Thursday, June 13, 2024

The Expert Witness re-visits a chest pain Malpractice case using the Queen of Hearts

This case was posted by Eric Funk (, editor of the Expert Witness Newsletter, on his site.

Here is the case on his site.   He has graciously allowed me to repost it here. It is a case he presented several years ago, but then want to re-visit in light of the Queen of Hearts AI system.

Don't miss his analysis and assessment of the Queen of Hearts AI OMI ECG bot 

       -- that assessment is at the very bottom of the post.

Click here to sign up for Queen of Hearts Access


A 58-year-old woman presented to the ED with burning chest pain that started 2-3 hours earlier while sitting on a porch swing.

Her first set of vitals were documented:

  • BP 116/57

  • Pulse 94bpm

  • Respiratory rate 24/min

  • O2 sat 90% on room air

  • Temp 97F

She had been cleaning a Jeep in the sun, and was sunburned.

Past medical history included Crohn’s disease, hyperlipidemia, hypothyroidism, and she smoked 1ppd.

Here is her ED EKG:

What do you think?
Since it is in the Expert Witness Newsletter, there must be something that was missed, right?

I would have very high suspicion for OMI on this ECG.  But which artery?  There is STE and a large T-wave in V1, with reciprocal ST depression in V5 and V6.  This is typical of LAD occlusion proximal to the first septal perforator, an ECG for which we have coined the term "Swirl".  

However, the pattern is also seen in inferior OMI with right ventricular OMI. (V1 sits over both the RV and the septum, so transmural ischemia of either one with give OMI pattern in V1 and reciprocal STD in V5 and V6.  

See this post: 

Septal STEMI with ST elevation in V1 and V4R, and reciprocal ST depression in V5, V6.  (Also seen in inferior + RV OMI.)

There appear to be hyperacute T-waves in inferior leads.  There is also STD in V2.  This suggests inferior-posterior OMI and makes me lean to a proximal RCA occlusion.  But there is also perhaps some STD in inferior leads -- this would support LAD.

In any case, it is diagnostic of OMI in a chest pain patient.

The Queen of Hearts agrees:

It is interesting to see what the Queen of Hearts highlights on this explainability map: ST depression in V5 is the most important finding for her.

Case progression:

The automated EKG interpretation was “sinus rhythm with sinus arrhythmia, right atrial enlargement, rightward axis, possible anterior infarct, age undetermined, abnormal ECG”.

The physician documented “normal sinus rhythm”.

The patient was given 0.4mg nitroglycerin SL, which did not change her pain.

She was given 500mL of normal saline and IV Toradol 30mg.

CBC and CMP were unremarkable.

A troponin was normal at 0.06 (normal range 0.00 - 0.10).

CK MB was 1.9 (normal range 0.0 - 3.6).

A chest x-ray was unremarkable.

The physician documented that she was “improved” and the patient was discharged.

The following morning, the patient collapsed in the kitchen in front of her family.

EMS arrived to a pulseless patient in V fib.

She was successfully defibrillated and taken back to the ED.

After being transferred to an academic center, she was taken to the cath lab:

Proximal RCA occlusion (causing inferior and RV OMI)

Unfortunately, she continued to decline despite aggressive measures.

The family elected to switch to comfort care and she died.

The family filed a lawsuit against the physician and the hospital.

They sued the the county-operated EMS service for allegedly not dispatching the call fast enough.

They also sued both EMTs for allegedly not starting chest compressions and defibrillating quickly enough.

The family sued.  

You can read the defense EM expert report on Dr. Funk's site:

Case Outcome

The EMTs were found to have personal immunity as part of their employment with the local government. They were dismissed.

The 911 call center and EMS service were also dismissed.

The hospital and doctor reached confidential settlements.

MedMalReviewer Analysis:

After originally working on this case several years ago, I became interested in re-analyzing it in light of The OMI Manifesto. In short, the OMI paradigm offers an improved method of EKG interpretation that seeks to replace the STEMI/NSTEMI dichotomy that currently dictates emergency cath lab activation (but is rife with false positives and false negatives). While initially promising, I grew frustrated with the OMI model due to the extreme nuance it mandates in EKG interpretation, and my own inability to identify findings that the OMI experts claimed were obvious. I remain skeptical that the average physician should be expected to unilaterally interpret EKGs to the level of the OMI experts.

However, my frustration quickly reversed with the release of the Queen of Hearts algorithm, which was trained by Dr. Stephen Smith using AI to detect occlusion myocardial infarction that is often missed using standard STEMI criteria. It allows users to take a picture of an EKG, which is digitized and then analyzed for OMI.

Naturally, I ran the EKG from this case through the Queen of Hearts algorithm. Sure enough, it detected OMI with high probability:

I was initially skeptical that the algorithm was just flagging everything as OMI. I repeatedly tested the algorithm with multiple other EKGs, and my anecdotal report is that I found it to be highly reliable.


