In patients with ST-segment elevation myocardial infarction treated with fibrinolytics, electrocardiogram-derived measures of ST-segment recovery guide therapy decisions and predict infarct size. The comprehension of these relationships in patients undergoing mechanical reperfusion is limited. We studied 144 patients treated with primary percutaneous coronary intervention. We aimed to define the association between infarct size as determined by cardiac magnetic resonance imaging and different metrics of ST-segment recovery. Electrocardiograms were assessed at baseline and 90 minutes after primary percutaneous coronary intervention. Three methods for calculating and categorizing ST-segment recovery were used: (1) summed ST-segment deviation (STD) resolution analyzed in 3 categories (≥70%, ≥30% to <70%, and <30%); (2) single-lead STD resolution analyzed in the same 3 categories; (3) worst-lead residual STD analyzed in 3 categories (<1 mm, 1 to <2 mm, and ≥2 mm). Infarct size and ejection fraction were assessed at 4 months by cardiac magnetic resonance imaging. All 3 ST-segment recovery algorithms predicted the final infarct size and cardiac function. Worst-lead residual STD performed the same as, or better than, the more complex methods and identified large subgroups at either end of the risk spectrum (median infarct size from the lowest to highest risk category (percentage of left ventricle: 7.7% [interquartile range 10.8], 13.1% [interquartile range 13.6]; 24.6% [interquartile range 21.1]); with adjusted odds ratios for infarct size greater than the median (reference <1 mm): 1 to <2 mm, odds ratio 2.3 (95% confidence interval 0.8 to 5.9); ≥2 mm, odds ratio 6.3 (95% confidence interval 1.7 to 23.7; c-index 0.781). In conclusion, an electrocardiogram obtained early after primary percutaneous coronary intervention analyzed by a simple algorithm provided prognostic information on the final infarct size and cardiac function.