Discussion about this post

User's avatar
The Skeptical Cardiologist's avatar

Titus,

I am not optimistic ML or AI will advance the field of echo interpretation. Attempts to automate the process of measuring the key parameter of ejection fraction using computer assistance have been going on for decades... and all have failed. Currently, there is a lot of buzz around measurement of strain as a better parameter for pump function of the heart. Clinical applicability is limited due to failure to visualize endocardium of the LV, something that is patient , machine and technician dependent and many other factors. A company Ultromics has developed a cloud-based AI platform which alleges superior interpretation . The studies are weak, biased and unlikely to be reproduced.

Garbage in creates garbage out and echo is just too operator dependent to be useful IMHO.

Dr P

Expand full comment
The Skeptical Cardiologist's avatar

JD,

Cardiologists utilize beta-blockers somewhat interchangeably for the indications I mentioned. We seek the most effective and well-tolerated type and dosage.

One area where they are not interchangeable is heart failure/cardiomyopathy/post MI where I utilize carvedilol or metoprolol succinate almost exclusively.

There are slight variations in beta-selectivity and alpha-selectivity among the various beta-blockers which play a roll in efficacy and tolerability.

Dr. P

Expand full comment
13 more comments...

No posts