Background: We hypothesized that the optimum margin after breast-conserving therapy (BCT) should depend on the original size of the tumor. We propose that "margin index"-a relationship of the margin obtained to the size of the tumor-is a better predictor of residual disease on reexcision than margin alone. Methods: We identified 475 consecutive patients with Stage I-II breast cancer, with or without ductal carcinoma in situ, who were treated with BCT from 1998-2008 who also underwent reexcision for close margins. Margin index was calculated as follows: margin index = closest margin (mm)/tumor size (mm) × 100. A receiver operating curve was created using the derived margin index and the presence or absence of residual disease in the reexcision specimen. Sensitivity and specificity were calculated at various margin indices to determine the optimum margin index. Results: Of the 475 patients, 102 (21%) had residual disease in the reexcision specimen. The optimum margin index was >5; the risk of residual disease for a margin index >5 was only 3.2%. The sensitivity and specificity of a margin index cutoff of 5 was 85 and 73%, respectively. The overall c index for the receiver operating curve was 0.88. The margin index was the only factor predictive of residual disease in multivariate analysis. Conclusions: Margin index is a reliable method for the prediction of residual disease after attempted BCT with close margins. This simple calculation may be helpful for identifying patients who require reexcision before radiation therapy and those who may be able to forego additional surgical interventions.