ECG Podcast-A.Kashou - Comparison ECGs-Computer-AI (Parts 3,4- 11-23.1-2023)-DRAFT

These 2 Podcasts are for Thursday, Dec. 14, 2023 @ 2pm EST !

=========================

This is the outline for my 2-Part Podcast on “Computerized ECG Interpretations and AI in 2024” and "All About Comparison ECGs for 12-leads and Arrhythmias". I will record these 2 sessions for Dr. Anthony Kashou on December 14, 2023.

• To see video of this podcast — CLICK HERE — 

=========================

TITLES of each Podcast:

• PART 3: — "Computerized ECG Interpretations and AI in 2024".
• 
  
• PART 4: — "All About Comparison ECGs for 12-Leads and Arrhythmias".

Brief “C.V.” for Dr. Kashou to Introduce Me (Could use same or similar Intro for both parts — Anthony, feel free to modify as you see fit! ):

• Dr. Ken Grauer is with us today. He is Professor Emeritus in Family Medicine. Following his Residency in Family Medicine — he worked for 2 years in a busy Florida ED in south Florida — before moving to Gainesville, where he was full-time Faculty in the University of Florida, Family Medicine Residency Program for his 30-year career, until he retired from his academic practice in 2010. Dr. Grauer has written over 10 books on ECG and Arrhythmia interpretation — presented literally hundreds of talks and workshops locally and nationally on ECG Interpretation and other cardiology topics over his career — and has been as active-as-ever since “retiring” in 2010 — with over 3 million views on his own ECG Blog — as an Associate Editor and active contributor to Dr. Stephen Smith’s ECG Blog (that many of you probably follow) — as well as answering numerous daily queries addressed to him on any of the many international ECG internet forums that he regularly contributes to.  — Dr. Ken Grauer …

QUESTIONS for ANTHONY KASHOU — for Part 3: (Draft done on 11/23)

• Ken — When did you first become interested in the Computerized Interpretation of ECGs?
• Is the Computerized Interpretation of ECGs a "GOOD" or "BAD" thing?
• How are current developments in AI changing the way we use computer applications to assist in the interpretation of ECGs?
• Should the clinician ever look at the AI interpretation before they make their own initial impression?
• Will there ever be an AI application able to assess complex arrhythmias?

My Answers to Part 3 Questions:

• ANTHONY — I will pretty much be on "autopilot" after your 1st Question — so you probably won't need to ask much else ... (although I of course always WELCOME any points you may want to raise!).
• The "flow" I am aiming for after you ask "When did I first become interested in Computerized Interpretation of ECGs?" — is to review the insights I gained from my study of computerized interpretations in the 1980s — and how those insights have served me (and evolved) over the years.
• Optimal use of the computer will vary depending on the skill and experience level of the interpreter.
• EVERYTHING has now changed — as per your interview a while back with Robert Herman (of Powerful Medical) — especially with ongoing development of his QOH ( = Queen Of Hearts) application for detection of OMIs that do not yet qualify as STEMIs.
• The potential of AI (via QOH and other applications) is HUGE — but I believe there will always be need for clinician oversight to ensure AI and QOH clinical accuracy!
• Clinicians should always make their own independent interpretation before they look at what the AI app says.
• An important adantage of the AI ECG interpretation is that it provides a 2nd opinion — and — 2 opinions ( = the Clinician's interpretation and the AI interpretation) — are better than one!
• I do not believe ( = my opinion) — there will never be an AI app able to supersede a true expert in arrhythmia interpretation (just too many variables to ever attain a large enough data base to cover all possibilities and rule out all forms of artifact and lead misplacements ... ).
• The above said — AI will continue to get better and better!

QUESTIONS for ANTHONY KASHOU — for Part 4: — (Draft done on 11/23)

• Ken — What can we learn from clinically-correlated Comparison of ECGs? — and — What is the optimal technique for comparing one tracing with another?
• Why is the clinical history so essential when comparing one tracing with another.
• What are some applications of comparison tracings that may be helpful with cardiac arrhythmias?

My Answers to Part 4 Questions:

• ANTHONY — I will pretty much be on "autopilot" after your 1st Question — so you probably won't need to ask much else ... (although I of course always WELCOME any points you may want to raise!).
• It’s essential to correlate clinically what was happening at the time each tracing you are comparing was recorded (ie, Was the ECG a "true baseline" — or was the patient having chest pain at the time?).
• Optimal technique for comparing 2 ECGs — is to put both 12-lead tracings you are comparing next to each other — and then to compare lead-by-lead.
• Depending on what you are looking for — comparison tracings may be very helpful in arrhythmia assessmen (ie, for finding P waves and/or atrial activity in other leads — for knowing what atrial activity and QRS morphology looked like in the past with sinus rhythm — for comparing QRS morphology during a WCT to baseline QRS morphology during sinus rhythm).

