Lior Jankelson (@ExtraStim) header
Lior Jankelson (@ExtraStim) profile picture

Cardiac Electrophysiologist| Director, Inherited Arrhythmia Program & AI/ML in Cardiology Group| Associate Professor of Medicine & Biomedical Engineering, NYU

New York City
Joined June 2013
Sharing here the #LAAO #LAAC simulator laa.liorjankelson.com. See the left atrial appendage with 3D CT correlation, explore LAAOD positioning, play with device alignment, see "shoulders" and leaks in each view and more. #eppeeps #afib
2
43
1
225
11,921
🔥🔥🔥let's go!
84
Replying to @drjohnm
To put your perspective in perspective, you are basing your conservativism argument on ONE small trial questioning QOL benefit of ablation - yet clearly demonstrating freedom from arrhythmia benefit - while we have multiple trials showing the opposite: with sham (see figure), AND with a mortality (!) benefit with early intervention (AFNET, CASTLE). It's good to always question, as long as you are honest about the answers.
1
9
786
If you’re an AI engineer and you care about doing something meaningful, please let me know. Join this 🚀
2
491
Huge interest in our work on opportunistic screening with #AI for Atrial Fibrillation from NON CONTRAST CALCIUM SCORE CTs and EKG presented by my fellow @jmjaslove Do more with what we already have! @nyulangone @nyugrossman @nyutandon
6
249
2/ As we ((nature.com/articles/s41598-0…) and others have shown, the prevalence of type 1 Brugada pattern is highly underestimated, and accordingly the risk associated with isolated type 1 is likely overestimated.
1
80
Replying to @jeffrey_vinocur
1/ The key and fundamental issue is the overly permissive definition of "spontaneous type 1 pattern" as BrS. In the 2013 consensus statement (Priori), asymptomatic patients with spontaneous type 1 pattern required additional supportive features:
1
1
53
Eric, this is an interesting study. However, freedom from atrial fibrillation is not an acceptable end point. Freedom from ANY atrial arrhythmia is.
10
247
Stellar presentation by @AviramHochstadt on our #AI tools for #genetics in #cardiology 💪💫 varcard.io #AHA24
2
1
16
529
Excited to present our work on AlphaFold/AlphaMissesne propelled variant annotation via Varcard.io! #AHA24 @nyugrossman @nyulangone @nyuniversity @nyutandon eppro01.ativ.me/appinfo.php?…
1
8
430
Introducing GuidelineGenetics. Leveraging recent breakthroughs in #AI, we developed an end-to-end system capable of intaking a note (with or without genetics) providing you with a detailed, *guideline directed*, decision support output. Try it at guidelinegenetics.com
6
344
Online in @JACCJournals CV Imaging: Our work on training/utilizing an LLM to simplify Echo reports for the benefits of patients. TL;DR: #AI generated reports are accurate. And, patients love it! Patient-Oriented Echocardiogram Reports Augmented by AI sciencedirect.com/science/ar…
1
6
19
1,877
Replying to @daardos
יש אסוציאציה בין אנשים שמשרתים ביחידות מובחרות להצלחה. זה לא אומר ש*בגלל* שהם שרתו ביחידות כאלה ואחרות הם מוצלחים.
1
54
⚡️Just out @JACCJournals⚡️ Prediction of 💊 induced QT prolongation with #AI 💥QTNet - 🆕 architecture for fusing ECG with EHR data ⏱️-dependent prediction from day 0- 180 & external validation 💥day 2 AUC > 0.9 sciencedirect.com/science/ar… 💫@neiljethani 💫@DrCardioCon 💫HaoZhang
1
8
1,006
Bifid appendage closed with 2X Amulet 🤠🤠
2
8
652
Replying to @tsoofbaror
כתיבה יפה. עיקר הערך בLLM נמצא באימון, ביצירת ה embeddings. זה בעצם ה world model וזו גם ה-מגבלה של LLMs: טקסט הוא מקור לא מספק ליצירת מודל עמוק. רוב ה״ידע״ או ״אינטואיציה״ (או אמבדינגז) באינטיליגנציה אנושית נוצרים בכלל ממקור ויזואלי. כל השאר (אינפירנס) זו בעיה מסדר שני. דוגמה:
37
📣New from varcard.io 🎉 We are now providing you with the @ClinGenResource curated gene-disease correlation(s). For any gene variant searched, lower bar will provide all known clinical correlation graded. #cardiotwitter #CancerResearch #Genetics #geneticmedicine
1
4
8
1,039