Wednesday, July 5, 2023

This was texted to me in real time. The patient has acute chest pain.

 This was texted to me in real time. The patient has acute chest pain.

What do you think?

Here was my answer:

"Not ischemia. Chronic. Maybe HOCM or another form of LVH.  I would not activate cath lab.  Get serial troponins"

It is a scary ECG, with a lot of ST Elevation and what appear to be hyperacute T-waves in inferior leads, and profound reciprocal ST Depression in aVL.  There are Q-waves in V4-V6, with what appear to be hyperacute T-waves.  

Any objective, rule-based analysis of this ECG would scream "STEMI" or "OMI".   But, alas, ECGs are like faces.  No measurements can tell you the identity of the face.  It must be recognized.  

And I recognized this as a STEMI mimic.  I told them so and thus they did NOT activate the cath lab, but measured serial troponins.

The patient ruled out for MI by serial troponins.  Further followup is below.

We are training the Queen of Hearts AI algorithm to recognized these cases, but we need to train her with a lot more false positives.

In this case, QOH was incorrect, saying "OMI High confidence."

I sent this ECG to the Queen of Hearts (PMcardio OMI), and here is the verdict:

Follow Up:

The young man had no CAD on his CT. HS Trop-I of 15 and 19. TTE below. Output cardiac MRI pending.

1. The left ventricle is normal in size. There is significant hypertrophy      
 noted of the anterior, anteroseptal and anterolateral wall segments. The left ventricular systolic function is lower limits of normal to mildly reduced with a visually estimated LVEF of 50-55%.                              
 2. The average global longintudinal strain value is -14.5% which is mildly abnormal with decreases strain noted in the anterior, anterolateral and anteroseptal segments.                                                         
Overall, findings may be consistent with an infiltrate cardiomyopathy.  May consider further evaluation with cardiac MRI based on clinical concern. 

MY Comment, by KEN GRAUER, MD (7/5/2023):
Today's case is instructive for a number of reasons: i) As good as the QOH (Queen OHearts) AI algorithm already is — it is not perfect; — andii) Both us as clinicians, as well as QOH — will continue to be challenged by "pseudo-infarct" patterns.
  • For clarity — I've reproduced today's ECG in Figure-1having deleted the distraction of the 3 long lead rhythm strips that appeared below the original ECG with confusing QRS overlap.

Figure-1: The initial ECG in today's case. (To improve visualization — I've digitized the original ECG using PMcardio).

WHY Did QOH Get It Wrong?
I do not view the wrong diagnosis put forth by QOH as an "error" in the usual sense of the word. Instead — this is a totally understandable response for an AI application, because: 
  • i) "Pseudo-infarct" patterns can be tricky. As a result — it will often not be possible to make a definitive diagnosis on the basis of a single ECG, especially in the absence of sufficient clinical correlation. 
  • ii) Many "pseudo-infarct" patterns are unique in the ECG findings that they manifest. As a result — there are a limited number of "practice" tracings with cath-confirmed diagnosis to represent the many potential ECG variations for any clinical condition that might produce a "pseudo-infarct" pattern. As might be imagined — this makes it extremely challenging to adequately "train" QOH by exposure to a sufficient number of practice tracings to which we know the answer.

WHY Did WE Get It Right?
I looked at today's tracing before reading Dr. Smith's response. My initial impression of this ECG was the same as Dr. Smith's — in that despite the alarming ST-T wave changes, I did not think ECG #1 was the result of an acute event. Instead — my thoughts were as follows:
  • The rhythm is sinus, with marked bradycardia and a component of sinus arrhythmia.
  • QRS amplitude is dramatically increased in a number of leads. This is most marked by the 26 mm R wave in lead I — with an almost equally deep S wave in lead III (in which the QRS complex is cut off by the bottom of the tracing).
  • QRS amplitude is also generous in the other limb leads — and in leads V1,V2 which manifest confusing overlap. The 10 mm R wave in lead V1 is excessive in an adult — with "eye-catching" early transition in the chest leads (ie, the R wave becomes predominant already by lead V2).
  • Surprisingly deep (albeit narrow) Q waves are present in no less than 9/12 leads! (ie, in leads I,II,III; aVL,aVF; and in V3-thru-V6). These narrow Q waves are surprisingly deep in leads V4,5,6 — in which the R that had become so positive early in transition barely reaches the baseline.
  • T waves are prominent in virtually all leads — with inferior lead, upward sloping ST elevation and disproportionate T wave positivity in the inferior and lateral chest leads. ST depression with a mirror-image opposite shape of the depressed ST-T waves is seen in high-lateral leads I,aVL compared to the ST-T wave picture in leads III and aVF.
  • There is almost "giant" T wave inversion in lead V2 (ie, the inverted T wave = 7 mm in this lead).

Putting It All Together:
  • The multiple unusual ECG findings on today's tracing "just don't fit" for acute OMI. 
  • As per Dr. Smith — I thought the enormous R wave amplitude with ST depression in lead I was diagnostic of some form of LVH
  • In the absence of QRS widening — the marked positivity of the R wave in lead V1 suggests either increased septal forces or RVH
  • The presence of deep, slender Q waves in so many leads — especially with the bizarre QrS morphology in leads V4,5,6 (with tiny r waves despite marked LVH in lead I) seems much more suggestive of some unusual form of underlying structural heart disease than of an acute MI pattern. 
  • I thought the very deep T wave inversion in lead V2 might simply reflect a "transition lead", given surprising R wave positivity in neighboring lead V1 — and bizarre R-wave lacking leads V3-thru-V6.
  • Despite the marked (>2 mm) ST segment deviation in 5 of the limb leads — I thought the shape of the ST-T waves in these leads much more consistent with LVH than of acute infarction.
Is there any Potential "Unifying" Diagnosis?
If I had to pick a single diagnosis for today's tracing — I'd pick HCM ( = Hypertrophic CardioMyopathy)Most patients with HCM do not have a normal ECG. Among the many ECG findings that may be seen in patients with HCM are the following: 
  • Increases in QRS amplitude.
  • Large septal Q waves (Sometimes known as "dagger" Q waves — because these are deep but narrow Q waves seen in lateral leads). 
  • Tall R wave in lead V1 and/or early transition in the chest leads (reflecting increased "septal" forces).
  • Abnormal ST-T wave abnormalities.
  • Conduction defects (ie, LBBB, IVCD).
  • WPW
  • Cardiac arrhythmias (especially AFib). 

  • The Problem: None of the above ECG findings are specific for HCM. It is also interesting (if not confusing) — how much of a variety one may see on the ECG of a patient with HCM. That said — ECG findings listed in the first 4 bullets above are found in today's tracing! HCM could explain all of the findings we see in ECG #1.
  • If not HCM — some unusual form of cardiomyopathy might explain the findings in today's ECG (ie, muscular dystrophy; infiltrative heart disease from amyloid or sarcoid; some unusual form of congenital heart disease, etc.).

  • The "Good" News: Echocardiography is likely to be diagnostic. If needed — cardiac MRI should provide any answers not explained by Echo.

  • Additional "Take Home" Point: When QOH makes a mistake for prediction of acute OMI — Try to figure out WHY! Optimal clinical assessment of today's patient is simply not possible from the single ECG shown. Instead — I would bet that "the Answer" to today's case will become obvious with a little bit of history — and with an Echo (possibly also a cardiac MRI). In my opinion — it's not enough to limit our assessment to "OMI or no OMI" without considering what the diagnosis is in non-OMI patients. 

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