What the Learner Should Take from Today’s Podcast? = for Part 3 — (Draft done on 11/23):

“Take-Home” Points I hope to emphasize with this Podcast on "Computerized ECG Interpretations and AI in 2024" — include the following:

• Pros and Cons of currently available computerized ECG interpretation programs.
• The need for the clinician to always make their own independent interpretation of any tracing before looking at what the computer or AI application said.
• The way I envision how AI may help to improve clinician recognition of acute OMI (Occlusion-based MIs). 

What the Learner Should Take from Today’s Podcast? = for Part 4 — (Draft done on 11/23):

“Take-Home” Points I hope to emphasize with this Podcast on "All About Comparison ECGs for 12-Leads and Arrhythmias" — include the following:

• The essentials for optimal technique when comparing one tracing with another.
• The importance of correlating the clinical history to each tracing interpreted. 
• Ways in which comparing one tracing to another may help clinically not only with assessment of 12-lead ECGs — but also with interpretation of cardiac arrhythmias.

========================

QOH APP — LINK = https://www.powerfulmedical.com/about

========================

• Highlighting KEY aspects of 12-lead ECG interpretation, as this relates to the diagnosis of acute ischemia, and especially to acute coronary occlusion. 
• To convey the concept of "OMI", emphasizing  appreciation of how the older (and outdated) STEMI paradigm misses at least 30% of acute coronary occlusions that can be picked up by using the newer OMI paradigm.
• Citing the ECG findings that facilitate early identification of an acute OMI — which include recognition of hyperacute T waves — looking for “mirror-image” ST-T wave changes (ie, between leads III and aVL with inferior OMI — and in leads V1-thru-V4 for posterior OMI) — and optimizing use of serial ECGs, with time-related correlation of each ECG to the presence and relative severity of chest pain.
• Applying a time-efficient technique for how to compare one ECG to another.

=================================

PART 3: Computerized ECG Interpretations and AI in 2024

==================================

Herman, Meyers, Smith et al — Eur Heart J: Digital Health — Nov. 2023

https://academic.oup.com/ehjdh/advance-article/doi/10.1093/ehjdh/ztad074/7453297 

An AI model (QOH) was developed using 18,000+ ECGs from patients with suspected ACS from an international data base — comparing STEMI criteria vs ECG experts for detecting OMI. Primary outcome of OMI was an acutely occluded or flow-limitin g culprit artery requiring emergent revascularization.

• QOH had superior performance (80% sensitivity; 93% specificity — vs STEMI criteria (32% sensitivity; 97% specificity).

=========================================

=========================================

Hi Anthony. THANKS for your invitation to talk about Computerized ECG Interpretations — including a look at “How AI might benefit ECG interpretation in 2024”. 

• DISCLAIMER:  — What follows is based on MY Observations — and represents MY “Opinion”.  Others may disagree — and that’s fine. My GOAL is to stimulate dialog.
• Application of AI (Artificial Intelligence) is creeping into our lives everywhere. To ME — This is at the same time tremendously exciting — yet still a “scary” topic in many of the areas in which it is used. The potential of AI for good is limitless — but so is it’s potential for misuse.
• 
  
• Regarding this potential for misuse — We need look no further than “DEEP-FAKE” — which is the AI-based technique that synthesizes media to allow superimposing human features — actions — and the voice of 1 person — with end-result of a very realistic-looking likeness or impersonation.
• In Politics — You can make it look like your opponent is saying whatever you want him (or her) to say ….
• 
  
• BOTTOM Line: We all are seeing this increasing influence of AI on our daily lives.

From my personal perspective:

• There is LOTS of GOOD that’s come with AI!
• For Example — In the area of 1 of my “passions” ( = foreign language study) — AI translations in almost any language now occur near instantaneously (either from the written word — or from scanned text — you can put in a screen shot of text) — with AI able to translate almost in “real time” — and then speak with surprisingly excellent foreign accents — just like I used to see on Star Trek: New Generation when Starship Enterprise would encounter a new alien nation — and their computers would instantly translate numerou alien languages.
• 
  
• That’s the GOOD. A personal example for ME of the “BAD” — I recently was involved in an ECG educational platform that regularly employed BOTS to generate via AI from my presentations my GOALS — Objectives and a SUMMARY of my talk. And while the AI-generated material “sounded good” — Suffice it to say that the errors made by AI were numerous, embarrassing — and alarming. 
• For general commentary — AI Bot-generated commentary can sometimes be quite good. But in my opinion — It was very inaccurate in its comments about my presentation — and it made a LOT of extra work for me, because it kept on generating and regenerating erroneous commentary that I had to correct ...

