Wednesday, April 24, 2024

Take the OMI Quiz and Test yourself against the Queen of Hearts

The PM Cardio Queen of Hearts AI model for ECG interpretation from Powerful Medical is still in its early days. 

Do you think you can outperform the toddler version of the AI model? 

Version 2.0 will soon be available with four times the training data.

The QoH groups ECGs into OMI and NOT OMI. Each category is subdivided into three levels of confidence. 

Thus you can get a reading of NOT OMI (low, mid or high). Or you can get sa reading of OMI (low, mid or high). 

In other words there are six outputs with NOT OMI high confidence on one end and OMI high confidence on the other end.

Take the Quiz below. It is not easy. Good luck! 

Get Queen of Hearts here for free: 

The Queen of Hearts PM Cardio App is now available in the European Union (CE approved) the App Store and on Google Play.  For Americans, you need to wait for the FDA.  But in the meantime:


If you want this bot to help you make the early diagnosis of OMI and save your patient and his/her myocardium, you can sign up to get an early beta version of the bot here.  It is not yet available, but this is your way to get on the list.


Monday, April 22, 2024

Chest pain, resolved. Does it need emergent cath lab activation (some controversy here)? And much much more.

50-something male with hypertension and 20- to 40-year smoking history presented with 1 week of stuttering chest pain that is worse with exertion, which takes many minutes to resolve after resting and never occurs at rest.  It is associated with mild dyspnea on exertion.  At times the pain does go to his left neck.  It was present on arrival at triage but then resolved before bed placement in the ED.

EKG from triage:

Queen of Hearts Interpretation:

Here is his previous ECG:

Normal ST Elevation

Queen of Hearts Interpretation:

Resident's interpretation: Reperfusion pattern/Wellens' with biphasic T waves in V2 and V3, and in comparison to an EKG in 2020 this is new.  

Course: Aspirin 325mg, chemistry, CBC, troponin panel all ordered.  Bedside ultrasound with no apparent wall motion abnormalities, no pericardial effusion, no right heart strain. Aorta briefly viewed, appears normal caliber and diameter. 

Repeat EKG:

Resident interpretation: ST elevation in V2 significantly different than his previous EKG.  Patient still not having chest pain however this is more concerning for OMI/STEMI.  He was transferred to the stabilization room for cath lab activation.  Labs ordered but not yet drawn.  Aspirin given.

Smith: I don't think that there is really any change in ST Elevation.  This change is due to a change in lead placement: in the first ECG (top), leads V1 and V2 were placed too high (P-wave inverted).  This leads to recording less ST elevation in lead V2.  In the later, 2nd, ECG, the leads are correctly placed and the ST Elevation appears to be greater.

This is a stable ECG.  Patient is pain free and clearly has Wellens' syndrome: 1) pain free episode following an episode of angina, typical Pattern A (biphasic, terminal T-wave inversion with an initial upsloping ST Segment) findings, preserved R-waves.  Pattern B is has deep symmetric T-wave inversion without the initial upwardly sloping ST Elevation.  Pattern A evolves into Pattern B.

This is a great series of ECGs demonstrating the evolution:

Classic Evolution of Wellens' T-waves over 26 hours


Does Wellens' syndrome require emergent cath lab activation??  Or is antiplatelent and antithrombotic therapy adequate to prevent re-occlusion before delayed intervention?

Wellens' syndrome is a syndrome of Transient OMI (old terminology would be transient STEMI).  

This is a demonstration of how Wellens' is transient OMI:

First ED ECG is Wellens' (pain free). What do you think the prehospital ECG showed (with pain)?

There are more examples here:

And this is literature from Wellens' himself showing how Wellens waves occur after reperfusion with thrombolytics. (first authors are Wehrens and Doevendans, respectively)

Wehrens XH, Doevendans PA, Ophuis TJ, Wellens HJ. A comparison of electrocardiographic changes during reperfusion of acute myocardial infarction by thrombolysis or percutaneous transluminal coronary angioplasty. Am Heart J. 2000;139:430–436.

Doevendans PA, Gorgels AP, van der Zee R, Partouns J, Bar FW, Wellens HJJ. Electrocardiographic diagnosis of reperfusion during thrombolytic therapy in acute myocardial infarction. Am J Cardiol. 1995;75:1206–1210.

As far as I can tell, there is only one randomized trial of immediate vs. delayed intervention for transient STEMI.  

(There is no randomized trial, or even observational trial, of immediate vs. delayed intervention for Wellens' syndrome.)

