Friday, March 31, 2023

A 50-something with chest pain. what to do? And get the PM Cardio app for your own use here!

This was sent to me by a friend.

It is from a 50-something with chest pain:

What do you think?

This was marked as "Not a STEMI" by the physicians.  

It is not a STEMI, but it is diagnostic of an LAD OMI (Occlusion MI).  There is subtle ST Elevation in V1-V4 and hyperacute T-waves in V2-V6.

There is also subtle but diagnostic INFERIOR ST Elevation, with reciprocal ST Depression in aVL.

I uploaded this to our new PM Cardio AI Bot app.  

We call the app the "Queen of Hearts"

First, it transforms it to a digital file and standardizes the image.  Here is the result:

Next, it interprets the digital data:

A perfect interpretation

Unfortunately, this ECG was not interpreted as OMI and the patient did not get another ECG recorded until 2 hours later.  

Here it is:
No one could miss this.

So there was an extra 2 hour delay to intervention for this patient, and a lot of lost myocardium.


If you want this bot to help you make the early diagnosis of OMI and save your patient and his/her myocardium, you can get the beta version of the bot here:

We will soon be submitting a manuscript of our study of the accuracy of this AI in diagnosing OMI.  The results are astounding, as you will see!

Just before publishing this post, I received this ECG, which was the first one recorded that day in triage on arrival, 70 minutes earlier than the first ECG above:

Interpreted as "Normal" by the conventional algorithm and by the Physician

Here is the PM Cardio transformation of the above ECG (simply by taking a photo of it off the screen with iPhone!!)

It is definitely not normal.  
There are Q-waves in V4 and lead III.
There are hyperacute T-waves in V3, V4.
This is nearly diagnostic of LAD OMI, but VERY difficult to see

And the PM Cardio Bot interpretation:


Frequently, when I say an ECG is diagnostic of OMI, people who are less attuned to the subtleties of OMI ECG interpretation object and say it is "non-specific".   I tell them that it is not the ECG that is nonspecific, but the interpreter who is nonspecific.

Similarly, the PM Cardio Bot is going to make some calls that physicians just do not believe.

But beware when it makes these calls.  It is as good or better than any human.

Finally, no matter what, it pays to record frequent serial ECGs, though using the bot will get the answer faster.

My Comment by KEN GRAUER, MD (3/31/2023):
Computers are amazing machines. They have been truly transformative to our way of living — having evolved from their initial versions (which in my college days literally filled an entire room — dependent on technicians punching in complicated Fortran commands) — to current increasingly smaller, user-friendly versions.

  • I first became interested in computerized ECG interpretation in the beginning of my academic days in the early 1980s (References to some of my work appear below — as I believe I may have been the first family physician to publish in this area).
  • Like all computer functions — the machine itself is amazing — but the accuracy of computerized ECG interpretations is only as good as the data fed into computer programs. The "balance" was always how much to increase sensitivity for detection of acute cardiac conditions — vs — how much of the resultant reduction in specificity from such increased sensitivity would be acceptable. 

Computerized ECG Interpretation in Emergency Medicine in 2023:
Regarding the benefits of computerized ECG interpretation in emergency care — Everything has recently changed! Ongoing evolution of increasingly sophisticated AI (Artificial Intelligence) applications — and especially the way in which computers are now programmed to interpret acute ECG conditions has been a "game-changer". 
  • The clinical areas of most potential benefit to emergency care from computerized interpretations are: i) Cardiac arrhythmias; andii) Rapid detection of acute coronary Occlusion (ie, detection of acute OMI) in cases for which easily recognizable STEMI-criteria are not present.

  • In my opinion — AI is not yet "there" with regard to interpretation of complex cardiac arrhythmias ...
  • That said — followers of Dr. Smith's ECG Blog have already seen numerous clinical cases that we have presented in which the PM Cardio AI Bot app. has outperformed many cardiologists in its ability to recognize with "high confidence" acute OMIs from ECGs not satisfying STEMI-criteria.

NOTE: There is a reason the PM Cardio AI Bot app. has performed so well clinically: It has been programmed using ECG criteria put forth by Drs. Meyers, Weingart and Smith in their 2018 OMI Manifesto.
  • Up until recently — all computerized ECG interpretation programs that I am aware of used standard millimeter-based STEMI critieria as the basis for determining which chest pain patients should "qualify" for prompt cath with PCI. The problem with this approach — is that a minimum of 25-30% of acute coronary occlusions are missed by sticking to the old (and now outdated) STEMI-criteria approach (See My Comment the July 31, 2020 post in Dr. Smith's ECG Blog).

  • As documented by the increasing number of cases we see (and cases sent to us worldwide) — the "Queen of Hearts" app referred to above by Dr. Smith is already amazingly accurate in recognizing acute OMIs in need of prompt cath — often long before consulting cardiologists become convinced of the diagnosis. Today's case is only one such example.
  • But as good as the "Queen of Hearts" app already is — it continues to get better. This is because more and more data of cath-proven acute coronary occlusions are fed into the program, thereby continually refining sophistication and accuracy of the program.

Learning Points from Today's CASE:
For clarity and ease of comparison — I've put the 2 ECGs from today's case together in Figure-1.
  • As noted by Dr. Smith — No one could miss the acute STEMI in the repeat ECG in today's case. We have to ask WHY it took 2 hours to repeat the ECG? Even if the interventionist you are working with is not convinced that the indication for acute cath is satisfied on the initial ECG — in most cases, serial ECGs repeated as often as every 15-30 minutes until a definite answer is reached will clearly establish the diagnosis of acute OMI long before the 2 hours it took in today's case.

