AIMS: Transmural extent (TME) of myocardial scar, contractile reserve, and perfusion all predict improvement in regional myocardial function after coronary revascularization. We sought their association with regional remodelling after infarction. METHODS AND RESULTS: We studied 89 patients (age 62 +/- 10 years) with left ventricular (LV) dysfunction, at least 1 month post infarction. Viability was identified by TME < 75% on contrast-enhanced magnetic resonance imaging (ce-MRI), augmentation at low-dose dobutamine echocardiography (DbE), or >60% uptake on delayed redistribution on TI-201 SPECT (single photon emission computed tomography). Coronary revascularization was performed in 36 patients. Regional LV end-diastolic volume (EDV) and end-systolic volume, and ejection fraction were measured with MRI at baseline and after a median follow-up of 18 months. Of 357 segments identified with subendocardial infarction (TME 0-25%) on ce-MRI, 176 were revascularized. Subendocardial scar segments were associated with reverse regional remodeling during follow-up. Revascularization was an independent correlate of change in EDV, but TME and revascularization showed no interaction with respect to their influence on regional volumes. Contractile reserve was present on DbE in 228 segments, of which 129 were TME 0-25%; remodelling was associated with intervention in non-transmural infarcts showing viability by DbE. Viability was identified by TI-201 SPECT in 381 segments (233 with TME 0-25%), but viability by SPECT was not associated with reverse remodelling. No significant reverse remodelling occurred in segments with intermediate scar thickness (TME 26-75%) or transmural scar, independent of revascularization or viability by DbE or TI-SPECT. CONCLUSION: Reverse regional remodelling is associated with subendocardial infarction, especially in the setting of contractile reserve and revascularization.