Lemkes JS, et al. Timing of revascularization in patients with transient ST-segment elevation myocardial infarction: a randomized clinical trial. Eur Heart J [Internet]. 2019;40:283–291. Available from:  

In this study, the major outcomes were the same for both groups, but of 70 patients in the delayed group, 4 required emergent intervention for sudden re-occlusion.  You can make your own conclusions.  

I think I would want to intervene before there is risk of re-occlusion.

However, one could make a reasonable argument for delaying, especially if you would need to awaken your cath team in the middle of the night.  It requires full antiplatelent and antithrombotic therapy, and, in my opinion, if you delay, you should institute continuous 12-lead ECG monitoring.  

This is why:

1. Why we need continuous 12-lead ST segment monitoring in Wellens' syndrome


Case Continued

The Cath lab was activated 70 minutes after ED arrival.

First hs troponin I returned 108 minutes after ED arrival and was normal: (12 ng/L)


No "upstream"  P2Y12 were given in the ED ("upstream" means "before the angiogram "defines" the coronary anatomy).  Upstream P2Y12 is often given, but my reading of the literature is that the benefit is limited, especially since intravenous Cangrelor can be given at the time of angiography, after the coronary anatomy is defined. This way, if there is need for CABG, surgery need not be delayed due to risk of bleeding from persistent P2Y12 inhibition.

Paper on upstream P2Y12:

Thus, in this case, no ticagrelor was given.  


--Culprit for the patient's unstable angina/Wellen syndrome is a ruptured plaque in the mid LAD.

--As suggested by the EKG, there is TIMI-3 flow on initial angiography

--LAD is a large-caliber vessel that extends to the apex

--There is an 80 to 90% stenosis in the mid LAD with TIMI-3 flow beyond on initial angiography

--This lesion has angiographic characteristics of plaque rupture and is likely the culprit for the patient's Wellens syndrome/unstable angina

Cangrelor given after coronary anatomy defined

The lesion was stented

Peak troponin was 108 ng/L.  Total coronary occlusion, if very brief, may have minimal infarction and yet be very dangerous.

Formal Echo:

Normal left ventricular cavity size, and normal LV systolic function.

Normal estimated left ventricular ejection fraction; 54%.

Regional wall motion abnormality-apical septum and apex, hypokinetic.

Increase in LV wall thickness, asymmetric (see below).

Normal right ventricular size and function.


 Thus, even with this very small infarct, there was myocardial stunning (wall motion abnormality in the absence of significant infarct.  This is common and the wall will almost always recover some time before about 6 weeks.

Here are other very interesting posts:

Wellens' syndrome: to stent or not? IVUS negative, Symptoms persist, Stress Testing, Instantaneous Wave Free Ratio, and Fractional Flow Reserve.

Wellens' has mimics and some are extremely difficult, as in this post:

Is it Wellens' Syndrome?

Here are many examples of a very common mimic, which we call "Benign T-wave Inversion" and is a variant of Early Repol:


MY Comment, by KEN GRAUER, MD (4/22/2024):

When we talk about Wellens’ Syndrome — It is all about timing. Despite description of Wellens’ Syndrome over 40 years ago — this syndrome remains misunderstood by all-to-many clinicians (See My Comment at the bottom of the page in the August 12, 2022 post in Dr. Smith’s ECG Blog).

Two ECG patterns were described by the original investigators in 1982 as being consistent with a Wellens’ Syndrome prediction of high-grade LAD stenosis.
  • Pattern B — was the more common form in the original Wellens’ report. In my experience — this ECG pattern is less specific for high-grade LAD stenosis, because other entities (including non-cardiac CNS conditions) may also be associated with symmetric T wave inversion. 
  • In contrast, Pattern A — was much less common in the original Wellens’ report. That said, when the history is “right” for Wellens’ Syndrome — the steep T wave descent from the T wave peak that Pattern A features is more specific for Wellens’ Syndrome in my experience, than the symmetric T wave inversion of Pattern B.

  • In the initial ECG from today’s case — the ST-T wave appearance within the light BLUE rectangles in Figure-1 is consistent with the more specific ECG Pattern A. This is especially true for the ST-T wave in lead V3 — in which descent of the T wave from the peak of the coved and slightly elevated ST segment, is extremely steep ( = another example of an ECG condition in which a picture is “worth 1,000 words” — because the shape of this steep T wave descent is unique).