So — WHAT has this all got to do with ECG interpretation?

• FIRST and FOREMOST: The “BOTTOM LINE” that I’ll be working up to is my strong opinion that while the potential Benefit of AI to assist in certain aspects of ECG interpretation is HUGE …
•   — I don’t think ( my opinion) — that we will ever totally eliminate the need for at least some human oversight (just as there remains a need for at least some human overight for almost everything accomplished by computer).
• 
  
• NOW — the amount of human oversight needed will continue to decrease — as AI programs for ECG interpretation continue to be “trained” with larger and larger data sets.
• But just as occurs for virtually any other computerized function — There should always be some human oversight to ensure that the computer is doing what it is supposed to be doing ...
• 
  
• The “2nd BOTTOM LINE” — is that the combination of AI plus Human oversight by a capable ECG interpreter will be better than either AI alone or Human interpretation alone …. — so THIS is what we are aiming for!
• 
  
• The "3rd BOTTOM LINE"  — is that from what I've observed reviewing countless tracings daily interpreted by medical providers of all levels of experience (including cardiologists) — a number of fundamental errors and oversight continue to be made. Recognition of these oversights CAN be taught  — and AI-guided interpretation can help to teach those clinicians who are open to this feedback.

So — WHERE are we "going" with respect to AI computer interpretations?

• To answer this question — I think it helps to look from where we’ve come in the field …
• 
  
• Use of computerized ECG interpretations is NOT new. 
• It has been around for decades !!! 
• 
  
• My involvement with computerized ECG interpretations began in the 80s — when as full-time faculty in our Family Mediine Residency Program — I was charged with interpretation and oversight of all interpretations of our 35 medical providers (25 Resident + up to 10 Faculty ...).
• My oversight was for all of the ECGs done in our ambulatory clinic — and for many of the ECGs done on our hospitalized patients.

What I Observed — was an increasing tendency to depend on, “What did the computer say?”

• IF the computer said there was an acute MI — then that is what the provider believed …
• IF the computer did not see an MI (but said, “Normal tracing” ) — then that  was what the provider believed.
• 
  
• This was not only a local phenomenon that occurred within our Training Program — but as a national educator — I was seeing this as a national trend.
• 
  
• Medical providers more and more — were looking 1st at what the computer said — and believing (without questioning) what the computer said — and that was a KEY mistake!
• 
  
• What I did not see — was any guidance for HOW best to use the computer. 
•   — Are the needs of the novice the same as an intermediate or advanced interpreter? No one addressed this …

===========================

How BEST to use the Computer should be obvious. What would YOU expect the computer to be real good at?

• Measurements — Intervals! — the PR interval — QRS duration — measuring the QTc — Axis …
•   — As long as the computer correctly determines where the interval begins and ends — it will be very accurate with calculations …
• 
  
• What will be more difficult for the computer?
•   — Incorporating the HISTORY 
•   — comparing SERIAL TRACINGS
•  — Complex cardiac arrhythmias (The computer can pick up the irregular irregularity of AFib — but HOW are you going to ever program all potential complexities of unusual AV blocks in which atrial activity is far from obvious?)

So — To make a long story short — I studied this in the mid-1980s ... (perhaps being the only primary care clinician to formally study computerized ECG interpretations).

• Motivation for my small study — was that I hated computer interpretations.
• I saw them being widely misused — and I wanted to prove to my residents (as well as to primary care clinicians nationwide) — that there was more harm than good from computer interpretations. 

Surprisingly — After completing my study — I learned to LOVE computer interpretations!

• WHY?
•   — Because I finally understood how BEST to use the Computer!
• 
  
• Optimal use of the computer to assist with ECG Interpretation — just like optimal use of the computer for any purpose — depend on WHAT your goals are …

What I found in my small Study:

• As an EXPERT interpreter — I found the computer GREATLY increased my SPEED for interpreting tracings.
• 
  
• I no longer had to calculate Rate — Axis — Intervals (The computer did this!).
• IF the computer said the tracing was normal — All it took me was a few seconds to take a quick look — and then sign my name ... 
• 
  
• BUT — If there were significant abnormalities — I'd then have to review things more closely ...

What I also found in my Study — was that the Benefits of computerized ECG interpretation — depend on the expertise of the interpreter.

