Diagnosis of non-ST-elevation acute myocardial infarction (AMI) by a 12-lead electrocardiogram has poor sensitivity and specificity and, therefore, relies on biochemical markers of myocardial necrosis, which can only be reliably detected within 14 to 16 hours from symptom onset. The body surface map (BSM) improves AMI detection but is limited by its interpretation by inexperienced medical staff. To facilitate interpretation, an automated BSM algorithm was developed and is evaluated in this study. One hundred three patients with ischemic-type chest pain were recruited for this study from December 2001 to April 2002. A 12-lead electrocardiogram (Marquette Mac 5K) and BSM (PRIME-ECG) were recorded at initial presentation, and cardiac troponin I and/or creatine kinase-MB levels measured at 12 hours after symptom onset. The admitting physician's 12-lead electrocardiographic (ECG) interpretation, 12-lead ECG algorithm (Marquette 12 SL V233) diagnosis, and BSM algorithm diagnosis were documented for each patient. AMI, defined by elevation of troponin I to >1 microg/L and/or creatine kinase-MB to >25U/L, occurred in 53 patients. The admitting physician diagnosed 24 patients with AMI (sensitivity 45%, specificity 94%), the 12-lead ECG algorithm diagnosed 17 patients with AMI (sensitivity 32%, specificity 98%), and the BSM algorithm diagnosed 34 patients with AMI (sensitivity 64%, specificity 94%). The BSM algorithm improved the diagnostic sensitivity by 2.0 (p <0.001) and 1.4 (p = 0.002) compared with the 12-lead ECG algorithm or the admitting physician, respectively. There was no significant difference in specificity. Thus, the BSM algorithm improves detection of AMI compared with the 12-lead ECG algorithm or physician's 12-lead ECG interpretation.