Central Illustration.

3- and 12-lead ECG Detection of Coronary Occlusion During Coronary Balloon Angioplasty

ST-segment shift computer analysis for ischemia detection (inflation) with and without using vectorcardiographic reconstruction (VCG) and blind consensus human reading performed in “comparative” (baseline/inflation, baseline/preinflation comparison) and “spot” (preinflation, inflation without baseline information) modes. Areas under the ROC curves (AUC) along with points representing “STEMI standard definition” (black square) and human consensus readings (2 independent “spot”, open triangles 1 and 2, and 1 “comparative, solid triangle) are presented. All ECG analyses in “comparative” mode were superior to “spot” mode (P < 0.05 for AUC differences) without significant difference between “abc” and 12 leads. Human “spot” readings 1 and 2 highlight low reproducibility of human reading. ECG = electrocardiogram; STEMI = ST-segment elevation myocardial infarction