Perioperative cardiac risk prediction

In this study, which is based on data collected from 1351 referred patients, we used myocardial perfusion single-photon emission computed tomography (SPECT) for risk stratification before non-cardiac surgery. We generated binary classifiers using support vector machine (SVM) and conventional linear models for predicting perioperative cardiac events. Input data comprising clinical and surgical risk factors as well as SPECT findings was interpreted to predict the occurrence of all and hard cardiac events. Our SVM models achieved 0.884 and 0.892 in terms of AUC (area under the receiver operating characteristic curve) in predicting all cardiac events and hard events respectively, and were shown to be superior to linear models for risk stratification. We also found an incremental prognostic value of SPECT results over information about clinical and surgical risk.