Accurate & dynamic estimates of #lifetime #ASCVD risk may enable earlier and more effective #cvPrev Now in Nature Communications, innovative work led by Mass General Heart 💫 fellow Sarah Urbut to develop and validate MSGene: a multistate lifetime risk model for #ASCVD! Broad Institute

Awesome 🧵by Sarah Urbut on her recent work developing MSGene - a novel multistate model incorporating dynamic transitions and treatment effects to estimate lifetime risk of #ASCVD ! #MedTwitter #CardioTwitter #EpiTwitter #cvPrev Paper Nature Communications ➡️ nature.com/articles/s4146…

Very excited to have this paper out in the world! We use "filtered corpus training" to show that LMs (transformer and LSTM, controlled for size, corpus, etc) can generalize from indirect evidence. Many interesting results! Great thread below, paper here: arxiv.org/abs/2405.15750

Interesting work by Gavidia et al. to develop WARN, a method of detecting #afib episodes ~30min in advance using 24-hr Holter #ECG monitors, now in Patterns journal Paper: cell.com/patterns/fullt… Need to identify ways to leverage early warnings... Sam Freesun Friedman and my thoughts ⬇️⬇️

To all my #EPeeps out there! May your PVs stay forever isolated… Image credit to the multitalented Tri N.

Great writeup on a nice study led by Lior Jankelson JACC Journals #Imaging on performance of GPT #LLM to create patient-facing summaries of clinical #echofirst reports. Encouraging accuracy but human review still needed in ~25%. jacc.org/doi/10.1016/j.…

I was today years old when I realized "Staples" (store) is a pun

Congrats to Sarnoff Foundation fellow and StanfordMedRes Amey Vrudhula for publishing in Circulation! Using deep learning, we view classify and quantify MR severity in #echofirst. Currently the only such project with open code and model weights. Paper: ahajournals.org/doi/10.1161/CI… Code:

Exceptional new work by James 🥥🏈 Pirruccello in European Society of Cardiology Journals #EHJ on AORTA Gene: a novel clinical and genetic model for aortic size AORTA Gene validates in 3 datasets using both #whyCMR & #echofirst, and ⬆️ scores assoc w/ ⬆️ risk of adverse aortic outcomes bit.ly/3yt6jXl

Innovative new work led by dynamo Shino Kany to develop Predict-AF3, a score leveraging clinical factors (CHARGE-AF) + genetics (#PRS) + #ECG-based #AI to estimate #afib risk among in ~50k people in the independent UK Biobank cohort Broad Institute ML4H Mass General Heart

Great presentation by Shino Kany on using an AI model to detect CAD from the 12 lead ECG at #ESCCongress Akl Fahed ML4H Mass General Heart

Simple question: Does ⬆️ LDL-C cause aortic stenosis (#AS)? NEW cutting edge work Circulation intersecting #WhyCMR / #AI & rare variant #genomics says YES Congrats Joel Rämö Sean Jurgens Shino Kany Patrick Ellinor James 🥥🏈 Pirruccello ! Broad Institute Mass General Heart UCSF Health