• As an Expert — and I call myself an "expert" — because I've spent my academic career (and the past 40+ years of my life) — studying and teaching ECG interpretation — 
•   — So as an expert — The computer can save me time.
• 
  
• For non-experts (which is >90% of clinicians reading ECGs) — the main benefit of the computer is the extra Opinion that it provides.
• 
  
• To EMPHASIZE — Many of those who I call "non-experts" — are good interpreters, but they don't have decades of experience reading countless tracings under all different circumstances — so these interpreters CAN still benefit more from the extra Opinion that AI can provide.

THAT was THEN ... — What about NOW?

• Lets looks at — What AI can offer us in 2024 as advantages for ECG interpretation?

==============================

I'll make a number of points about this ...

Previous ECG computer programs are of limited benefit in 2024 — especially with regard to emergency care.

• Prevous ECG computer programs may still provide a useful 2nd opinion with regard to measurements — and with regard to recognizing chamber enlargement and conduction defects ....

BUT — Previous ECG Computer Programs are not helpful (and are potentially harmful) in Emergency Care in 2024 — because:

• #1) They STILL miss too many arrhythmias to be relied upon ...
• 
  
• #2) They are not GOOD at diagnosing Acute Coronary Occlusion
•                — Far too many false positives and false negatives!
• 
  
• As to Terminology — From this point forth, I'm going to refer to acute coronary occlusion — as an OMI (Occlusion Myocardial infarction)

What MATTERS in Emergency Care?

• What matters — is IF a patient with CP has an OMI or not? (ie, acute coronary occlusion) — because IF a patient with new CP has an OMI — then we KNOW that they can benefit from prompt cardiac cath and reperfusion with PCI.
• 
  
•   — and there IS a GOLD Standard for determining IF a patient with new CP does or does not have acute coronary occlusion (ie, IF you do prompt cath — and IF there is a "culprit" vessel that is closed — then this is an OMI! ... ).
• 
  
• So — This has been studied!
• IF you depend on millimeter-based STEMI criteria for determining acute coronary oclcusion — You'll miss at least 30% of all acute OMIs!

As I talked about in my Podcast with you about ERRORS in 12-lead ECG Interpretation — certain ECG features in a patient with new CP — DO correlate well with OMI ( = ie, with Acute Coronary Occlusion). These ECG features include:

• Hyperacute T waves
• Max ST depression in lead V2,V3 or V4
• Seeing the "magical" mirror-image opposite reciprocal relationship between ST elevation in lead III — and reciprocal ST depression in lead aVL.
• Seeing dynamic ST-T wave changes on serial ECGs ...
• 
  
• And you can program AI to detect these changes of acute OMI (as Dr. Robert Herman shared with us on a previous Podcast you did with him).
• 
  
• Drs. Stephen Smith and Pendell Meyers — have been doing precisely this for the past few years —  teaching the AI program (Dr. Herman's program is known as QOH = Queen Of Hearts). 
• Drs. Smith & Meyers — they have literally been programming thousands of tracings with cath-proven findings on HOW to recognize acute OMIs.
• 
  
• And THIS is the FUTURE as I see it for AI with ECG interpretation. And the future is NOW ... (ie, The QOH AI program is already available in Europe — and will be available in the U.S. as soon as it is approved by the FDA ....).

I'll conclude with the following — To EMPHASIZE:

• There are still all-too-many clinicians (including all-too-many cardiologists) worldwide who are "stuck" on the STEMI paradigm that is based on satisfying millimeter-based criteria for ST elevation — despite the fact that doing so will miss at least 30% of all acute OMIs!
• 
  
• The QOH AI program instantly becomes SUPERIOR to those cardiologists who are "stuck" on the STEMI Paradigm — for picking up acute OMIs ....
• 
  
• To EMPHASIZE — the AI QOH program has "learned" how to recognize acute OMIs by being fed data from thousands of tracings with cath-proven determination whether or not there was acute coronary occlusion.
• 
  
• Therefore — Clinicians with lesser experience can learn from recommendations put for by QOH.
• 
  
• That said — in my opinion — there still needs to be clinician oversight — to ensure technical issues are not missed. QOH isn't perfect (She does not yet integrate the clinical history — and she doesn't yet compare tracings — ).  This is coming with future versions of QOH  — but it is not here yet ...
• 
  
• As an expert — I find that I depend less on QOH. That said — I definitely find it comforting when the QOH AI interpretation agrees with my impression.
• And IF the QOH interpretation is different than mine — this helps me — because it tells me to go back and make sure that I wasn't missing something.

