According to this paper it can help despite recent teaching which says it is unreliable:
ST-segment elevation predicted AMI with 88% sensitivity and 84% specificity. The criterion including ST-segment elevation and/or depression had 95% sensitivity and 62% specificity. The combined criterion including ST-segment elevation and/or depression, and/or non-specific wide QRS complex and/or left bundle branch block provided a sensitivity and negative predictive value of 100%, a specificity of 46% and a positive predictive value of 52%.
So reliance on classic STEMI criteria is not enough because it will miss at least 22% of patients who need to go to the lab. On the other hand the use of combined criteria (meaning the patient's EKG has to be nearly stone-cold normal to avoid a trip to the lab) has very high sensitivity.