Home Creators Posts Import Register Favorites Logout
haven't archived this post yet. have a subscription? use the importer!

Content

This is astonishing. The most highly trained profession is now surpassed not only by the combination of cardiologists with AI but also by AI alone.

Remarkably, an electrocardiogram can detect previously undiagnosed structural heart disease using AI by itself, outperforming cardiologists even with AI assistance.

Files

Previews only

Comments

John L

I insisted my cardiologist do a CT of my heart w AI interpretation (Heartflow). The doc begrudgingly did saying I didn’t need it , and it showed that my atrial appendage was really large and convoluted and I’m is an extremely high risk category for stroke and I need treatment : insertion of a mesh into the opening of the appendage to prevent clots from disrupting or blood thinners the rest of my life. I’m going to have to find another cardiologist who is more preventative orientated.

Norse Gael

Is this cardiology including surgery as a whole or just electrocardiograms???

MVesting

James, what is the source of this data? I am a physician, but have no skin in the game about cardiology and job security.

Miran

John I’m a doctor and your doctor is right . You don’t close an appendage just because you have one. It’s a risky procedure . The indications for closure are having documented atrial fibrillation + LA appendage + high risk stroke + contraindication to blood thinners . Also you need a trans esophageal echo to confirm the appendage before anything else . My recommendation is wear a heart monitor first to see if you have atrial fibrillation . BTC 120k btw

Horsegal33

how do we invest in it??

CryptoToRetire

Most patients have had an ECG prior to seeing a Cardiologist - detecting abnormalities from that leads to a referral in the UK. If patients can actually get an accurate ECG interpretation from AI rather than misdiagnosis by incompetent GPs it may lead to more appropriate referrals to Cardiologists. We have used algorithms within ECG interpretation for years - as with a lot of flash headlines, it is the question of how this improves clinical outcomes rather than who does it better. As above, an ECG does not define a ‘Cardiologist’. I just hope I have enough TSLA before AI can do the echos I report ☺️

Aegis87

This is a good question. The picks and shovels component is straight forward. Picking a winning individual company is not.

Crypto-Cephalon

I'm not saying this isn't true but as a scientist this figure makes me suspicious. There need to be error bars on the green bars to know whether they're statistically different or not, despite what it may look like. If there is a huge amount of variability in the AI alone performance the black error bars would be large and in fact, on average the difference might not be as meaningful as it seems. Nonetheless, the first and third measures are clear benefits of adding AI to the clinician's toolbox. The middle one is likely not better than chance (statistically).

GloMa

Exercise & avoid doctors.

Dr.Bones

Did an AI generate this just to make us stop trying? I give up

Raddoc

Doc here… while in the long-term I think it’s possible we could have Star Trek like AI, we are a long way away from this. AI is making rapid progress, but if you’re an expert in your field, you understand that it’s really having trouble closing the gap on that last 20% of expertise. Specifically, the areas that involves a nuanced combination of skill, experience, and knowledge. AI does not excel at this, and I think it’s gonna be a long time before it completely closes the gap. Like many things, there is rapid progress with the curve flattening as it gets closer to being perfect. Most likely we are looking at physicians being augmented by AI for quite a while. All in all, I think it’s a win-win for doctors and patients.