FINAL Thoughts: — 

• This last point is the reason QOH continues to get better — because if QOH misses a diagnosis that Drs. Smith and Meyers see — they go back and reprogram QOH  so that she "learns" from her mistakes.
• 
  
• Future versions of QOH will continue to improve — as Drs. Smith and Meyers continue to add to her data set!
• 
  
• That said — My strong feeling is that clinicians can learn to greatly improve their ability to recognize acute OMI much earlier in the process. 
• This is not too difficult.
• Clinicians merely need to be open to feedback — and to accept new literature that clearly shows the STEMI Paradigm is outdated.
• 
  
• And — The BEST way for clinicians to get better — is to always force themselves to interpret the ECG 1st by themselves — and then to profit from the 2nd opinion that the AI interpretation may provide.

================================

Can AI “Fix” these Errors?

• AI is a powerful tool! But there are PROs and CONs to its use …
• Up until recently — computerized programs offered the advantage of providing an additional opinion — in that they make you STOP and think about your interpretation if it differs from the computer interpretation.
• But such computer ECG reports were not optimally accurate. They were good for number, intervals, axis — but in my experience — routinely missed complex rhythms and subtle MIs
• 
  
• A while back — you did a podcast with Robert Herman of Powerful Medical. Robert’s company is in process of changing the “playing field” — with assistance from Drs. Stephen Smith and Pendell Meyers, who are helping to program the AI Bot known as “Queen of Hearts” to pick up subtle signs of acute coronary occlusion that even cardiologists miss.
• The potential of this new AI app is tremendous! 
• That said — I still believe that there always needs to be clinician oversight to ensure that AI readings make sense. 
• Like any computer — AI only can do what it is programmed to do — and we’ve not yet figured out all of the data that needs to be programmed. So VERY PROMISING — and ALREADY very good at picking up OMIs — but not yet perfect.
• 
  
• My Opinion — there will never be an AI program able to pick up all complex arrhythmias. I feel there are simply too many rhythm possibilities — and as one who loves interpreting complex arrhythmias — You often have to think “out of the box” (with there not being enough data to provide adequate experience in this to AI programs).

What is Your Experience in Teaching ECG Interpretation?

• XXXX
• I “get sent” tracings 

WEBINAR:

Robert Herman, MD — Powerful Medical (SSmith, Pendell Meyers)

• Only 1/3 of OMIs are STEMIs.
• Requiring strict STEMI criteria will miss at least 30% of OMIs
• The focus of his ECG-based AI model is on detecting OMI vs NOMI (and not on STEMI vs NSTEMI).
• The data base is correlated to cath data showing acute coronary occlusion.
• Huge data base to train AI systems.
• Similar to AI "Fake" — in that you can recognize a face, but it may be hard to describe the features ....
• RESULTS — QOH (programmed by Smith/Meyers) — 2X higher sensitivity in picking up the angiographic endpoint ( = acute occlusion) — and does it up to 3 hours faster (compared to detection time in patients who ultimately DO develop a STEMI).
• QOH — Version #1 does not compare to prior ECGs — or compare to serial ECGs — or incorporate the history — nor does Version #1 distinguish between acute OMI vs a recent and now reperfuse OMI ...
• AI QOH will work best with a "high prevalence" population (ie, someone who presents with new CP who is "higher risk" )
• Posterior Leads — another reason why they may show LESS ST depression — is that they are often applied a little bit later (ie, at a time when there has been some reperfusion — so it is not "negative" for STE — but rather now a reperfused OMI)

 

ECG FEATURES:

• Hyperacute T waves 
• Pathologic Q waves, along with subtle STE
• Terminal QRS Distortion
• Reciprocal STD and/or Reciprocal T wave inversion
• Subtle STE not meeting criteria, but with other features
• ANY amount of STD maximal in V1-V4
• ANy STE in inferior leds with any STD/T wave inversion in aVL

==========================

ECG Blog #207

======================

Computerized ECG Interpretations: Friend or Foe

A frequent question that arises is, “How best to use (or not use) the computerized ECG interpretation?” Opinions vary. We feel the answer depends on the goals and experience level of the interpreter.

• Computerized ECG analysis systems are not infallible. They are far from perfect at ECG interpretation. Our task is to appreciate what computer systems do well — and what they do not do well.

At the current time — virtually all modern ECG machines automatically provide a computerized interpretation. This has benefits and drawbacks. Consider the following:

• Computerized systems excel at computing values. That’s what computers do. As a result — computerized systems are extremely accurate in calculating: i) Rate; ii) Intervals (PR/QRS/QT); and iii) Axis.
• They are usually reliable in recognizing sinus rhythm mechanisms and normal tracings.
• For the Expert Interpreter — the most advantageous feature of computerized systems is saving time! There is no longer need to calculate rate, intervals or axis — since the computer instantly provides legible and accurate printout of these values. IF the computer says, “Normal ECG” — it may literally take no more than 2-3 seconds for an experienced interpreter to overread and sign the report (provided they agree with the interpretation).
• For the Non-Expert Interpreter — the major benefit of computerized systems is the backup opinion the system provides. The computer may suggest findings not initially thought of by a less experienced interpreter. This encourages more careful, targeted review of the tracing. It may also be educational by the suggestions it makes. Finally — confidence may be boosted when computer analysis agrees with the clinician’s interpretation.
• NOTE: The computer backup opinion may help the expert-in-a-hurry from overlooking any ECG findings.

