Development of an automated Selvester Scoring System for estimating the size of myocardial infarction from the electrocardiogram

Although the Selvester Scoring System for estimating the size of myocardial infarction from the standard 12-lead electrocardiogram (ECG) has potential clinical value, it has found limited application because of the difficulties in making precise and reproducible measurements. The objective of this study was to develop software to automate the Selvester Scoring System, thus allowing wider application of the technique. The study was carried out using a training set consisting of ECG data recorded from 705 individuals with and without previous myocardial infarction. Algorithms for the 50 criteria in the Selvester Scoring System were iteratively improved by comparison of scores obtained by 2 experienced cardiologist investigators with those generated by the program. The final version was evaluated in a test set consisting of 60 ECGs by comparing scores derived by cardiologist investigator with those obtained by the program. The disagreements occurred only in 1.1% of the score comparisons and in 1.6% of the specific measurements. In all cases in which a disagreement occurred, it resulted from very small differences in measurements. These results indicate that the algorithm for automated application of the Selvester Scoring System is adequate for both clinical and research applications.