Mintu Turakhia, MD MS

5.9K posts

Mintu Turakhia, MD MS banner
Mintu Turakhia, MD MS

Mintu Turakhia, MD MS

@leftbundle

CMO/CSO/EVP Adv Tech/Product @irhythmtech. Professor and founding director @stanfordcdh. Cardiac EP, scientist, trialist, AI builder. Views mine.

San Francisco, CA เข้าร่วม Mayıs 2013
1.4K กำลังติดตาม6.5K ผู้ติดตาม
Edward J Schloss MD
Edward J Schloss MD@EJSMD·
@SergioPinski @drjohnm ICE makes more sense to me than 3D map. Suspect one main issue is lab dependence on mappers to do any ablation. Before I gave it up, I watched ablation costs skyrocket due to unnecessary mapping and regulatory barriers to catheter reuse. All this from industry influence.
English
2
2
9
1.5K
Mintu Turakhia, MD MS
Mintu Turakhia, MD MS@leftbundle·
Delays in Orlando due to storms. Interestingly, Dubai still listed as the only flight on time.
Mintu Turakhia, MD MS tweet media
English
0
0
1
484
Mintu Turakhia, MD MS
Mintu Turakhia, MD MS@leftbundle·
You can’t generalize whether you should categorically not switch patients from warfarin to DOACs from these studies. Here are the the most important factors: 1. Time in INR therapeutic range (TTR) and time above INR range. This remains the gold standard process measure linked to outcomes and why specialized warfarin clinics save lives. I don’t see TTR used here. 2. DOAC dose-patient mismatch. GFR is incredibly important, and some DOACs with high renal elimination also have the longest half life and are twice daily (dabigatran). CKD is associated with worse time in INR range. Going from well-managed warfarin to DOAC overdosing is a problem. This can happen when GFR is close but has not hit a threshold for dose adjustment. Going from poorly-managed warfarin to appropriate DOAC dosing is rational based on the totality of safety evidence. 3. Concomitant aspirin. Should be stopped unless there is a very strong case not to. TL;DR - when dealing with edge cases, use pharmacology, math, and judgement to tailor to your patient rather than a measure of central tendency of observational studies of highly-selected patients that may not generalize to yours.
John Mandrola, MD@drjohnm

Super interesting. This aligns well with FRAIL AF which found > bleeding from switching But, neither these authors nor the editorialists, cite Nicolau et al [JACC 2025] COMBINE AF substudy, which was an elegant substudy of the 4 DOAC v warfarin trials, which found no net diff in frail Vit K naive pts randomized to DOAC. COMBINE was a substudy but an elegant one in that it preserved randomization What I worry about w the below obs study is confounding because maybe the switching group was inherently more ill Will cover on TWIC Friday. See my Aug 8 2025 coverage of COMBINE AF medscape.com/viewarticle/10…

English
2
4
14
12.8K
Mintu Turakhia, MD MS
Mintu Turakhia, MD MS@leftbundle·
@MKIttlesonMD IMO, it’s less about being a direct endorsement and more about validating the narrative that the candidate offers in their own candidate statement. Promotion is harder when the candidate’s self-view differs from those in field.
English
0
0
3
741
Michelle Kittleson MD PhD
Michelle Kittleson MD PhD@MKIttlesonMD·
So, if basically no one ever writes a negative letter for academic promotions, doesn't that make the recommendation letter component of the process performative and a waste of the letter writers' time?
Michelle Kittleson MD PhD@MKIttlesonMD

I don't sit on any academic promotions committees, and sometimes I wonder if the process is somewhat performative. So I have to ask: when you agree to write a letter of support for a colleague's promotion, have you ever written a negative one that argues against promotion?

English
12
4
44
28.1K
Harriette Van Spall, MD MPH 🇨🇦
Harriette Van Spall, MD MPH 🇨🇦@hvanspall·
He was a mentor But departing from evidence, he had my dad w new angina wait 4wks for CABG: Cath/IVUS had shown 70% LM, tight ostial LAD+LCx lesions, RCA occlusion My beloved dad died before CABG The ensuing silence left me w #lessons I carry everyday1/ acpjournals.org/doi/abs/10.732…
Harriette Van Spall, MD MPH 🇨🇦 tweet media
Subodh Verma@SubodhVermaMD

So much fun, interviewing, Dr. David Latter on his incredible journey over four decades as a heart surgeon, lessons learned and words of wisdom for the future generation. The full video will be posted soon.

English
43
53
283
96.4K
Mintu Turakhia, MD MS
Mintu Turakhia, MD MS@leftbundle·
@hvanspall Thanks for sharing and resurfacing. I’m sorry for your loss and for having to live with the “what ifs”. It’s brave of you to share and important that people see it. ❤️‍🩹
English
1
0
1
213
Mintu Turakhia, MD MS
Mintu Turakhia, MD MS@leftbundle·
Privileged to deliver the Dr. Louis F. Bishop Keynote at #ACC26 and share the stage with some absolute giants in the field. Thank you, @AmericanColl.
American College of Cardiology@ACCinTouch

#ACC26 brings together visionary leaders whose ideas, research & innovation are redefining what's possible in CV care. This year's Keynote lineup spans the cutting edge of #AI, the evolution of #ACHD, the power of team-based care, & more. Don't miss! ➡️ bit.ly/3NhyBuZ

English
0
0
3
520
Mintu Turakhia, MD MS รีทวีตแล้ว
Dr Anastasia Mihailidou FAHA FCSANZ FESC
Dr Anastasia Mihailidou FAHA FCSANZ FESC@AnastasiaSMihai·
Congratulations @rohan_khera @CathieBiga @leftbundle & everyone👏👏👏👏👏 Looking fwd to your @ACCinTouch #ACC26 Keynotes & vision!
American College of Cardiology@ACCinTouch