CAVEATS: Computerized systems do not do as well in evaluation of abnormal tracings as they do with normal or relatively normal tracings (especially when complex abnormalities are present ).

• They are far less accurate interpreting rhythms that do not have a sinus mechanism.
• They may miss subtle infarctions ...
• They tend to overinterpret the J-point ST elevation that is commonly seen with early repolarization patterns (at times mislabeling these normal variants as “acute MI”).
• They may miss pacemaker spikes / WPW / tall R in V1.

Suggested Approach: How to Use the Computer

Expertise of the interpreter dictates the approach:

• For the Non-Expert Interpreter — Do not initially read the computer report (because you do not want to bias your interpretation by what the computer says). Instead — WRITE OUT (or at least think out) your interpretation before you look at the computer readout. Check findings you note with each computer statement. Then delete, modify and/or add to the computer interpretation as needed. Admittedly — it will take a little longer to interpret ECGs in this manner — but given the limited number of tracings most non-expert interpreters must assess, time-saving is far less important than optimizing accuracy. Checking your interpretation with what the computer says may suggest findings not initially thought of — as well as forcing you to relook at any questionable findings. This increases accuracy.
• For the Expert Interpreter — Review the computer report either before or after evaluation of the ECG itself. Minimize time devoted to determination of heart rate, intervals and axis (since the computer is very accurate for these parameters). Consider more careful evaluation IF the rhythm is not sinus — or IF the ECG is interpreted by the computer as abnormal. Overread each computer statement. Place a check mark next to those that are accurate. Delete, modify or add to incorrect statements. Optimal speed is attained by quick overview of the tracing in conjunction with review of the computer report. Normal ECGs should literally take the expert no more than 2-3 seconds to complete. Awareness of computer strengths and weaknesses in interpretation allows the expert to quickly determine when the computer report should be partially or completely ignored.
• For any Interpreter — If you disagree with one or more computer statements — cross these out! The computer reading is part of the medical record — so incorrect statements should not stand unmarked.

------------------------------------------

Editorial Comment by the Author:  Prior to researching this topic (references of my work below) — I thought computerized interpretations were a waste of time. However, once I learned to appreciate their benefits and drawbacks — I learned to love computerized interpretations. They literally tripled my speed of interpretation — especially when sinus rhythm and minimal abnormalities are present. On the other hand — I ignore what the computer says when the rhythm is anything other than sinus. I’m fully aware of the need to carefully overread early repolarization patterns and tracings on patients with chest pain. Finally — I’ve observed that when less experienced interpreters truly make an honest attempt to interpret the ECG first before they look at what the computer says — that accuracy is increased.

-----------------------------------

 

=================================

PART 4: All About Comparison ECGs for 12-Leads and Arrhythmias

==================================

“All About Comparison ECGs for 12-Leads and Arrhythmias ... "

From what I’ve observed — potential benefits of comparison tracings are not optimally appreciated by all too many clinicians of all experience levels —

• — and even when clinicians do appreciate the benefits of comparison tracings — their technique for comparison is lacking.
• 
  
• NOTE: This applies both to use of comparison 12-lead ECGs — as well as — use of comparison tracings for arrhythmia detection.

===============================

I’m going to start with the use of comparison tracings for 12-Lead ECGs:

Benefits from Comparison of 12-Lead ECGs:

• I think of 2 “Types” of comparison tracings — i) With a prior tracing on the patient (done some time in the past); and, ii) With serial tracings (most commonly done when assessing CP patients for acute MI.
• 
  
• KEY Point for Both — WHAT are you comparing?  — By this I mean …
• IF using a prior tracing: How long ago was the prior tracing obtained?
• This is esp. important IF you are assessing a pt. with new CP!
• I’ve seen clinicians compare a “baseline” tracing from 5 years earlier with the initial ECG from a pt. with new CP — without any mention that since this “baseline” ECG — the pt. had another MI … — OR — without looking in the chart to see IF at the time of the “baseline” ECG the pt. was (or was not) having CP … 
• Now — it may not be easy to figure out from the chart what was going on at the time of the last ECG – but for the “prior” tracing you find to be useful — You gotta know —  What are you comparing?
• 
  
• KEY POINT:  — Write on the chart or the tracing — What was going on at the time !!!