MY Comment, by KEN GRAUER, MD (6/13/2024):
Interpretation of the subtle ECG changes seen in a significant percentage of acute OMIs can be difficult. This task is made even more challenging when technical issues such as erroneous lead placement or excessive artifact — call into question our ability to accurately interpret a number of leads on the tracing in front of us. 
  • To illustrate this point — I've labeled the initial ECG in today's case.
Before looking at my labeled Figure-1 — Take Another LOOK at the unlabeled initial tracing shown above at the beginning of today’s case.
  • Is there ST depression in the inferior leads in Figure-1?
  • Is there ST elevation in lead aVL?
  • Are there any leads with definite abnormal findings?

MY Thoughts on Today’s Initial ECG:
The history in today’s case is concerning (ie, = a 58-year old woman with longterm tobacco use, who presented with new-onset CP [Chest Pain]).
  • As we frequently emphasize — this type of history automatically places today’s patient in a higher-prevalence group at risk of an ongoing cardiac event. As a result — our threshold for interpreting ECG changes as “abnormal” should be lowered. Given this history — the onus falls on us is to rule out ACS (Acute Coronary Syndrome) — rather than ruling it in.

  • That said — technical issues with today’s initial ECG (in the form of an extremely wavy baseline in 4 of the limb leads) — make it difficult to know IF we truly are seeing ST depression in the inferior leads? — and, if there is truly ST elevation in lead aVL?
  • KEY Point: I think it important to appreciate how we are handicapped in our interpretation of today’s initial ECG by the technical issue of a wavy baseline that calls into question the ECG information conveyed from 4 of the 12 leads. While I thought it likely that there is “real” inferior lead ST depression — and Q waves with subtle-but-real ST elevation in lead aVL — I was less than certain about this given the excessively wavy baseline in these limb leads.

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

Continuing Assessment of ECG #1:
Despite uncertainty about the ECG information provided by leads II,III,avF and aVL — there are other leads that definitely are abnormal.
  • In a patient like the woman in today’s case, who presents with new CP — there is no way the ST-T wave appearance of the 2 leads within the RED rectangle can be normal.
  • In lead V1 — The ST segment straightening, with slight-but-real J-point ST elevation and hyperacute T wave appearance are never “normal” findings in this lead.
  • Lead V1’s appearance is in sharp contrast to what we see in lead V2 — in which the usual slight, gently upsloping ST elevation that is normally seen in this lead — has been replaced by an isoelectric, straightened ST segment (RED arrow). In a patient with new CP — this suggests posterior OMI until proven otherwise.
  • Additional definitely abnormal ST-T waves are seen in the 2 leads within the BLUE rectangle. In leads V5 and V6 — BLUE arrows highlight “ledge-like” and significant ST depression.
  • BLUE arrows in lead I add support to the validity of the finding that there clearly is abnormal, straightened ST depression in the lateral leads.

In a patient with new CP — today’s initial ECG has to be interpreted as clearly abnormal until proven otherwise.
  • As enthusiastic as I am about the amazing (and ever increasing) accuracy of the QOH AI application — I believe that optimal clinical management  should also depend on capable provider assessment of emergency ECGs. Optimal use of QOH can best be achieved by joint effort (ie, Pre-hospital and hospital providers seeing the ECG first — and then supplementing their interpretation by what they learn from QOH input).
  • Therefore — an important benefit of QOH — is that providers can (and should) learn from her — so that with time, the interpretations of all providers (pre-hospital, emergency clinicians and cardiologists) will improve regarding more timely recognition of acute OMI.

  • Regarding specifics of today's case — I fully acknowledge that I found myself unable to predict the “culprit” artery based solely on the initial ECG. That said — I nevertheless was able to strongly suspect acute OMI based on the history of new CP and the initial ECG. I believe that recognition of the ST-T wave abnormalities I highlight in Figure-1 should be within the skill set of experienced prehospital providers, emergency physicians, and the cardiologists called on for acute cath lab activation.
  • At the least — if technical issues (such as the overly wavy baseline in the limb leads from today’s case) prevent valid interpretation — then the ECG should be promptly repeated. The decision of whether or not to activate the cath lab hangs in the balance.
  • And, if after assessment of a technically adequate tracing there is still uncertainty about whether this higher risk patient with new CP is having an acute OMI — then serial ECGs should be repeated frequently (within every 10-30 minutes, or sountil such time that joint decision-making between clinician with QOH input can comfortably rule in or rule out acute OMI.


Tuesday, June 11, 2024

What if your system adopted the recommendation that a computer "normal" ECG need not be shown to the doctor?

