About OMI:
Occlusion Myocardial Infarction

The current standard of care misidentifies 50% of patients with OMI causing 2x more short and long-term deaths.

What is OMI?


Occlusion Myocardial Infarction (OMI) refers to an acute coronary occlusion or near occlusion of a culprit artery with insufficient collateral circulation, resulting in transmural myocardial infarction and cardiac tissue death.


OMI is not an exclusively electrocardiographic (ECG) pattern. Thus it is important to determine OMI on the basis of more clinical investigations, including clinical presentation, cardiac troponin and coronary angiography vessel stenosis and perfusion.


The primary OMI endpoint is defined as an acute culprit vessel identified angiographically with:

  1. Thrombolysis In Myocardial Infarction (TIMI) 0-2 Flow

  2. TIMI 3 Flow and very high Biomarker elevation (indicating that the vessel was occluded causing a large infarction)

OMI-FAQ

OMI ECG criteria

 

Notably, a patient’s lack of ST-elevation on the initial ECG does not necessarily preclude the presence of OMI. Approximately half of OMI patients may never exhibit ST elevation. Under the OMI paradigm, there are no ECG millimeter criteria. Instead, it requires a thorough examination of the ECG, a comprehensive evaluation of the clinical context, and diligent scrutiny of laboratory findings.

 

Emre Aslanger has created an excellent flowchart that effectively illustrates the process of identifying OMI ECG criteria. It includes:

  • Recognition of patterns with ST elevation only in one lead
  • Subtle ST elevation with minimal reciprocal changes
  • Exclusively identifying ST depression
  • Bulky (hyper-acute) T-waves

OMI ECG criteria

 

Notably, a patient’s lack of ST-elevation on the initial ECG does not necessarily preclude the presence of OMI. Approximately half of OMI patients may never exhibit ST elevation. Under the OMI paradigm, there are no ECG millimeter criteria. Instead, it requires a thorough examination of the ECG, a comprehensive evaluation of the clinical context, and diligent scrutiny of laboratory findings.

 

Emre Aslanger has created an excellent flowchart that effectively illustrates the process of identifying OMI ECG criteria. It includes:

  • Recognition of patterns with ST elevation only in one lead
  • Subtle ST elevation with minimal reciprocal changes
  • Exclusively identifying ST depression
  • Bulky (hyper-acute) T-waves
OMI About

Accuracy of OMI ECG criteria

 

In a substantial cohort of 808 patients suspected of acute coronary syndrome (ACS) and confirmed angiographic outcomes, trained ECG experts have rigorously tested this novel ECG interpretation approach. The results of their study revealed a remarkable two-fold increase in sensitivity compared to STEMI criteria (86% vs. 41%) while maintaining statistically equivalent specificity (91% vs. 94%)⁸.

 

These findings strongly indicate that these refined criteria are remarkably more effective in precisely identifying patients with OMI (occlusion myocardial infarction). The study’s results emphasize the importance of adopting these more sophisticated ECG criteria to enhance the diagnostic accuracy of OMI cases and improve patient management and outcomes.

The Role of AI in OMI Detection

 

Recognizing OMI on a 12-lead ECG largely relies on pattern recognition, a process that experts have honed through examining thousands of past cases. The ever-changing attributes of acute coronary syndromes necessitate many years of training to understand intricate ECG patterns.

 

Effectively training first responders to recognize these complex patterns remains challenging. Therefore, overcoming these constraints to widespread adoption of this paradigm could be aided by artificial intelligence (AI) models deployed to augment current clinical workflows.

 

The potential of AI in analyzing ECG waveforms has been demonstrated in various patient groups, showcasing significant progress in diagnosing conditions such as heart failure, hypertrophic cardiomyopathy, and sudden cardiac arrest.

 

QOH (Queen Of Hearts) algorithm, used for OMI diagnostic in PMcardio application, represents this effort with already admirable results. Overcoming most medical practitioners, even cardiologists. The table below summarizes the AI model performance on the overall Validation set. In addition to industry standard metrics, summative metrics, such as the area under the curve (AUC) and Matthew’s correlation coefficient (MCC) are reported, when applicable.

Comparison of the AI model performance to STEMI criteria and Experts for the whole Validation dataset (aggregated to patient level)

Ref - Ref + Accuracy Sensitivity Specificity NPV PPV MCC
OMI AI Model - 1655 99 0.905 (0.892-0.917) 0.798 (0.762-0.833) 0.935 (0.923-0.946) 0.944 (0.933-0.954) 0.771 (0.733-0.807) 0.723 (0.688-0.758)
+ 116 390
STEMI Criteria - 1730 330 0.836 (0.82-0.851) 0.325 (0.284-0.368) 0.977 (0.969-0.984) 0.84 (0.824-0.856) 0.795 (0.735-0.85) 0.438 (0.391-0.483)
+ 41 159
ECG Experts - 1694 132 0.908 (0.896-0.919) 0.73 (0.69-0.769) 0.957 (0.947-0.966) 0.928 (0.916-0.94) 0.823 (0.786-0.857) 0.718 (0.681-0.753)
+ 77 357