Misunderstood Features of Wellens:
  • For Wellens’ Syndrome to be present — the Hx must be of prior CP (Chest Pain) that has now resolved at the time the ECG is recorded. If CP is still present — then this is not Wellens’ Syndrome. Instead, persistent CP may indicate ongoing infarction and/or reperfusion T waves that develop after a completed event.
  • It is the pathophysiology of a true "Wellens' Syndrome" that is commonly misunderstood — in that there has been transient coronary occlusion — that then spontaneously resolves, with resultant resolution of CP (and no more than minimal myocardial damage).
  • The above pathophysiology is the reason Dr. Smith's recognition that the repeat ECG in today's case was not evolving is important. This is because by the above rationale — recognition of Wellens' Syndrome implies a temporarily "stable" situation after there has been spontaneous reperfusion of the "culprit" LAD lesion.
  • As per Dr. Smith’s above review of current literature on this issue — it would seem that prompt cath with PCI of a patient with Wellens' Syndrome is preferable to close monitoring while you wait and hope that you don't have to do emergency cath IF the "culprit vessel" reoccludes. And, given that such reocclusion of the "culprit" vessel might occur at almost any time after spontaneous reperfusion (ie, hours or even days later) — What IF the spontaneous reocclusion only occurs after the patient has been sent home because "nothing happened" during the 1-2 days that the patient was monitored in the hospital?

Figure-1: I've labeled the initial ECG in today's case.

Additional Take-Home Points:
I thought the KEY clinical point underscored by Dr. Smith's insightful presentation of today's case is how quickly the diagnosis of Wellens' Syndrome can be made.
  • Clearly — comparison of ECG #1 with the previous ECG on today's patient confirmed that the steep T wave downsloping, in association with anterior T wave inversion was a new finding. BUT — prior tracings are not always available — and even IF this prior ECG would not have been found — the diagnosis of Wellens' Syndrome was already confirmed by brief history that the patient's CP had resolved at the time the uniquely steep anterior T wave downsloping was seen on the initial ECG.
  • Sobering Note: The need for prompt cath in today's case was only appreciated after the 2nd ECG was done — in which a change in lead placement led to the team mistakenly concluding that there was increased ST elevation. (See My Comment in the April 17, 2022 post of Dr. Smith's ECG Blog for review on how to recognize too-high placement of the V1,V2 electrodes).

Final Thoughts:
  • The diagnosis of Wellens' Syndrome in today's case was made more difficult because this 50-something man with hypertension has LVH on ECG (by Peguero Criteria — as sum of the 25 mm S in V3 + the 13 mm S in V4 >28 mm — as per LVH Criteria that I detail in my Figure-3 of the April 17, 2022 post).
  • That said —  LV "strain" should never look like the steep T wave descent that we see in lead V3 of Figure-1, in which there is no indication of LV "strain" in the lateral chest leads.

  • BOTTOM Line: The history in today's case of new CP that had resolved by the time the initial ECG showing the unique steep T wave downsloping with anterior T wave inversion was done — should allow immediate recognition of Wellens' Syndrome. From what we know about the pathophysiology of Wellens' Syndrome — it would seem that the sooner prompt cath with PCI can be arranged, the better the chance for optimal outcome.

Friday, April 19, 2024

Is OMI an ECG Diagnosis?

Written by Jesse McLaren


A 70 year old with prior MIs and stents to LAD and RCA presented to the emergency department with 2 weeks of increasing exertional chest pain radiating to the left arm, associated with nausea. The pain recurred at rest 90 minutes prior to presentation, felt like the patient’s prior MIs, and was not relieved by 6 sprays of nitro. Paramedics provided another 3 sprays of nitro, and 6mg of morphine, which reduced but did not resolve the pain.  What do you think of the ECG, and does it matter?

There’s normal sinus rhythm, LAFB, old anterior Q waves, and no diagnostic sign of OMI. I sent this to the Queen of Hearts

So the ECG is both STEMI negative and has no subtle diagnostic signs of occlusion. But does this matter?


The ECG is just a test: a Bayesian approach to acute coronary occlusion


If a patient with a recent femur fracture has sudden onset of pleuritic chest pain, shortness of breath, and hemoptysis, the D-dimer doesn’t matter: the patient’s pre-test likelihood for PE is so high that they need a CT.  Similarly, if a patient with known CAD presents with refractory ischemic chest pain, the ECG barely matters: the pre-test likelihood of acute coronary occlusion is so high that they need an emergent angiogram.  Non-STEMI guidelines call for “urgent/immediate invasive strategy is indicated in patients with NSTE-ACS who have refractory angina or hemodynamic or electrical instability,” regardless of ECG findings.[1]  European guidelines add "regardless of biomarkers".