Written by Pendell Meyers with edits by Smith.  Sent by anonymous

A man in his 40s with no previous heart disease presented within 30 minutes of onset of acute chest pain that started while exercising. There was associated shortness of breath and left arm radiation. 

This Triage ECG (ECG #1) was recorded on a chest pain patient at triage at 1906 (top highest quality image, bottom photo including computer algorithm interpretation):

"Sinus rhythm, normal ECG" (This was performed on a Mortara machine, most likely using Veritas algorithm, but I do not have perfect confirmation of that)

This ECG shows an obvious inferior acute coronary occlusion (OMI).  It is clearly missed by the conventional algorithm.  

Many systems now refrain from showing computer "normal" ECGs to the busy emergency physicians at triage because of very poorly conceived articles that say that if the computer algorithm says "normal," the emergency physician should not be bothered.  

Luckily, this institution does show all triage ECGs to the physician, who in this case immediately recognized OMI and activated emergent transfer to the local PCI center.

Here is the ECG interpreted by the Queen of Hearts:

Click here to sign up for Queen of Hearts Access

Here is ECG #2 at 1959 (I believe this is the time of arrival to the PCI center): 

Again, outrageous conventional algorithm interpretation!
Now it is a full blown STEMI of 3 myocardial territories: inferior, posterior, and lateral
But at least it does not call it "Normal."

Queen of Hearts:

The initial troponin (high sensitivity troponin I) returned less than 6 ng/L.  Below the limit of detection.

Angiogram findings included:

95% mid RCA stenosis with occluded distal right PDA secondary to thrombus (peristent OMI). Successful drug-eluting stent placement opening up 95% mid RCA stenosis to 0% residual

Nonobstructive left system disease. 

Left-ventriculogram showed severe infero-apical hypokinesis with LV ejection fraction 50 to 55%. LVEDP 25.   This is a significant loss of myocardium and ejection fraction.  Some function might possibly recover over weeks.

Medical therapy for thrombotic occlusion of distal right PDA.

Formal echocardiogram:

Systolic function is at the lower limits of normal. The ejection fraction is 50% +/- 5% , calculated using biplane MOD. Severe hypokinesis of the mid-apical inferior and inferoseptal myocardium.

Troponin trend:
less than 6 ng/L
933 ng/L
13,386 ng/L, typical of STEMI
(none further measured -- it might have peaked at a much higher level if it had been measured to peak)

Repeat ECGs after PCI:

These are diagnostic of reperfusion.

The patient was discharged home the next day. No further follow up is available.

Learning Points:

You cannot trust conventional algorithms even to find STEMI(+) OMI, even when they say "normal ECG." We have shown many examples of this on this blog.

Queen of Hearts is available and performs well.

Click here to sign up for Queen of Hearts Access

You should not wait for the troponin when the history and ECG is diagnostic. Even in obvious STEMI(+) OMI, the initial troponin can easily be negative in the initial short time from onset of OMI.

The Queen of Hearts diagnoses almost all of these so called "Normal" ECGs with are in fact OMI and she does so with High Confidence:  

See this post of 10 cases:

When the conventional algorithm diagnoses the ECG as COMPLETELY NORMAL, but there is in fact OMI, what does the Queen of Hearts PM Cardio AI app say? (with 10 case examples)

See other relevant posts:

An undergraduate who is an EKG tech sees something. The computer calls it completely normal. How about the physicians?

Three patients with chest pain and “normal” ECGs: which had OMI? Which were normal? And how did the Queen of Hearts perform?

Four patients with chest pain and ‘normal’ ECG: can you trust the computer interpretation?

And literature:

McLaren, Meyers, Smith and Chartier. Emergency department Code STEMI patients with initial electrocardiogram labeled ‘normal’ by computer interpretation: a 7-year retrospective review. Acad Emerg Med 2024;31:296-300


MY Comment, by KEN GRAUER, MD (6/11/2024):
Recognition of repolarization variants can be challenging. At times, the distinction between a repolarization variant vs the early stage of acute OMI may not be possible solely on the basis of a single ECG.
  • To add to this complexity (as per My Comment in the August 22, 2020 post in Dr. Smith's ECG Blog) — the ST-T wave appearance in repolarization variants may be dynamic! On occasion — ST-T wave appearance with repolarization variants may change from one-hour-to-the-next — or, ST-T wave appearance may change due to a difference in heart rate, performance of exercise, or variation in vagal tone — and, sometimes even without any obvious explanation.
  • Finally — there is the clinical reality that a patient who has a "baseline" ECG that manifests a repolarization variant — may at some point develop acute coronary occlusion that in part is masked by benign-appearing ECG characteristics of the underlying repolarization variant.
It is for the above reasons that I was at first uncertain about the ST-T wave appearance in the inferior leads of the initial ECG in today's case (within the light BLUE rectangles in leads II,III,aVF in Figure-1).
  • Although the amount of J-point ST elevation in leads II,III,aVF in Figure-1 is more than is usually seen with repolarization variants, and the peak of the T wave in these leads seemed "bulkier" than usual — the upward-concavity shape of the ST segment (ie, "smiley"-configuration) was not unlike that seen in many repolarization variants.
  • Small and narrow q waves are seen in each of the inferior leads of ECG #1 — but in this patient with a relatively vertical frontal plane axis, narrow inferior lead q waves are a common normal manifestation of septal depolarization.
  • BOTTOM Line: I would not have diagnosed an acute OMI on the sole basis of inferior lead appearance in today's initial tracing.

