Non-invasive diagnosis of acute myocardial infarction (AMI) associated with significant left main stem (LMS) stenosis remains challenging. Consecutive patients presenting with acute ischaemic-type chest pain from 2000 to 2010 were analysed. Entry criteria: 12-lead ECG and Body Surface Potential Map (BSPM) at presentation, cardiac troponin T (cTnT) ≥12 h and coronary angiography during admission. cTnT ≥0.03 µg/l defined AMI. ECG abnormalities assessed: STEMI by Minnesota criteria; ST elevation (STE) aVR ≥0.5 mm; ST depression (STD) ≥0.5 mm in ≥2 contiguous leads (CL); T-wave inversion (TWI) ≥1 mm in ≥2 CL. BSPM STE was ≥2 mm in anterior, ≥1 mm in lateral, inferior, right ventricular or high right anterior and ≥0.5 mm in posterior territories. Significant LMS stenosis was ≥70%. Enrolled were 2810 patients (aged 60 ± 12 years; 71% male). Of these, 116 (4.1%) had significant LMS stenosis with AMI occurring in 92 (79%). STEMI by Minnesota criteria occurred in 13 (11%) (sensitivity 12%, specificity 92%), STE in lead aVR in 23 (20%) (sensitivity 23%, specificity 92%), TWI in 38 (33%) (sensitivity 34%, specificity 71%) and STD in 51 (44%) (sensitivity 49%, specificity 75%). BSPM STE occurred in 85 (73%): sensitivity 88%, specificity 83%, positive predictive value 95% and negative predictive value 65%. Of those with AMI, 74% had STE in either the high right anterior or right ventricular territories not identified by the 12-lead ECG. C-Statistic for AMI diagnosis using BSPM STE was 0.800 (P < 0.001). In patients with significant LMS stenosis presenting with chest pain, BSPM STE has improved sensitivity (88%), with specificity 83%, over 12-lead ECG in the diagnosis of AMI.