Objectives: To evaluate the accuracy of 2D and 3D strain by speckle tracking in detecting early and subclinical changes in myocardial contractility during ischemia. Methods: Twenty consecutive patients (mean age 65±11 years, mean LV ejection fraction 55±7%) with severe coronary artery stenosis (12 left anterior descending) referred for percutaneous coronary intervention were included. A comprehensive 2D and 3D echocardiography data was recorded before and during PCI (mean time occlusion 51±14 seconds). 2D and 3D speckle tracking analysis were computed only from data acquired before visual wall motion changes. After PCI, intracoronary ultrasound contrast agent was used to accurately delineate ischemic segments. Results: Global longitudinal strain by 2D and 3D decreased during PCI (+1.7%, P=0.01 for 2D and +1.6%, P=0.03 for 3D longitudinal strain). Similar changes were also observed for 3D circumferential (+2.2%, P=0.02), 3D area (+2.9%, P=0.009) and 3D radial global strain (-5.3%, P = 0.04). In addition, regional systolic peak strain by 2D and 3D speckle tracking also decreased in ischemic segments for longitudinal (+2.5% vs. +0.1%, P=0.01 for 2D and +1.7% vs. +1.0%, P=0.03 for 3D), circumferential (+2.2% vs +0.2%, P=0.01) and area strain component (+3.0% vs +0.3%, P=0.009). In contrast, time to peak longitudinal strain was prolonged during ischemia only for 2D (+55ms vs +13ms, P=0.0001) and not for 3D strain. Conclusions: Global and regional peak strain by 2D and 3D speckle tracking are sensitive to detect early ischemia before visual wall motion changes. However, 3D volume rate remains insufficient to identify post systolic shortening during PCI.