Regarding Serial ECGs (as I mentioned in my Podcast Talk on OMIs) — 

• In a pt. you are evaluating for new CP — You want to know IF the “culprit” vessel is OPEN or CLOSED — so it is KEY to WRITE DOWN for each ECG the time and IF the pt. had CP (and if so — HOW SEVERE?).
• That’s because IF — CP is decreased (or gone) at the same time that the ECG shows the ST elevation is gone — this tells you that the “culprit” artery has opened.
• 
  
• But IF pt.’s CP returns at the same time that the repeat ECG shows ST elevation (and reciprocal ST dep.) is returning — there has been re-occlusion!
• 
  
• Clinically — This is critical info. I can’t tell you HOW MANY TIMES clinicians think “All is Well” with no need for cath because pt. no longer has CP and the ST elevation is gone.
• 
  
• Again — the Clinical Point — is that spontaneous reperfusion (that is, reopening) of the “culprit” artery may occur (without any treatment! ) 
•   — so if CP resolves, the patient feels better — and the ST elevation is gone — what spontaneously opened, may just as easily spontaneously reclose! — which is why the patient still needs to go to the cath lab! 

HOW OFTEN should you repeat ECGs in a patient with new CP?

• The ANSWER: As often as you need until you become certain of your diagnosis.
• 
  
• REMEMBER — Even IF the ECG does not show acute changes — IF your patient has persistent (unrelieved) CP and Troponin is elevated (even slightly) — that IS an indication for prompt cath.
• 
  
• Also REMEMBER that in an acutely evolving event — you can go from NO ST elevation — to marked ST elevation in minutes! 
•   — So IF you suspect an acutely evolving process — Repeat the ECG within 10-20 minutes!
• 
  
• Repeat the ECG whenever there is a clinical change (The patient's CP goes away — OR — the CP had gone away and now comes back!).

 

What about the Technique for comparing ECGs?

• The most common ERROR (by far! ) — Clinicians are in a rush — so they look at 1 ECG — and then they look at the next ECG — BUT they never go lead-by-lead when they compare tracings.
• 
  
• What do I MEAN by “lead-by-lead”? Simply — that I look at lead I on 1 tracing — and then I compare this to the QRST in lead I in the 2nd tracing.
• Then do the same for lead II — and for each of the 12-leads on an ECG …
• 
  
• So — PICK 1 ECG — either the prior ECG — or the current ECG.
• Interpret 1 ECG in its entirety (ie, Rate-Rhythm-Intervals-Axis-Hypertrophy-QRST Changes).
• THEN — Go lead-by-lead ...

What IF you don't go "lead-by-lead" when you compare ... ?

• Then you might look at ECG #1 — and see nothing more than nonspecific ST-T wave changes.
•   — and then you might look entirely at ECG #2 (and see no more than nonspecific ST-T wave changes) 
•  
•   — whereas IF you would have gone lead-by-lead — even though neither ECG by itself looks acute — When you go lead-by-lead — You might THEN see that there has been a significant change between the 2 ECGs — and this might indicate "dynamic" ST-T wave changes that confirm an acute event!

KEY Point: 

   — Be SURE you are comparing “Apples with Apples 

           — and not with Oranges” !!!!  To do this:

• Frontal plane AXIS similar? 
•    — R wave progression similar? 

========== 

A Word about Artifact !!! (as well as lead reversals or misplacement)

• These technical “misadventures” — are all-too-common (No time to delve into all of them … )
• In my experience — These often go unrecognized ...

The clinician sees something that looks unusual

• As a general rule — IF the ECG looks “funny” (ie, Odd, unusual shapes — for example, the QRS should never be all neg in lead I — or if you have a flat line in 1 of the standard leads) — Verify lead placement — and — REPEAT the ECG.
• — So here … — the “comparison” tracing — is the repeat ECG after you’ve verified lead placement and have fixed the leads … and you’re looking to see if this resolves the unusual findings ….
• 
  
• For some reason — many clinicians don't want to repeat the ECG ...
• So — You are evaluating a patient with new CP for possible a possible acute OMI. IF there is any doubt about a technical misadventure — then CHECK the leads and Repeat the ECG!
• 
  
• I’ll just mention 1 type of artifact we seem to be seeing more of (probably because many clinicians never realized what they were seeing before) — which is PTA (Pulse-Tap Artifact) — when one of the extremity electrodes makes contact with a pulsating artery — and therefore produces a large geometric deflection that is related to the QRS (because it’s caused by a pulsating artery underneath and in contact with this recording electrode).
• So — because by definition — PTA is the result of contact with a pulsating artery — the artifact will have a fixed relation to the QRS (usually occurring a short distance after the QRS) — so as you might imagine, this can simulate ST elevation or depression except that the elevation or depression may be quite large and round.
• 
  
