Automated discrimination of proximal right coronary artery occlusion from middle-to-distal right coronary artery occlusion and left circumflex occlusion in ST-elevation myocardial infarction.

BACKGROUND: Classifying the location of an occlusion in the culprit artery during ST-elevation myocardial infarction (STEMI) is important for risk stratification to optimize treatment. We developed a new logistic regression (LR) algorithm for 3-group classification of occlusion location as proximal right coronary artery (RCA), middle-to-distal RCA or left circumflex (LCx) coronary artery with inferior myocardial infarction. We compared the performance of the new LR algorithm with the recently introduced decision tree classifier of Fiol et al (Ann Noninvasive Electrocardiol. 2004;4:383-388) in the classification of the same 3 categories. METHODS: The new algorithm was developed on a set of electrocardiograms from an emergency department setting (n = 64) and tested on a different set from a prehospital setting (n = 68). All patients met the current STEMI criteria with angiographic confirmation of culprit artery and occlusion location. Using LR, 4 ST-segment deviation features were chosen by forward stepwise selection. Final LR coefficients were obtained by averaging more than 200 bootstrap iterations on the training set. In addition, a separate 4-feature classifier was designed adding ST features of V4R and V8, only available in the training set. RESULTS: The LR algorithm classified proximal RCA occlusion vs combined LCx occlusion and middle-to-distal RCA occlusion, with a sensitivity of 76% and specificity of 81% as compared with 71% and 62% for the Fiol classifier. The difference in specificity was statistically significant. The LR classifier trained with additional ST features of V4R and V8, but still limited to 4, improved the overall agreement in the training set from 65% to 70%. CONCLUSION: Discrimination of proximal RCA lesion location from LCx or middle-to-distal RCA using the new LR classifier shows improvement over decision tree­type classification criteria. Automated identification of proximal RCA occlusion could speed up the risk stratification of patients with STEMI. The addition of leads V4R and V8 should further improve the automated classification of the occlusion site in RCA and LCx.

Automated discrimination of proximal right coronary artery occlusion from middle-to-distal right coronary artery occlusion and left circumflex occlusion in ST-elevation myocardial infarction.

BACKGROUND: Classifying the location of an occlusion in the culprit artery during ST-elevation myocardial infarction (STEMI) is important for risk stratification to optimize treatment. We developed a new logistic regression (LR) algorithm for 3-group classification of occlusion location as proximal right coronary artery (RCA), middle-to-distal RCA or left circumflex (LCx) coronary artery with inferior myocardial infarction. We compared the performance of the new LR algorithm with the recently introduced decision tree classifier of Fiol et al (Ann Noninvasive Electrocardiol. 2004;4:383-388) in the classification of the same 3 categories. METHODS: The new algorithm was developed on a set of electrocardiograms from an emergency department setting (n = 64) and tested on a different set from a prehospital setting (n = 68). All patients met the current STEMI criteria with angiographic confirmation of culprit artery and occlusion location. Using LR, 4 ST-segment deviation features were chosen by forward stepwise selection. Final LR coefficients were obtained by averaging more than 200 bootstrap iterations on the training set. In addition, a separate 4-feature classifier was designed adding ST features of V4R and V8, only available in the training set. RESULTS: The LR algorithm classified proximal RCA occlusion vs combined LCx occlusion and middle-to-distal RCA occlusion, with a sensitivity of 76% and specificity of 81% as compared with 71% and 62% for the Fiol classifier. The difference in specificity was statistically significant. The LR classifier trained with additional ST features of V4R and V8, but still limited to 4, improved the overall agreement in the training set from 65% to 70%. CONCLUSION: Discrimination of proximal RCA lesion location from LCx or middle-to-distal RCA using the new LR classifier shows improvement over decision tree-type classification criteria. Automated identification of proximal RCA occlusion could speed up the risk stratification of patients with STEMI. The addition of leads V4R and V8 should further improve the automated classification of the occlusion site in RCA and LCx.

Computerized classification of proximal occlusion in the left anterior descending coronary artery.

Proximal occlusion within the left anterior descending (LAD) coronary artery in patients with acute myocardial infarction leads to higher mortality than does nonproximal occlusion. We evaluated an automated program to detect proximal LAD occlusion. All patients with suspected acute coronary syndrome (n = 7,710) presenting consecutively to the emergency department of a local hospital with a coronary angiogram­confirmed flow-limiting lesion and notation of occlusion site were included in the study (n = 711). Electrocardiograms (ECGs) that met ST-segment elevation myocardial infarction (STEMI) criteria were included in the training set (n = 183). Paired angiographic location of proximal LAD and ECGs with ST elevation in the anterolateral region were used for the computer program development (n = 36). The test set was based on ECG criteria for anterolateral STEMI only without angiographic reports (n = 162). Tested against 2 expert cardiologists' agreed reading of proximal LAD occlusion, the algorithm has a sensitivity of 95% and a specificity of 82%. The algorithm is designed to have high sensitivity rather than high specificity for the purpose of not missing any proximal LAD in the STEMI population. Our preliminary evaluation suggests that the algorithm can detect proximal LAD occlusion as an additional interpretation to STEMI detection with similar accuracy as cardiologist readers.

Usefulness of p-wave duration to identify myocardial ischemia during exercise testing.

It is well recognized that ST-segment depression is due to subendocardial ischemia secondary to an increase in left ventricular end-diastolic pressure. The increase in left ventricular end-diastolic pressure is associated with increased left atrial pressure, resulting in left atrial wall distension that contributes to increasing P-wave duration (PWD). The objective of this study was to determine if PWD measured in leads II and V(5) during maximum exercise stress testing could be a reliable predictor of myocardial ischemia. Patients with suspected coronary disease underwent maximum exercise stress testing with myocardial perfusion imaging. PWD was measured using leads II and V(5) at rest and after exercise, with electrocardiographic complexes magnified 4 times (100 mm/s, 40 mm/mV). The change in PWD was calculated as Delta = PWD(recovery) - PWD(rest). DeltaPWD and ST-segment changes were related to the absence or presence of ischemia (localized reversible perfusion abnormalities) on myocardial perfusion imaging scans. DeltaPWD had sensitivity of 72%, specificity of 82%, negative predictive power (NPP) of 90%, and positive predictive power of 57%. ST-segment change had sensitivity of 34%, specificity of 87%, NPP of 80%, and positive predictive power of 47%. When DeltaPWD and ST changes were combined, sensitivity increased to 79% and NPP increased to 91%. In conclusion, DeltaPWD outperformed ST-segment changes in predicting myocardial ischemia on myocardial perfusion imaging scans. Furthermore, when DeltaPWD and ST-segment changes were combined, sensitivity and NPP were also significantly increased. In this study population, measuring DeltaPWD substantially increased the diagnostic value of maximum exercise stress testing.