You too can have the Queen of Hearts AI model
Case 1.
Case 1 is another case sent by and written by the undergraduate who works as an EKG tech.
He sent me this great case this great case and also this one this one.
I (Smith) have made a lot of comments.
A 72 year old male with a history of “CAD with stent placement” presented to a rural hospital (PCI-incapable) with chest pain. The pain started 2 hours prior to presentation while performing manual labor. The man described the pain as substernal, constant, and non-radiating. An EKG was taken right away:
Smith: For those not familiar, the T-waves in V2-V5 (and even V6) are hyperacute T-waves (HATW). That is to say, they have a large area under the curve relative to the QRS size. They are bulky, not just tall.
Four features make up a hyperacute T-wave:
1) height,
2) width (QT interal minus QRS duration)
3) loss of upward concavity of the ST segment (making it more straight or, in really obvious cases, convex)
4) Ratio of that overall "bulkiness" to the QRS size (height or AUC or both of the QRS)
The HATW in this case are a subset of the much larger set of HATW called De Winter’s T waves. De Winter's T-waves are anterior HATW with a depressed ST takeoff (preceded by ST depression). Such HATW with depressed ST takeoff can also occur in inferior and/or lateral leads if those are the infarct territories, but de Winter only described the LAD version.
Of important note: there is still well preserved R wave progression. Immediate PCI on a patient with this EKG may save a lot of still-viable myocardium.
The emergency provider at the rural hospital did not recognize this EKG. They noted “T waves in V2 and V3 are somewhat tall. I do not have an old for comparison”.
The Queen of Hearts recognizes the De Winter’s phenomenon:
Because the EKG went unrecognized, the patient had to rely on troponins to catch the provider’s attention. Unfortunately, the first troponin I, drawn on arrival, resulted at only 0.013 (URL for this assay is 0.034).
Smith: At least 25% of full blown STEMI have an initially negative troponin. YOU CANNOT USE TROPONIN TO RULE OUT OMI!!!
There is not much information about his stay in the ED. The patient was given aspirin and nitroglycerin without much pain relief. The patient was eventually given morphine and the pain subsided.
Smith: Never give morphine in ACS unless you are already committed to the cath lab. It just delays reperfusion
There is a bit more information on the patient’s past MI. Twenty years ago, the patient received a stent to his proximal LAD for a 95% occlusion. A formal echocardiogram after the MI showed an ejection fraction of 50%.
After 4 hours of waiting in the ED, a second troponin I came back at 0.296 ng/mL.
This prompted a repeat EKG:
The ED provider updated their impression: “Chest pain with initial negative troponin and unremarkable EKG, now with elevated troponin and EKG concerning for STEMI”
The emergency provider called for a STAT cardiology consult with the nearest PCI-capable facility. Cardiology accepted the patient for transfer, noting “He presented to [the rural facility] where his first EKG revealed peaked T waves, but no ST elevation. His subsequent EKG revealed ST elevation”.
The
patient arrived to the PCI-capable facility by helicopter, but had to
wait in the ED for another case to finish in the cath lab. Another EKG
was taken while he waited:
Now the tiny R wave in V4 has disappeared.
Reciprocal ST depression has developed in the inferior leads.
Smith: this ECG with QS-waves shows a nearly completed anterior MI. A huge amount of myocardium is now infarcted (dead, necrotic).
A troponin T was drawn while he waited. It resulted at 504 ng/L (the URL is 16 for men in this assay).
The patient was finally brought to the cath lab about 7 hours after he first presented to the rural hospital and 9 hours after the onset of his chest pain. A 99% in stent-restenosis of the proximal LAD with TIMI 1 flow was found and stented.
A post-cath EKG was taken:
This is an ECG that is typical of a left ventricular aneurysm, though an actual anatomic aneurysm takes weeks to form. This is a terrible ECG.
Unfortunately, there are now no R waves in any of the precordial leads! Contrast this with the patient’s presenting EKG at the rural facility, where R wave progression is more or less normal. Also notable is the new rightward axis. The depolarization vector towards the left ventricle has dwindled enough to shift the mean ventricular depolarization vector towards the right ventricle.
After catheterization, another troponin T was drawn. It resulted at >10,000 ng/L. A value >1,000 ng/L is typical of an OMI. This is more than 10x as high. This was a massive OMI.
A formal echocardiogram done the day after catheterization showed:
Severely reduced global systolic function with an estimated EF of 20-25%
Mid and distal anterior wall, entire apex, and mid anterior septum segment are abnormal- consistent with large LAD infarct
This man lost 50-60% of his ejection fraction to this OMI. Of course, some of his EF may recover, but with a troponin T above 10,000, much of this damage will be permanent.
The patient is undergoing evaluation for heart failure.
Case 2
This was sent by a good friend with whom I trained in EM residency 34 years ago. He works at a small hospital without PCI capability. He is very very smart and good with EKGs.
This a 70 year old man has no h/o coronary disease, but does have a history of ablation of atrial fibrillation 10 years ago. He presented with chest pain and tachycardia. He was in atrial fibrillation with a ventricular rate of 170. My friend did not send that EKG. My friend gave the patient amiodarone and his rate decreased.
