Session S54.4

Detection of Acute Myocardial Ischemia by Vessel-Specific Leads Derived from Reduced Lead Sets

JY Wang*, M Mirmoghisi, JW Warren, GS Wagner, BM Horacek

Philips Healthcare
Andover, MA, USA

Background: Currently used electrocardiographic criteria for identifying patients with ST-elevation myocardial infarction (STEMI) perform with high specificity (SP), but low sensitivity (SE). Our aim was to enhance ischemia-detection ability of conventional ACC/ESC STEMI criteria based on 12-lead electrocardiogram (ECG) by adding new criteria applied to 3 vessel-specific leads (VSLs) derived from subsets of 12-lead ECG using limb leads and dual chest leads.
Methods: Study data consisted of 12-lead ECGs from STAFF3 database acquired during 99 angioplasty-induced ischemic episodes caused by balloon inflation in, respectively, left anterior descending coronary artery (LAD; n = 35), right coronary artery (RCA; n = 47), and left circumflex coronary artery (LCx; n = 17). In each recording, 10-second non-ischemic state and 10-second ischemic state were identified. ST deviation was automatically measured at J point in 12 standard leads, and for 3 VSLs its value was derived from 12-lead ECG by using just 4 predictor leads (a pair of precordial leads combined with a pair of limb leads). Coefficients for derivation of VSLs were developed from an independent dataset (n = 892). Mean values of SE and SP of ischemia detection achieved with conventional versus VSL criteria were obtained from bootstrap trials. The detection of ischemia in VSLs was performed with various threshold values to obtain a complete picture of SE/SP trade-off.
Results: We found that the detection of ischemic state, in any vessel, by means of the ACC/ESC STEMI criteria achieved the SE/SP of 62/96%, i.e., the mean diagnostic performance DP = (SE + SP)/2 was 79%. In comparison, the mean DP of VSLs derived from 8 independent leads of 12-lead ECG was 87% (at SE = SP) and the mean DP of VSLs derived from just 4 predictor leads was 87% with precordial leads V4 & V5; 86% with leads V3 & V5 or V3 & V6 or V4 & V6; 85% with V1 & V3 or V2 & V3 or V3 & V4; 83% with V2 & V4; 82% with V2 & V5 or V2 & V6; 81% with V1 & V4 or V5 & V6; 80% with V1 & V2 or V1 & V5; and 78% with V1 & V6.
Conclusion: Our results demonstrate that VSLs derived from the reduced lead sets using just 6 electrodes can identify, in our test set, acute myocardial ischemia with better diagnostic performance than existing ACC/ESC STEMI criteria based on 12-lead ECG requiring 10 electrodes.

(Abstract Control Number: 20)