• HOW to Recognize PTA (in addition to the unusual geometric shape) — There will be a mathematical relationship between the various limb lead based on Einthoven’s Triangle. 
• 2 of the Standard Leads ( = I,II,III ) — will show equal amplitude artifact — and the 3rd Standard lead will look relatively normal! It is this 3rd normal-looking lead that really clues you in!
• In the Augmented leads — the lead in which the artifact is greatest tells you which is the “culprit” extremity.
• For example — IF the bizarre deflection appears to be of equal amplitude in leads II and III — but lead I looks very normal — then I’m thinking the source of the artifact is the LL (Left Leg) — because electrical activity in lead I (which looks normal) is not dependent on the LL recording. And then IF maximal artifact is seen in lead aVF — this confirms that the LL is the source of the artifact ( = the "culprit" extremity ...).

====================================

Use of Comparison Tracings for Arrhythmias:

====================================

• Know what you are looking for!
• It’s a lot easier to find what you are looking for — IF you know in advance what it is you are looking for!

Let’s assume that for your single or double lead rhythm strip — You’ve already gotten a 12-lead!

• Advantages of a 12-lead for tachycardias is that you have 11 more chances to find atrial activity.
• BEST leads for seeing atrial activity = lead II (upright P if sinus) — III,aVF — aVR — V1
• Confirms IF the QRS is wide or narrow !
• 
  
• Is atrial activity retrograde? (IF so — P waves should be neg in inferior leads! — often pos in aVR,V1 … )
• 
  
• CAUTION: 1:1 VA ( = retrograde) conduction does not distinguish VT from SVT !!!

Say you have a regular SVT rhythm at a rate ~140-to-160/minute (so that the rhythm could be Sinus Tach — AFlutter — ATach — Reentry SVT = AVNRT or orthodromic AVRT) — then identifying presence or absence of atrial activity is KEY!

• It’s sometimes hard to identify neg notching of retrograde conduction (Look to see if the notch is GONE after conversion back to sinus rhythm … ).
• 
  
• So — Especially IF the QRS is narrow ( = an “SVT” ) — IF retrograde P waves — What is the RP’ interval ???
• 
  
•   — IF a short RP’ (seen at end of QRS) = AVNRT
•   — IF a moderate RP’ (seen from mid-ST-to-T ) = orthodromic AVRT likely (There’s an AP! ).
•   — IF a long RP’ (seen close to the next P wave) = AVNRT of the “short-long” type (similar Rx — but sometimes more resistent to Rx).
• 
  
• Are flutter waves hidden? Comparing to the ST-T wave during sinus rhythm may tell you if P waves were hidden during the SVT rhythm …
• 
  
• Is there WPW on the post-conversion tracing? (as may occur if you had orthodromic AVRT ..).

Say you have a regular WCT …

• Sometimes your BEST hint a to whether this was VT vs an SVT — only comes after converting to sinus rhythm.
• 
  
• Is the post-conversion 12-lead normal? (as should be expected for an idiopathic VT — in which the patient does not have underlying heart disease).
• 
  
• Is there suggestion on post-conversion ECG of coronary disease? (either “old” or “acute or recent MI”).
•   — Perhaps the post-conversion 12-lead tells you WHY there was VT ??? (ie, the post-conversion tracing shows an acute STEMI ...).
• 
  
•   — Are there post-conversion PVCs with identical QRS morphology as during the regular WCT ??? (could prove that the reg WCT was indeed VT !!! ).
• 
  
• Is there a prior ECG available ??? (Can prove SVT if identical prior QRS morphology during sinus rhythm … !!! ).

==============================================

Failure to Recognize ARTIFACT!

• The reason we often fail to recognize these — is that they are not real common — but they definitely DO occur!
• Lead Reversal DOES Occur
• The most common = LA-RA Reversal (Global negativity in lead I — I looks like R — P neg in II )But the 2nd-most common that I’ve seen which is very easy to overlook = LA-LL Reversal (P in I > II )
• 
  
• Pulse Tap Artifact (often bizarre QRS or ST-T wave — artifact related to the QRST, because it is from electrode contact with a pulsating artery)
• How to tell WHICH Extremity? — Based on Einthoven’s Triangle — Max artifact in 2 limb leads but nothing in the 3rd limb lead — and Max in 1 Augmented lead ( ~ 1/2 max in the other 2 Aug leads!)
• THINK Artifact — IF the funny-looking deflections look geometric (too straight) and esp. IF underneath the artifact you can see a normal rhythm.

==============================================