KEY Point: Despite what I felt was the nondiagnostic picture presented by the ST-T wave appearance in leads II,III,aVF — Definitive ECG diagnosis of an acute OMI is present in ECG #1 for the following reasons:
  • The clinical history immediately places today's patient in a higher-prevalence group of patients likely to be evolving an acute OMI (ie, a middle-aged man who presents to the ED for new-onset CP [Chest Pain] that occurs during exercise, and lasts for at least 30 minutes!).
  • There is no way the ST-T wave appearance in lead aVL can be normal (within the RED rectangle in this lead). True reciprocal ST-T wave depression does not develop with repolarization variants. While some T wave inversion may normally be seen in lead aVL when the QRS is predominantly negative — there should not be J-point depression in such cases (the RED arrow in lead aVL) — and the inverted T wave should not be as "bulky" as it appears to be in lead aVL of ECG #1.
  • Once we know in this patient with new CP that the ST-T wave appearance in lead aVL is definitely abnormal (and consistent with reciprocal ST depression) — we then have to presume that the upward-concavity ST elevation in each of the inferior leads is not simply due to a repolarization variant — but instead, must be interpreted as an acute inferior OMI until proven otherwise.
  • PEARL: Acute posterior involvement is a common accompaniment of acute inferior OMI. Therefore, the fact that the limb leads show an acute inferior OMI — should prompt us to carefully scrutinize anterior leads for any suggestion of posterior involvement.
  • Having said this — there is no way the ST-T wave appearance in lead V3 of ECG #1 can be normal (within the RED rectangle in this lead). As we often emphasize on Dr. Smith's ECG Blog — there is normally slight, upward sloping ST elevation in leads V2 and V3. The RED arrow in lead V3 highlights the isoelectric (ie, non-elevated) baseline of the J-point in this lead — which in the context of the above ECG findings, strongly suggests acute posterior OMI until proven otherwise.
  • By the concept of neighboring leads — I strongly suspected that the ST-T wave appearance in lead V2 was also abnormal because: i) There is no more than the most minimal J-point elevation in this lead; and, ii) The T wave appears more pointed than expected (potentially suggesting some posterior reperfusion).
  • To EMPHASIZE: I would not perceive lead V2 by itself to be abnormal (especially given how deep the S wave is in this lead) — but because the history and limb lead appearance are diagnostic of acute inferior OMI — and — because lead V3 truly suggests associated acute posterior involvement — I suspected that the subtle changes in lead V2 were probably also abnormal.

Figure-1: I've labeled the initial ECG in today's case — and compare it with the repeat ECG done 53 minutes later.

What Do We Learn from ECG #2?
The diagnosis of acute infero-postero-lateral OMI becomes obvious with the evolution seen in ECG #2: 
  • The ST elevation in the inferior leads of ECG #2 has increased — and the shape of this ST elevation has clearly taken on a more acute appearance.
  • There is more ST depression in lead aVL — and the shape of this reciprocal ST depression has taken on a "shelf-like" appearance.
  • The peaked T wave previously seen in lead V2 has been replaced by ST depression. The upright T wave seen in lead V1 of ECG #1 is also gone.
  • There is now frank ST segment straightening, with some definite ST elevation in leads V3-thru-V6.
At this point — a STEMI was diagnosed, and cardiac cath with PCI was performed.
  • Given the history of new-onset worrisome CP — the initial ECG in today's case was diagnostic of acute infero-postero OMI for the reasons detailed above.
  • The fact that the initial troponin was normal does not in any way rule out acute OMI — as the initial troponin is not always elevated.
  • KEY Point: The comparison between ECG #1 and ECG #2 done just 53 minutes later — shows a dramatic change! Significant ST-T wave abnormalities may evolve over a period of minutes in an actively ongoing OMI. If either the initial ECG in today's case was not recognized as diagnostic of acute OMI — or — the interventionist cardiologist On-Call could not yet be convinced of the need for prompt cath from the initial ECG — it is likely that a repeat ECG done much sooner than 53 minutes later would have satisfied their criteria for cath.
  • In cases like this — repeating the ECG every ~10-20 minutes until there are ECG changes sufficient for the patient to be accepted for cath is advised.

Recommended Resources