Background: For the detection of coronary artery disease (CAD) with supine bicycle echocardiography (SBE), some laboratories image at rest, during intermediate stages, and peak exercise (RIP), while others image at rest and during peak exercise only (RP). We postulated that using atation of images of RIP protocol improves the diagnostic accuracy of SBE by allowing the detection of subtle ischemia through a biphasic response. Methods and Results: We studied supine bicycle echocariograms of 104 patients (mean age 57 짹 11 years, 37 women). Ninety-one patients underwent both SBE and coronary angiography. The SBE protocol started at 25 W with increments of 25 watts every 3 minutes. We interpreted imaging without intermediate stage (RP) in first session, then reinterpreted n a random order at second session with all stages of exercise displayed (RIP). 86% of patients exercised 75 watts or more; 77% of patients achieved a double product > 20,000. Thirty patients had 1-vessel disease, 31 had multivessel disease. Compared to imaging at RP only, interpretation of RIP imaging improved the sensitivity for the detection of the individual vessel stenosis (RP vs RIP: LAD 64% vs.83%; circumflex 52% vs. 67%; RCA 56% vs. 84%). Specificity was either unchanged or improved slightly. Sensitivity for all vessels and al patients increased form 58% to 78% (p=0.004) and 74% to 94% without a change in specificity. The majority of improvement in detection of CAD with RIP (16 of27 vessels in 27 patients) was in the identification of additional ischemic territories to those detected with RP because of more subtle ischemia assessed with a biphasic response. Conclusion: Interpretation of intermediate stages of exercise in addition to peak exercise improves the diagnostic accuracy of supine bicycle echocardiography in the detection of CAD.