  • Learning Point: The RED arrows in leads V3,V4 of ECG #2 provide an excellent example of Terminal QRS Distortion (T-QRS-D). While T-QRS-D is clearly not needed for the diagnosis of a STEMI in ECG #2 — awareness of this important ECG sign may on occasion when it is seen, provide invaluable assistance for distinguishing between early repolarization vs acute OMI (See My Comment in the November 14, 2019 post in Dr. Smith's ECG Blog).

What About the Initial ECG?
The 1st point to emphasize about the initial ECG in today's case — is that it was obtained from a patient who presents to the ED with new chest pain. This history by itself immediately places today's patient in a "higher prevalence" group for an acute event. Our role immediately becomes having to rule out an acute cardiac event (rather than having to rule it "in" ).
  • I like to start my ECG assessment in patients with new chest pain by looking for at least 1 or 2 leads that I know are definitely abnormal. In a patient with new chest pain — there is no way that the ST-T wave in lead V3 can be normal. In addition to subtle ST elevation — the T wave in lead V3 is clearly disproportionate to QRS amplitude in this lead (ie, This T wave is hyperacute — as recognized by being much taller-than-expected, as well as "fatter" at-its-peak and wider-at-its-base than expected considering QRS amplitude in this lead).
  • Knowing that the ST-T wave in lead V3 is definitely abnormal — facilitates recognizing that the T waves in neighboring leads V2 and V4,V5 are also hyperacute (ie, "hypervoluminous" with respect to QRS dimensions in these leads).

  • Learning Point: Note the shape of the short ST segment in leads V4,V5,V6 before steep upslope of the T wave in these leads. This short ST segment shape is clearly abnormal! Note that this short ST segment is straightened in V4 — horizontal in V5 — and downsloping in lead V6. For me — this confirmed acuity of the ST-T waves in these leads.

  • Learning Point: The coved ST elevation in lead V1 is clearly abnormal! While very slight ST elevation is sometimes a normal finding in lead V1 — in a patient with a narrow QRS and no LVH — you should not see ST coving with the amount of ST elevation present in lead V1 of ECG #1.

  • I was less certain about QRST changes in the limb leads of ECG #1. I thought the exceeding deep and wide Q wave in lead III might represent a prior inferior infarction — especially given relatively modest ST elevation in this lead. Q waves were also present in leads II and aVF. And while considering the tiny QRS size in lead aVF — the ST-T wave in this lead could clearly represent an acute event (as could the reciprocal ST depression in lead aVL). That said — I was not initially sure about whether the inferior MI might be newold — or new superimposed on a prior inferior infarction.

  • BOTTOM Line: As per Dr. Smith, in this patient with new chest pain — ECG #1 is clearly diagnostic of acute LAD OMI. Whether or not associated acute inferior infarction was ongoing would not in any way change management and dissuade the need for prompt cath.

Learning Points: I thought it interesting to compare the repeat ECG with today's initial tracing.
  • The acute inferior wall event clearly declares itself in the repeat tracing! So in retrospect — ST-T wave findings in the limb leads of ECG #1 were indeed acute.
  • Even without the benefit of serial ECGs during the 2 hours since ECG #1 was recorded — we can see how the hyperacute chest lead T waves (as well as those abnormally shaped short ST segments in V4,V5,V6) were harbingers for the dramatic STEMI elevation that followed.
  • It becomes easy to imagine this ongoing chest lead ST elevation "lifting" up the S wave above the baseline in leads V3,V4 to produce T-QRS-D.

FINAL Reminder: As amazing as the "Queen of Hearts" app is — it is still a "tool" that needs to be used by a thinking clinician. The app isn't perfect — so all tracings still need to be overread by a human clinician. That said — this amazing technology already provides an invaluable aid to learn from — that can help expedite much more rapid recognition of acute OMIs not associated with frank ST elevation. 

Figure-1: Comparison between the 2 ECG in today's case. RED arrows in leads V3,V4 highlight T-QRS-D in ECG #2(To improve visualization — I've digitized the original ECG using PMcardio)

My Publications on Computerized ECG Interpretation from my early academic years include the following:

  • Grauer K: Chapter 21 — Does the Computer Know Better? — from Grauer K: Practical Guide to ECG Interpretation (2nd Edition) — Mosby, St. Louis, 1998, pp 374-379.
  • Grauer K, Kravitz L, Ariet M, Curry RW, Nelson WP, Marriott HJL: Potential Benefits of a Computer ECG Interpretation System for Primary Care Physicians in a Community Hospital. J Am Bd Fam Prac 1:17-24, 1989.
  • Grauer K, Kravitz L, Curry RW, Ariet M: Computerized Electrocardiogram Interpretations: Are They Useful for the Family Physician? J Fam Prac 24:39-43, 1987.
  • Grauer K, Curry RW: Chapter 11: Use of Computerized ECG Interpretation Programs — from Clinical Electrocardiography (Grauer & Curry) — Blackwell Scientific Publications, Boston, 1992, pp 418-425.

No comments:

Post a Comment

DEAR READER: I have loved receiving your comments, but I am no longer able to moderate them. Since the vast majority are SPAM, I need to moderate them all. Therefore, comments will rarely be published any more. So Sorry.

Recommended Resources