An artificial intelligence (AI)-guided single-lead electrocardiogram (EKG) algorithm showed promise in allowing for the detection of possible ST-elevation myocardial infarction (STEMI) using wearable devices, new study results show. C. Michael Gibson, MD, of Beth Israel Deaconess Medical Center, Boston, presented the results Tuesday at Cardiovascular Research Technologies (CRT) 2022 in Washington, D.C., as a CRT First Report. Reducing door-to-balloon time has been the traditional focus for improving care for STEMI patients. However, the study investigators said reducing symptom-to-balloon time, while more challenging because it relies on patient perception and initiative, would provide even more benefit for patients. Therefore, the study, prepared by Sameer Mehta, MD, of the Lumen Foundation, Miami, and colleagues and presented by Gibson, evaluated whether an intelligent and wearable AI-driven single-lead EKG detection algorithm might be a solution. The investigators analyzed EKG records from the Latin America Telemedicine Infarct Network (covering Mexico, Colombia, Argentina and Brazil) from April 2014 to December 2019. The data set included 11,567 12-lead EKG records of 10 seconds in length with a sampling frequency of 500 Hz, including the following balanced classes: angiographically confirmed and unconfirmed STEMI, branch blocks, non-specific ST-T abnormalities, and normal and abnormal (200+ CPT codes, excluding those mentioned above). Cardiologists manually checked the label of each record to ensure precision. The combined STEMI data (testing sample size n=1,156) had 91.2% accuracy, 89.6% sensitivity, and 92.9% specificity. In the confirmed STEMI data set (testing sample size n=723), accuracy was 92.4%, sensitivity was 93.4%, and specificity was 91.4%. The investigators concluded that efforts to consistently improve their AI-guided single-lead EKG algorithm “seemed to continuously yield auspicious results, particularly with Lead V2, which remains consistently superior for STEMI detection throughout our experiments.” “By consistently improving the quality of the model's input, we continue to assess our algorithm's performance and reliability for future clinical validation as a potential remote monitoring and early STEMI detection device,” the investigators said.