Clinical data versus noninvasive testing as an estimate of coronary status.

ABSTRACT

The step from coronary angiography to surgery is taken primarily on the basis of the ejection fraction (EF) and number of diseased vessels (NV), taken jointly as coronary status (CS). This study reviewed (1) how well EF, NV and CS are estimated from clinical data, (2) how much improvement is gained by adding noninvasive data, (3) how good this model is when applied to a larger cohort and (4) whether the decision to angiography can be made on the basis of CS estimated from clinical data alone. A quantitative definition of CS was established as a principal component of EF and NV. Estimates of EF, NV and CS were made from clinical data of 60 patients. Correlations with the actual values were EF 87%, NV 93% and CS 93%. When noninvasive data were added to this model, the correlations increased to EF 95%, NV 99% and CS 98%. When the clinical data model was applied to 169 patients, the correlations decreased to EF 77%, NV 71% and CS 74%. CS estimated from clinical data alone was set up as a test for angiography for 169 patients. This test had a sensitivity of 98% and a specificity of 63%, demonstrating the feasibility of using clinical data alone in deciding angiography.