A gender-specific blood-based gene expression score for assessing obstructive coronary artery disease in nondiabetic patients: results of the Personalized Risk Evaluation and Diagnosis in the Coronary Tree (PREDICT) trial.

BACKGROUND Currently available noninvasive tests to risk stratify patients for obstructive coronary disease result in many unnecessary cardiac catheterizations, especially in women. We sought to compare the diagnostic accuracy of presenting symptoms, noninvasive test results, and a gene expression score (GES) in identifying obstructive coronary artery disease (CAD) according to gender, using quantitative coronary angiography as the criterion standard. METHODS The PREDICT trial is a prospective multicenter observational study designed to develop and validate gene expression algorithms to assess obstructive CAD, defined as at least one ≥50% diameter stenosis measured by quantitative coronary angiography. Patients referred for diagnostic cardiac catheterization with suspected but previously unknown CAD were enrolled. Noninvasive myocardial perfusion imaging (MPI) was available in 60% of patients. The GES, comprising gender-specific age functions and 6 gene expression terms containing 23 genes, was performed for all patients. RESULTS A total of 1,160 consecutive patients (57.6% men and 42.4% women) were enrolled in PREDICT. The prevalence of obstructive CAD was 46.7% in men and 22.0% in women. Chest pain symptoms were a discriminator of obstructive CAD in men (P < .001) but not in women. The positive predictive value of MPI was significantly higher in men (45%) than in women (22%). An abnormal site-read MPI was not significantly associated with obstructive or severity of CAD. The GES was significantly associated with a 2-fold increase in the odds of obstructive CAD for every 10-point increment in the GES and had a significant association with all measures of severity and burden of CAD. By multivariable analysis, GES was an independent predictor of obstructive CAD in the overall population (odds ratio [OR] 2.53, P = .001) and in the male (OR 1.99, P = .001) and female (OR 3.45, P = .001) subgroups separately, whereas MPI was not. CONCLUSIONS Commonly used diagnostic approaches including symptom evaluation and MPI performed less well in women than in men for identifying significant CAD. In contrast, gender-specific GES performed similarly in women and men. Gene expression score offers a reliable diagnostic approach for the assessment of nondiabetic patients and, in particular, women with suspected obstructive CAD.