Abstract
Accurate differentiation between angina with no obstructive coronary arteries (ANOCA) and mental stress-induced myocardial ischemia (MSIMI) is crucial for tailored treatment strategies, yet public data scarcity hampers understanding. Given the higher incidence of both conditions in women, this study prospectively enrolled 80 female ANOCA and 39 age-matched female controls, subjecting them to three types of mental stress tasks. ECGs were continuously monitored across Rest, Stress, and Recover stages of the mental stress tasks, with PET/CT imaging during the Stress stage to evaluate myocardial perfusion. With PET/CT serving as the gold standard for MSIMI diagnosis, 35 of the 80 ANOCA patients were diagnosed as MSIMI. Using ECG variables from different stages of mental stress tasks, we developed five machine learning models to diagnose MSIMI. The results showed that ECG data from different stages provide valuable information for MSIMI classification. Additionally, the dataset encompassed demographic details, physiological, and blood sample test results of the patients. We anticipate this new dataset will significantly push further progress in ANOCA and MSIMI research.