But only 6.4% of such ‘high risk Non-STEMI’ patients get angiography within 2 hours.[2] This is because, contrary to Bayesian reasoning, the STEMI paradigm is named after and defined by one part of one test: ST elevation on ECG. Emergent cath lab activation is also named after this test (code STEMI), so patients whose ECGs don’t meet STEMI criteria don’t get emergent angiograms, despite guidelines. 


The sensitivity of STEMI criteria for acute coronary occlusion is much worse than the sensitivity of D-dimer for PE: a recent meta-analysis of the only three studies that have assessed STEMI criteria found sensitivity of only 43.6%”[3] Expert ECG interpretation for subtle ECG signs of Occlusion MI, and AI trained to identify these signs, have twice the sensitivity of STEMI criteria with preserved specificity.[4] This is a major improvement, but it means that the ECG at its best is still only 80% sensitive for OMI. This is why the OMI paradigm is not named after the ECG or any other test - all of which have their limitations, including angiography - but rather the pathology in the patient. The OMI paradigm shift both maximizes the test characteristics of the ECG, while putting them in context of the patient – using Bayesian reasoning.[5]


Back to the case


The patient had serial ECGs over the next hour with no significant change:

The first troponin came back at 1,400 ng/L (normal <26 in males and <16 in females), confirming MI – and the patient’s refractory ischemia indicated this was an Occlusion MI. But no ECG met STEMI criteria so the patient was referred to cardiology as Non-STEMI.


Cardiology started a nitropatch (ACC/AHA guidelines specifically state that they are ineffective and should not be used), with a plan for nitro infusion (rather than emergent cath) if the pain worsened. But when the repeat troponin two hours later rose to 9,000 ng/L,  the patient was transferred for urgent angiogram. Door-to-cath time was 7 hours, and found a complex 99% ostial LAD lesion and 80% OM lesion. Echo showed new anterior regional wall motion abnormality and decrease EF from 60% to 45%.


The patient was transferred to CCU to consider surgical options. Refractory ischemic chest pain continued and trop increased to 160,000ng/L, with subtle convex anterior ST elevation:

The patient was brought back to cath lab for stenting of LAD and balloon angioplasty to OM. Peak troponin was 225,000 ng/L and discharge ECG showed anterior reperfusion T wave inversion

This was a massive infarct from an acutely occluded coronary artery, yet no ECG met STEMI (or OMI) criteria. And because there was no Code STEMI, the discharge diagnosis was “non-STEMI”, so this case will not be flagged as an opportunity for improvement. If instead this was considered an OMI with delayed reperfusion, then there could be steps towards improvement. For example, the patient could have been identified at triage (or even earlier, by EMS) as having a high likelihood of OMI, so despite a non-diagnostic ECG there could have been a stat cardiology consult with bedside echo, and cath lab activation before waiting for serial troponins to rise. But the only way to actually meet 'high risk NSTEMI' guidelines is to shift the paradigm to Occlusion MI.



Take away

1.     STEMI criteria miss the majority of OMI, and ‘high risk NSTEMI’ guidelines are not followed because both the disease (STEMI) and the treatment protocol (code STEMI) are named after the poor surrogate marker of STE

2.     OMI is a clinical diagnosis that incorporates advanced ECG interpretation, complementary echo, and treatment for refractory ischemia regardless of the ECG

3.     Queen of Hearts can double the sensitivity of STEMI criteria with preserved specificity, but needs to be applied in clinical context and doesn’t rule out OMI if there is high pre-test likelihood.

You can get the app here:

The Queen of Hearts PM Cardio App is now available in the European Union (CE approved) the App Store and on Google Play.  For Americans, you need to wait for the FDA.  But in the meantime:


If you want this bot to help you make the early diagnosis of OMI and save your patient and his/her myocardium, you can sign up to get an early beta version of the bot here.  It is not yet available, but this is your way to get on the list.



