Nimish Shah, MD

Excited to be able to share this work from @Parkland @PCCInnovation with the support of @sandeepdasmd @TimothyJBrownMD @dhjutsw1 @UTSWIMchief @UTSWInternalMed!! sciencedirect.com/science/articl… A simple EHR change reduced expensive inpatient IV Mag supplementation

2026 Dietary Guidance to Improve Cardiovascular Health ahajrnls.org/4v0o6gL

...to enable clinicians to be more efficient not just with charting, but with thinking and acting. But EHR-based multi-agentic AI shouldn't be locked down to admins or IT. Clinicians should be able to impart their own clinical expertise to drive their own agents.

This. AI in healthcare is not just AI scribing or automatic interpretation of studies (EKGs, imaging etc). It's also not just for large health system research or big data. The practical power comes from automating all of these little clinical steps and cognitive links.

@PraveenRangana9 Thank you! That's what I was thinking to do. Started with Copilot for the framework and thought to move to Claude but I was anticipating the DICOM transfer issue. Need to see what can be built within institutional data ecosystem.

@NNShahMD Had Gemini build the architecture for the eventual prompt to Claude for an integrated DICOM viewer and reporting solution but never ended up executing. Issue is more secure DICOM transfer to your viewer/app, less building a functional app

Has anyone out there used AI vibe coding to redesign their medical image interpretation and reporting software? Specifically interested in echo interpretation. Curious to know how one would go about this within the constraints of institutional IT.

Can AI read an ECG like a cardiologist - from just an image? We built ECG-GPT, a vision-text transformer that generates complete diagnostic reports directly from photos of 12-lead ECGs Now out in @ESC_Journals #EHJDigitalHealth Kudos to @aakhunte & @Veer_Sangha_ for leading this @cards_lab 🧵

Check out our study in @JAHA_AHA on short and long term outcomes of pericarditis, myocarditis and myopericarditis in patients on ICI therapy! Couldn’t have done without @AkivaRosenzveig Thank you, Dr. @AllanLKleinMD1 for the immense support! ahajournals.org/doi/10.1161/JA…

Been saying this since residency

Magnesium supplementation showed no significant impact on 24-hour risk of tachyarrhythmias, hypotension or death in critically ill patients with hypomagnesemia with serum values close to hospital thresholds. bit.ly/4rGieab

🫀💊 Lipid management after statins: more tools, more complexity, more responsibility This 2025 JACC Advances expert panel offers a timely, pragmatic appraisal of lipid management in the post-statin era, comparing international guidelines and translating them into real-world, case-based decisions . 📌 Core message LDL-C remains causal for ASCVD, but the therapeutic landscape has expanded dramatically. Statins are still the foundation, yet non-statin therapies are now essential to reach modern LDL-C targets—especially in high- and very-high-risk patients. 🌍 Guidelines: same evidence, different philosophies ESC guidelines are the most aggressive (LDL-C <55 mg/dL, and <40 mg/dL after recurrent events). ACC/AHA remain more conservative, prioritizing randomized trial evidence. Canadian guidelines sit in between.The result? Different thresholds, same patients, increasing clinical ambiguity. 🧠 Key advances highlighted Combination therapy beats dose escalation: doubling statins yields ~6% LDL-C reduction, while adding ezetimibe, PCSK9 inhibitors, bempedoic acid or inclisiran achieves 20–60%. Therapeutic inertia is a major problem—many high-risk patients remain far above target. Imaging matters: CAC ≥300 AU confers a risk similar to secondary prevention and should trigger aggressive LDL-C lowering. Lp(a) and ApoB deserve broader screening, even if outcome-driven targets are still evolving. 👩‍⚕️ Why the case-based approach matters Real patients rarely fit guideline boxes: young adults, women planning pregnancy, South Asian populations, HIV patients, severe hypertriglyceridemia. The paper shows how rigid guideline application fails without clinical judgment. 🔮 Bottom line We are no longer in a “statin-only” world. Modern lipid management requires earlier risk detection, imaging-guided stratification, combination therapy, and individualized targets. Guidelines inform decisions—but expertise personalizes them 🚀

Honored to have our Cardiology program featured in Cath Lab Digest: rebuilding our labs around advanced imaging, physiology, structural heart, and research so patients in western New England get world-class PCI/structural care close to home. hmpgloballearningnetwork.com/site/cathlab/i… #CardioTwitter

A classic RAC sign from 📢TTE (📷&🔗) vs “novel”🧲 CMR-based RAC (tweet) Which one have you seen more frequently? jacc.org/doi/10.1016/j.…

Thrilled by the success of our 12th Annual Western New England Acute Cardiology Conference at MGM Springfield — over 250 in attendance and incredible talks on acute CV care. Huge thanks to our speakers, attendees & sponsors for an amazing experience @ACCinTouch @MGMSpringfield

New multi-society Advanced Training Statement on Advanced #cvImaging addresses the core competencies & training requirements necessary for advanced CV imagers across all 4 imaging modalities – echocardiography, CCT, nuclear cardiology & #CMR. jacc.org/doi/10.1016/j.… #JACC

📰 HOT OFF THE PRESSES 📰 Immune Checkpoint Inhibitor (#ICIs) treatments are known to potentially cause harm to the heart, particularly through conditions like #myocarditis. Today in @JAMAOnc, we stressed the importance of interdisciplinary collaboration as the best way to mitigate effects and protect the #hearts of #cancer patients. Read the positioning statement: jamanetwork.com/journals/jamao…

2025 ESC/EACTS Guidelines for the Management of Valvular Heart Disease Key Change: 🔹In patients with low-flow, low-gradient (LFLG) AS with reduced EF, the 2025 guidelines now place CT aortic valve calcium scoring (AVCS) on equal footing with dobutamine stress echocardiography (DSE) for confirming severity. 🔹AVCS thresholds remain the same: >1200 AU in women and >2000 AU in men. 🔹If CT calcium score is above these cutoffs → this is sufficient for diagnosis of true severe AS, without needing DSE. 🔹DSE is no longer mandatory — it’s optional and equivalent to CT in decision-making. If one test is not conclusive, complement diagnostics with the other test.

🔥of the press! @ASE360 Guidelines for Standardization of Adult #EchoFirst Reporting are out 🙏Grateful to have worked with this amazing group led by @CynthiaTaub & @ray_stainback 📣Don’t miss a great session on this important📄this weekend at #ASE2025! 👉onlinejase.com/article/S0894-…

It’s true that current food systems undermine health & contribute to chronic disease. But the evidence-based solution isn’t beef tallow or raw milk - it’s whole-food plant-forward nutrition. @JACCJournals @ACCinTouch @YaleMed ⁦@YaleCardiology⁩ sciencedirect.com/science/articl…

I passed the Lipid Boards earlier this year. Here are some take-away from the major guidelines that you can use in your practice. Part 1: 10 Take-Aways #arjuncardiology #cardiotwitter

A new meta-analysis confirms that CCTA is more effective than TEE at detecting residual leaks after LAAO. With higher detection rates, fewer risks, and less operator dependence, CCTA is emerging as the preferred post-LAAO imaging tool—supporting 2023 SCAI/HRS recommendations.