#ACC26 brings together visionary leaders whose ideas, research & innovation are redefining what's possible in CV care. This year's Keynote lineup spans the cutting edge of #AI, the evolution of #ACHD, the power of team-based care, & more. Don't miss! ➡️ bit.ly/3NhyBuZ

English
1
2
8
573
Mintu Turakhia, MD MS
Mintu Turakhia, MD MS@leftbundle·
@nmarrouche I was registered and with tickets and hotel booked — but had to change plans for Family First. Big weekend for both kids. I’m so sorry to miss. Pics look amazing, and team iRhythm will still support and represent!
English
0
0
3
504
JK Han MD
JK Han MD@netta_doc·
💥Starting by off his talk with a bang! “WE SHOULD JETTISON THE BLANKING PERIOD” No other trials have a blanking period! We wear 2 hats: clinician & scientist- We can use judgement, *and* we should count all recurrences. #WAFib2025
JK Han MD tweet mediaJK Han MD tweet media
English
3
5
21
4.1K
Ross Prager
Ross Prager@ross_prager·
Here's a more full explanation. AUC looks at all cutoffs, not just one. Example test scores: Diseased patients: 71, 72, 73, 74, 75 Non-diseased patients: 10, 20, 30, 40, 50 Every diseased patient has a higher score than every non-diseased patient = perfect separation, so AUC = 1.0. Now pick a single cutoff of 74: Sensitivity = 3/5 = 60% (you miss the two lowest diseased scores) Specificity = 5/5 = 100% The ROC curve does reach the top-left corner (100% sensitivity and specificity) at a different cutoff (e.g., 71), which is why the AUC is 1.0. Sensitivity = 0.70 just reflects where you chose to operate on that curve, not the shape of the curve itself.
English
3
0
0
838
Venk Murthy MD PhD
Venk Murthy MD PhD@venkmurthy·
I've thought about these measures for decades and I don't understand how this can be true AUC of exactly 1.0 means sensitivity and specificity both must also be 1.0 exactly
Ross Prager@ross_prager

How can you have a test with an AUC of 1.0 yet a sensitivity of 0.70? AUC describes how well a test separates positives from negatives overall (ranking across all possible cutoffs). Sensitivity describes how many positives the test detects at ONE chosen cutoff. A test can separate cases perfectly (AUC = 1.0) but still have sensitivity = 0.70 if the cutoff is set conservatively. Want to learn more about diagnostic tests? Check out these calculators 👇

English
5
1
15
11.6K
Mintu Turakhia, MD MS รีทวีตแล้ว
Bo Wang
Bo Wang@BoWang87·
After leading AI & Health at @UHN — Canada’s largest hospital network — for the past few years, this is the first time it genuinely feels different. A cardiologist just placed 3rd out of 13,000 at Anthropic’s hackathon. He built it between hospital shifts. In a week. An AI agent that follows patients home — reverse scribe, full history, devices, evidence — all in one place. A few years ago this would’ve taken a full team and months. Now? One doctor. That’s the shift.
English
16
56
339
48.5K
Mintu Turakhia, MD MS รีทวีตแล้ว
Caleb Watney
Caleb Watney@calebwatney·
25-40% of clinical trial costs come from an excessive quality-check process that the FDA itself has recommended against for over a decade. Great piece from a former FDA official on why the whole system is stuck in a "too big to fail" loop. learninghealthadam.substack.com/p/why-clinical…
Caleb Watney tweet mediaCaleb Watney tweet media
English
12
94
510
50.1K
Chris Treadaway
Chris Treadaway@ctreada·
@cljack I've been building this watching parents struggle w/ the most stressful time of the day -- 6am! I would love your feedback! parentpoint.ai
English
1
0
1
477
Charlotte Lee
Charlotte Lee@cljack·
I'm trying to train Claude to read the weekly emails from my kids school and reliably summarize them and print a list of action items. It is losing its damn mind and rapidly spiraling into madness. I feel vindicated
English
142
289
8K
401.4K
Mintu Turakhia, MD MS
Mintu Turakhia, MD MS@leftbundle·
@cljack The solution is to for the schools (not the parents) to use AI to make their emails clear and concise. Or move to slack.
English
0
0
1
224
Mintu Turakhia, MD MS
Mintu Turakhia, MD MS@leftbundle·
Will central illustrations in journals get better or worse?
Akshay 🚀@akshay_pachaar

Google just dropped another banger! the figures in this paper were drawn by the system described in the paper. PaperBanana is an agentic framework that generates publication-ready academic illustrations from methodology descriptions. no manual design, no Figma, just your method section and a caption. here's how it works: five specialized agents collaborate in sequence: > Retriever: finds relevant reference diagrams from a curated set of NeurIPS papers. matches by visual structure, not topic. > Planner: translates your methodology text into a detailed visual description using in-context learning. > Stylist: applies aesthetic guidelines (color palettes, typography, layout) auto-summarized from hundreds of top-tier papers. > Visualizer + Critic loop: generates the image, critiques it against source text, and refines. repeats for 3 rounds. one surprising finding: randomly selected examples work nearly as well as semantically matched ones. what matters is showing the model what good diagrams look like, not finding the topically perfect reference. in blind evaluations, humans preferred PaperBanana outputs nearly 3 out of 4 times. it also extends to statistical plots using code-based generation for numerical precision. link in the next tweet.

English
0
0
1
688

ค้นพบ

@Kpakanati3 @ACCinTouch @EJSMD @ftrae @SergioPinski @drjohnm @MKIttlesonMD @hvanspall