Early angiography may not adequately subgroup patients with myocardial infarction if cyclic changes in coronary flow occur frequently. From a pilot experience using a new 12-lead ST-segment monitor, a continuously updated, self-referenced ST-recovery analysis method was developed to quantify both instantaneous recovery, as a noninvasive marker of patency, and cumulative ST recovery over time, as a marker of the speed, stability and duration of reperfusion. In 22 patients with acute infarction in whom 44 observations of unique angiographic patency were noted within 6 hours of presentation, serial patency assessments simultaneous with all angiographic observations predicted coronary occlusion with 90% sensitivity and 92% specificity. Of the 22 patients, 11 (50%) had multiple ST trend transitions suggesting cyclic changes in coronary flow before catheterization. Speed, stability and duration of ST-segment recovery were defined by the time to first 50% ST recovery, total number of ST-trend transitions and patent physiology index (percentage of monitoring period showing ST recovery), respectively. Subgrouped angiographically, the median (interquartile range) for cumulative ST parameters with patent (n = 8) versus occluded (n = 14) arteries were, respectively--time to 50% recovery, 1.57 (1.16, 1.70) versus 0.17 (-0.47, 0.32) hours; number of reelevation/recovery events, 1.5 (1, 3) versus 3 (1, 3); and patent physiology index, 52 (47, 59) versus 50 (5, 73). Thus, continuous ST-segment recovery analysis appears to predict simultaneous angiographic patency over serial assessments, whereas cumulative parameters appear to contain independent information, probably because of patency changes before or after angiography.