1.     Amsterdam et al. 2014 AHA/ACC guideline for the management of patients with non-ST elevation acute coronary syndromes. Circulation 2014

2.     Alencar et al. Systematic review and meta-analysis of diagnostic test accuracy of ST-segment elevation for acute coronary occlusion. Int J Cardiol 2024

3.     Lupu et al. Immediate and early percutaneous coronary intervention in very high-risk and high-risk non-ST segment elevation myocardial infarction patients. Clin Cardiol 2022

4.     Herman, Meyers, Smith et al. International evaluation of an artificial-intelligence-powered electrocardiogram model detecting acute coronary occlusion myocardial infarction. Eur Heart J Digital Health 2024 

5.     McLaren and Smith. A Bayesian approach to acute coronary occlusion. J Electrocardiol 2023


MY Comment, by KEN GRAUER, MD (4/19/2024):

Insightful case by Dr. McLaren — that emphasizes the importance of the application of Bayes' Theorem to the clinical history. In other words — "When you hear hoofbeats — Think of horses (not zebras)."
  • As per Dr. McLaren — regardless of what the initial ECG in today's case shows, the onus is on us to rule out acute OMI, rather than the other way around. That's because "Time is muscle" — and regardless what the initial troponin and initial ECG show — IF the "pre-ECG likelihood" for acute OMI is high in a patient with new CP (Chest Pain) — then the threshold for performing prompt cath needs to be lowered (if we are to have any hope of lowering door-to-cath time from the unjustifiable 7 hours that it was in today's case).

  • As Drs. Smith and Meyers have so often emphasized — many (perhaps most"NSTEMIs" end up being OMIs (despite the fact that most of the time the discharge diagnosis remains NSTEMI) — so much so, that the term, "NSTEMI" has in essence become a useless term (See the October 11, 2020 and the September 10, 2023 posts in Dr. Smith's ECG Blog).

I would add the following PEARL to Dr. McLarens excellent presentation:
  • Given the very high pre-ECG likelihood of acute OMI in today's case — I immediately lowered my "threshold" in the assessment of the initial ECG for "suspicious" ECG findings.
  • While there is no indication of acute OMI from sinus-conducted beats in today's initial ECG — Even before I looked at the answer, I strongly suspected an acute MI from the appearance of the PVC in lead I of Figure-1 (1st beat in lead I within the BLUE rectangle).
  • We have previously shown examples in which acute OMI is only recognized by the appearance of ST-T waves in a PVC, despite no indication of OMI from sinus-conducted beats (See My Comment at the bottom of the page in the October 8, 2018 and September 13, 2022 posts in Dr. Smith's ECG Blog).

  • While specific criteria for when ST-T wave appearance in a PVC is indicative of acute OMI do not exist (probably being impossible to objectively study this phenomenon in random, prospective, controlled trials) — awareness of ST-T wave changes that just-shouldn't-be-there is the subjective criterion that I have found most helpful to me.
  • Whereas I would not call the ST-T wave depression of the PVC in leads II and III of Figure-1 diagnostic (even though the relative amount of ST depression in these leads is marked) — the overly tall, "fatter"-at-its-peak and wider-at-its-base T wave in lead I (compared to the tiny QRS of the PVC in this lead) clearly should not be there (BLUE arrow in lead I that suggests a hyperacute T wave for this PVC)Given the history in today's case — even by itself, this PVC appearance in this initial ECG seems strongly suggestive of acute OMI.

Figure-1: I’ve labeled the initial ECG in today's case.

Additional points (Beyond-the-Core) regarding interpretation of today's initial ECG:
  • Optimal description of the 12-lead in Figure-1 is challenging. There is sinus rhythm with 1 PVC — slight QRS widening (to 0.11-to-0.12 second) — with marked LAD (Left Axis Deviation).
  • The QRS is not wide enough — and QRS morphology lacks the all-upright monophasic R wave in lead V6 — therefore not qualifying as complete LBBB. Instead (as noted by Dr. McLaren) — LAHB (Left Anterior HemiBlockcriteria are satisfied.
  • The overly tall R wave (far exceeding 12 mm) qualifies as LVH (supported by an ST-T wave appearance in lead aVL completely typical for LV "strain").
  • To NOTE: LVH and LAHB are each estimated to potentially increase QRS duration by 0.01-to-0.02 second (due to slight delay in depolarization of a thicker LV and/or an LV in which the left anterior hemidivision of the conduction system is not functioning). This accounts in Figure-1 for more QRS widening that we might expect with either LVH or LAHB alone.
  • While the QS waves in leads V1-thru-V3 in today's initial ECG do represent prior anterior infarction — it is good to appreciate that both LVH and LAHB may result in delayed R wave progression, including the presence of anterior QS waves (LVH by predominance of posterior forces from the large LV that attenuates anterior r wave forces — and LAHB by posterior depolarization via the left posterior hemifascicle that is no longer "opposed" by the blocked left anterior hemifascicle)
  • Despite joint LVH and LAHB — the fragmentation that is clearly seen on the upslope of the S waves in leads V1 and V2 (RED arrows) supports the premise that the anterior QS waves indicate prior anterior infarction. 

Recommended Resources