定特征重要性。
在37个月的随访中，共有365例患者（4.1%）达到联合主要结局。随机森林分类器在测试集上的AUC为0.71。自动算法按重要性顺序选择的10个最重要特征包括尿白蛋白/肌酐（CR）比值、估计肾小球滤过率、年龄、血清CR、亚临床心血管疾病（CVD）病史、胆固醇、使用小波变换表示SBP信号的变量、高密度脂蛋白、SBP的第90百分位数和甘油三酯水平。
我们成功证明了使用随机森林算法来定义最佳预后纵向SBP表现。除了已知的CVD危险因素外，发现时间序列SBP测量的转换变量在预测不良心血管结局方面很重要，需要进一步评估。
