Application of machine learning algorithms to predict heat-sensitive angina (HSA) attacks: a multicentric observational cohort study

Jincheng Wang, Conghui Zhou, Yue Zhao, Jingqing Hu

Background The association between air temperature and angina has been confirmed by several studies, which show that some patients with cardiovascular disease are “heat sensitive” and experience angina more frequently in high-temperature environments. Although several predictive models for cardiovascular risk stratification and angina-related outcomes have been reported, predictive tools specifically designed to identify susceptibility to heat-sensitive angina (HSA) exacerbation under hot weather conditions remain limited. Methods The derivation cohort consisted of 1,246 individuals with stable angina treated at 43 clinical research centers in China. Variable selection was performed using the Boruta algorithm. Seven machine learning algorithms were developed and evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, F1 score, calibration analysis, and decision curve analysis. An independent external validation cohort comprising 120 patients from 5 additional clinical research centers was used to assess model generalizability. Furthermore, we conducted a post hoc analysis of a multi-center clinical trial to validate the value of the model in guiding clinical strategies. Results Through variable selection, 14 predictors were included as the risk factors to develop ML model. A random forest (RF) model demonstrated strong performance during the training and internal validation cohorts. External validation further supported the RF model's predictive performance and robustness. Exploratory post hoc analysis suggested that patients identified by the RF model as having high HSA probability exhibited increased angina frequency during summer months. Conclusions This study innovatively employs region, MPA, DBP, BMI, constipation, body fat distribution indicators, and seven serological markers related to inflammation, lipid metabolism to construct an RF model. The proposed RF model demonstrated promising predictive performance for identifying patients with potential susceptibility to HSA. The model may serve as an exploratory decision-support tool for individualized heat-related cardiovascular risk assessment, although further prospective multicenter validation is required before routine clinical application. Clinical Trial Registration https://www.chictr.org.cn/showproj.html?proj=166685 , identifier ChiCTR2200060267 and https://clinicaltrials.gov/study/NCT02967718?term=NCT02967718&rank=1 , identifier NCT02967718.