Amazing progress in treating Huntington’s disease https://youtu.be/g66x9funPc0?si=qBcoW_oOD5UZ_x0q

3 replies

0 recast

12 reactions

Read a great paper on energy resistance in the body by Martin Picard. Essentially this idea of how inflammatory states or metabolic disease impact flow of electrons (such as in the mitochondria) and cause high resistance states. Its quite complex and so much of it was over my head so had NotebookLM made an explainer I found useful. Sharing if anyone finds it useful. Also a recent podcast Picard did: https://www.youtube.com/watch?v=GiwDfsIgziA Link to paper: https://www.sciencedirect.com/science/article/abs/pii/S1550413125003870?via=ihub (reach out if you want a PDF please)

1 reply

0 recast

5 reactions

One of the byproducts (or may have been intended) of this whole ‘longevity’ trend is that it has created a sense of optionality that those who are sick chose to be such. I had a patient last night who was normal weight, ex-college football player, and a family sports history. He has heart failure and had already had a bypass by 49. His 3 brothers, older and younger had already died from heart attacks and failure. And so had his uncles. Came in because now his rhythms were going whack. His genetics are trash, yet you’d be surprised at how much he will hear ‘if you had done X’…

1 reply

0 recast

2 reactions

Tahoe dropped some cool stuff. An open source 3B-parameter (YES 3B!!), foundation model trained on 100M single cells (which is an amazing dataset already) across 50 cancer models and 1,100 drug perturbations. Tx1 learns unified representations of genes, cells, and drugs, then delivering state-of-the-art performance on gene essentiality (DepMap) and hallmark oncogenic programs (MSigDB), with strong zero-shot generalization for drug response across unseen cell types which means finding response and perturbation to cells before medication is even used and new regimens. Surprised this hasnt received more coverage. Can see this expanding beyond cancer response into other specialties. Also opens the door to n of 1 treatments. https://www.tahoebio.ai/news/tahoe-x1-blog

0 reply

0 recast

4 reactions

Something doctors must learn to reconcile with as our roles change with technology, public opinion and simply as we go through different stages of our practice. Medicine can be practiced in many ways take multiple forms not just writing prescriptions or with a scalpel. We are skilled healers and listeners refined through years of practice. Nearly every aspect of life has a need for that skill. Wajahat articulates eloquently here. https://substack.com/home/post/p-172448030

0 reply

0 recast

0 reaction

PSA that if you live in the Western US, esp in the Southwest and California, and have any odd weakness, fatigue with respiratory symptoms, chances are it may be silent Valley Fever more than even long Covid. Always good to get tested for this as early detection and treatment is key while it’s localized to the lungs. Once in the bloodstream, it’s latches onto organs and becomes very hard to eliminate fully. As temps heat up and droughts become more frequent (see map below), this will increase in areas affected. Have even seen a couple of cases in NY. Had a rough case today of a young guy. https://apple.news/A_4hTLdTCR_62ZaoKRw3Aag

0 reply

0 recast

5 reactions

An issue with LinkedIn leading the AI implementation discourse is that there is little originality or disruption. Seeing this especially with healthcare use cases. Rather than create new rails and think from first principles, it’s all ‘let’s use AI to replace X’. Rather than do we even need this whole flow if X can be freed up to do something new’ Not sure that’s a good thing. Corporate speak only mean more bloat. Sadly most startups funded right now are tapping into insurance reimbursement for revenue which only reinforces that flow, hence optimizing for billing (just as what Epic did). Not looking at what benefits patients most.

0 reply

0 recast

2 reactions

GLP-1s might be the biggest threat to the LDL hypothesis and #1 status of statins in cardiovascular protection. It may prove that insulin resistance >>>> LDL optimization in preventing cardiovascular disease. Long term studies pending still.

2 replies

0 recast

4 reactions

Great stuff! The soft robot for intubation is interesting and could be big for those in the field such as EMT and non-specialized first responders (police, even laymen). One commonly unspoken consequence of the rise in obesity is the rise in difficult airways (rated by whats called a Mallampati score, with score of 3-4 being difficult airways) Though I think an easier path for non-specialized personnel is augmenting the Glidescope with AI directions similar to how the face is adjusted on ID verification. I really like using the glidescope which is what I trained on for the most part and considering their availability now even with first responders (Glidescope Go 2), it would be a good way until the robots are ubiquitous. I edited a picture of a glidescope with nannobanana below to show how it could look.

0 reply

0 recast

4 reactions

Out of the recent AI/ML development in med devices/imaging Cleerly really stands out. It is an AI/ML platform that helps provide coronary insights such as plaque analysis/burden and calculate stenosis using a Coronary CT Angio which is non-invasive. Coronary arteries of the heart are a black box as few tests can really give the full picture of what it going on. Calcium scans help, but there are more a rough picture. Functional tests are positive when stenosis is quite far progressed. The current and only method today for getting a picture of the stenosis and plaque burden is doing a coronary angiogram or cath. Most angiograms are done when red flags are present either in labs or symptoms. A vague picture can be ascertained using a combination of echocardiogram/EKG and blood markers, but cath is gold standard. Cardiac angiograms carry real risks, but also require specialists and incur high costs. Tools like Cleerly may lead to earlier detection, especially in patients with high genetic risk, positive Calcium scans or long history of metabolic disease and can be done at a younger age.

1 reply

0 recast

2 reactions

OpenEvidence touts a 100% score on USMLE then goes and raises a round 2 days later. Something seems off about this. I use OpenEvidence daily and run into misjudged decisions and errors nearly every 3rd query. It really struggles with higher degree reasoning for complex patients with multiple differentials. Similar to HLE, almost feels like they targeted optimizing for USMLE rather than real life patient care. https://www.openevidence.com/announcements/openevidence-creates-the-first-ai-in-history-to-score-a-perfect-100percent-on-the-united-states-medical-licensing-examination-usmle

3 replies

0 recast

3 reactions

I love my youtube Indian lecturers

2 replies

0 recast

3 reactions

About 585,000 women die annually from pregnancy related events!!! That’s a lot!! 😪😪 There is so much room for growth in terms of reducing maternal mortality rate.

1 reply

0 recast

3 reactions

Q1-Q2 2026 is lining up to be big for next gen GLP-1s with triple agonists expected to get approval. It will be led by the triple Gs Lily has retarutide in the pipeline with phase 3 finishing soon Novo just signed a deal with Chinese pharma company United Biotech for global launch. Most looking forward to the lean muscle mass preservation data on these 3Gs which many insiders are telling me may actually increase muscle mass! A big reason for this acceleration is that Ozempic will be out of patent in China in Q1 2026 so incentivizing a lot of combos as well such as CagriSema https://newatlas.com/disease/obesity/obesity-novo-nordisk-lilly/

1 reply

0 recast

4 reactions

I was today years old when I learned that Tylenol and Doliprane are the same thing. Apparently Acetaminophen is just how North Americans call paracetamol.

0 reply

0 recast

2 reactions