He then obtained this ECG:
Angiogram done 2 days later showed a culprit in the proximal LAD with 95% stenosis, stented. This is of course consistent with a reperfused LAD OMI.
Here is the Queen of Hearts interpretation of both ECGs, with explainability:
First ECG, with full occlusion:
2nd ECG after resolution of chest pain, with reperfusion:
You too can have the Queen of Hearts AI model
My friend asked why it is that cardiologists so often miss these NSTEMI-OMI.
I can only speculate:
1) they have an enormous specialty and a lot to know
2) they believe that the ECG is nonspecific. If they do not see OMI on the ECG even when it is there, that confirms the bias. (It is really the interpreter, not the ECG, which is nonspecific)
3) reading OMI on the ECG is unbelievably hard, requires years and years of intense study with known outcomes,
4) one needs to understand how hard it is and that one probably does not know how to do it very well. That is, one must start with total humility.
5) Cardiologists do not see patients with undifferentiated chest pain.
6) The culture seems to be to avoid false positives more than avoiding false negatives.
7) a true belief (in spite of contrary evidence) that any patient who needs emergent coronary intervention must hav an ECG that meets STEMI criteria.
8) The idea of OMI is new.
MY Comment, by KEN GRAUER, MD (6/3/2024):
- For clarity in Figure-1 — I have labeled the initial ECG, and added a laddergram to illustrate the mechanism of the rhythm
 |
| Figure-1: I’ve labeled the initial ECG in Cast #2 — and have added a laddergram. (To improve visualization — I've digitized the original ECG using PMcardio). |
- The reason diagnosis of acute LAD OMI is challenging in Case #2 — is superposition of flutter waves on top of the underlying ST-T wave. That said — RED arrows in leads V1 and V2 in Figure-1 clearly show ST segment coving and elevation (that should not be there with RBBB — which normally shows ST depression in these leads). Careful scrutiny of the ECG baseline in these leads (within the RED rectangle) reveals a clear difference compared to the baseline away from the ST segments.
- More subtle, but still definitely present — is the ST elevation also in lead aVL (RED arrows in this lead) — with this subtle ST elevation in aVL most noticeable when compared to the ECG baseline away from the QRS in lead aVL.
- BLUE arrows in lead III and lead aVF — reveal subtle-but-definitely present reciprocal ST depression (again, best appreciated when the "dip" in the ST segments highlighted by the BLUE arrows in these leads is compared to the ECG baseline away from QRS complexes).
- BOTTOM Line: The combination of abnormal ST coving and elevation in leads V1,V2 and aVL — in association with reciprocal ST depression in inferior leads III and aVF — in this patient with new chest pain — is absolutely diagnostic of proximal LAD OMI.
- Although not all R-R intervals in Figure-1 are equal — there are only 2 different R-R interval durations — and these 2 R-R interval durations repeat! The shorter R-R interval separates beats #2-3; 4-5; and 7-8. The longer R-R interval separates beats #1-2; 3-4; 5-6; 6-7; 8-9; 9-10; 10-11; 11-12; and 12-13.
- KEY Point: As I've illustrated in a number of previous Blog posts — AFlutter commonly manifests Wenckebach conduction. Technically — this is not a conduction "block", but rather expected reduced conduction by an AV node presented with extremely rapid flutter impulses that are too fast to conduct them all! (~300/minute in today's case!).
- More than just 2 families of similar R-R intervals — the PR interval in front of each of the QRS complexes in the larger R-R interval group is the same! (ie, the PR interval that begins with each of the BLUE lines).
- Similarly — the PR interval in front of each of the QRS complexes in the shorter R-R interval group is also the same! (ie, the PR interval that begins with each of the PURPLE lines).
- PEARL: It took me less than 5 seconds to recognize this repetitive relationship between these 2 families of R-R intervals — and the finding of 2 similar PR intervals preceding both shorter and longer R-R intervals. This is not by chance — and with AFlutter, is virtually diagnostic of dual-level AV Wenckebach conduction in a patient with AFlutter! (See My Comment and laddergram in the November 24, 2020 post and the May 26, 2023 post, regarding the concept of dual AV nodal Wenckebach conduction).
- I've drawn a laddergram in Figure-1 to illustrate my proposed mechanism for this rhythm. Keep in mind that because of "concealed" conduction (ie, atrial impulses bombarding the AV node) — it is usually not the flutter wave closest to a QRS that conducts. Thus, there is dual-level Wencekbach conduction out of the AV node — with 2:1 conduction in the upper AV nodal level — and 3:2 alternating with 2:1 conduction in the lower AV nodal level.
- To EMPHASIZE — There is no need to dwell on the sophisticated mechanism of the rhythm in this 2nd Case. What I do want to emphasize — is to appreciate that AFlutter (as well as ATach) will commonly manifest Wenckebach conduction at 1 or more levels out of the AV node — and that simply by appreciating that a pattern does exist to the rhythm (as I illustrate above) — you can within seconds identify Wenckebach conduction as an expected consequence of these rhythms.
- Finally, to emphasize — This Wenckebach conduction out of the AV node is not an AV "block" per se — and chances are excellent that normal AV conduction will resume once sinus rhythm is restored!
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.