Predictive potential of noninvasive methods, inclusive of exercise SPECT Tc99m MIBI imaging, in recognition of high-risk patients with left main coronary artery stenosis

The aim of the present study was to determine which clinical, exercise and perfusion variables can be instrumental in the identification of left main coronary artery. A multivariate model for prediction of left main disease w as developed, based on a number of clinical, exercise and MIBI perfusion va riables in two groups of patients. Group I consisted of 38 patients (30 men and eight women) with angiographically proven left main stenosis, while gr oup II consisted of 41 patients (27 men and 14 women) with multivessel coro nary artery diseases. A multivariate logistic regression analysis demonstra ted that clinical variables including diabetes, hypertension, together with history of typical angina were the only independent predictors of left mai n stenosis. It was found that p value was 0.05 for hypertension, 0.01 for d iabetes as well as 0.01 for the history of typical angina in clinical exami nation. No exercise or perfusion variables may be instrumental in predictio n of left main stenosis, when considered in isolation. Myocardial perfusion by itself is therefore not deemed sufficiently specific to attempt its pos itive identification. High-risk patients with left main disease can be iden tified noninvasively by exercise perfusion imaging using a model based on t he proposed logistic regression analysis with clinical variables.