We have shown recently that tissue characterization of myocardium with ultrasound reflects changes associated with contractile function throughout the cardiac cycle. To determine whether ultrasonic tissue characterization can sensitively detect the impact of ischemic injury and reperfusion on contractile properties of the heart, we studied the time course of change of backscatter after 5, 20, and 60 min of coronary occlusion followed by reperfusion in 15 dogs. The time-averaged integrated backscatter (IB) and the amplitude and phase of cyclic variation of IB (phase relative to the left ventricular pressure waveform) were measured. A novel ultrasonic index of acute injury was identified, the phase-weighted amplitude of cyclic variation, and calculated by weighting the amplitude of cyclic variation of IB with respect to the phase. We hypothesized that backscatter variables would change dramatically after occlusion and that their restitution after reperfusion would sensitively reflect the extent and time course of reversibility of ischemic injury. After coronary occlusion, segmental wall thickening decreased from approximately 55% to 5% regardless of the duration of ischemia. Changes in backscatter associated with this decrease included an increase in time-averaged IB of approximately 5 dB, a 5 dB decrease in cyclic variation, an 80 degree phase shift, and a 7 dB decrease in phaseweighted amplitude. Wall thickening after reperfusion immediately after the 5, 20, or 60 min occlusions recovered to 45%, 27%, and 12% of baseline values, respectively. Within 3 hr it recovered to 53%, 44%, and 22%. Time-averaged IB recovered initially by 89%, 61%, and 44% (all p < .05) and continued to recover subsequently although more slowly. Ultimate recovery was virtually complete. In contrast to the rapid recovery of time-averaged IB, phase-weighted amplitude recovered initially to only 72%, 41%, and - 7% of baseline (all p < .05) and manifested slower and incomplete recovery when ischemia had been present for 20 or 60 min. After reperfusion, the time course of both cyclic variation and phase were reflected by changes in the phase-weighted amplitude. The backscatter variables assessed appear to sensitively delineate the duration, time course of recovery, and reversibility of ischemic injury in response to reperfusion. The results suggest that early recovery of timeaveraged IB corresponds in part to the restoration of tissue ultrastructural integrity. Accordingly, ultrasonic characterization of reperfused myocardium should be useful in defining the myocardium's response to reflow and to interventions such as thrombolysis designed to enhance salvage of ischemic myocardium.