CDI Centro Diagnostico Italiano

🤖⚠️ “AI will replace physicians.” This paper basically says: wrong problem. This JAMA viewpoint flips the narrative in a very uncomfortable way: 👉 AI didn’t become more human. 👉 Medicine became less human. We’re all impressed that AI scores higher on “empathy”. But look closely: 👉 those studies are text-based 👉 no real patients 👉 no physical exam 👉 no uncertainty 👉 no responsibility The real issue is elsewhere Over decades, medicine has been buried under: - documentation - billing - prior authorization - inbox management - performance metrics Result? 👉 physicians spend ~50% of time on screens 👉 ~27% with patients So what happened? We didn’t lose empathy because AI gained it. We lost it because: 👉 we engineered physicians away from the bedside. The key concept (and it’s a good one) The authors call it: 👉 “excavation” AI doesn’t replace doctors. 👉 it removes what replaced them. But here’s the critical nuance (not fully addressed) AI can: ✔ reduce administrative burden ✔ assist reasoning ✔ improve efficiency But it can also: ❌ increase throughput pressure ❌ generate more data, more summaries ❌ create new layers of digital noise Translation AI does not automatically bring physicians back to patients. It can also: 👉 make absence more efficient. The real fork in the road As the paper says: 👉 one path → back to bedside 👉 one path → further industrialization of care My take This is not a technology problem. It’s a system design problem. If you optimize for: 👉 volume 👉 speed 👉 revenue AI will amplify: 👉 depersonalized medicine If you optimize for: 👉 presence 👉 judgment 👉 patient interaction AI can finally: 👉 restore what medicine was supposed to be Bottom line AI is not the end of the physician. But it will expose very clearly: 👉 what we actually value in healthcare. #AIinMedicine #Healthcare #Radiology #DigitalHealth #MedicalAI #FutureOfMedicine

🫀⚠️ We used to say: “It’s non-obstructive, don’t worry.” This paper suggests we may have been… very wrong. Photon Counting CT answers NEW questions. A simple but powerful case: 👉 74-year-old patient 👉 unstable angina 👉 non-obstructive disease on CT And yet: 👉 Photon Counting CT detected plaque ulceration. Confirmed by OCT. Let that sink in. This was NOT a tight stenosis. This was: 👉 biologically dangerous plaque 👉 invisible to conventional paradigms 👉 clinically relevant The key finding Ultra-high-resolution PCCT showed: 👉 non-calcified plaque 👉 spotty calcification 👉 positive remodeling 👉 sharp angular protrusion (ulceration) —all non-invasively. Translation We are starting to see: 👉 plaque rupture biology without a catheter. And this matters. Because plaque ulceration (cap disruption + cavity formation) is: 👉 one of the main substrates of ACS Traditionally detected only with: ❌ invasive angiography ❌ OCT This is the real shift From: ❌ “Is there a stenosis?” To: 👉 “Is this plaque dangerous?” The uncomfortable implication A patient can have: 👉 “non-obstructive CAD” and still carry: 👉 rupture-prone disease. My take Photon Counting CT is not just improving image quality. It is: 👉 exposing the blind spots of lumen-based cardiology. Bottom line If you only look for stenosis: 👉 you will miss biology 👉 you will miss risk 👉 you will miss patients ⚡ The future is not: “better detection of narrowing” It is: 👉 detection of instability #PCCT #PhotonCounting #CCTA #PlaqueImaging #Atherosclerosis #Cardiology #PrecisionMedicine #CardiacCT

🫀⚠️ Not all hypertensive hearts fail the same way. And this paper quietly destroys the idea that “LVH is LVH.” Because hypertension does not produce one cardiac phenotype. It produces: 👉 different geometries 👉 different fibrosis patterns 👉 different functional trajectories 👉 dramatically different outcomes —all hiding behind the same blood pressure number. Using CMR in >24,000 hypertensive patients from UK Biobank, the authors identified four phenotypes: - normal LV - LV remodeling - eccentric LVH - concentric LVH And the differences were striking. The most dangerous phenotype? ⚠️ Eccentric LVH. These patients had: 👉 the worst ventricular function 👉 impaired strain 👉 chamber dilation 👉 larger atria 👉 the highest event rates Including: 👉 9-fold higher heart failure risk. Meanwhile concentric LVH showed something different: 👉 highest native T1 👉 greatest wall thickness 👉 stronger fibrosis signature Basically: pressure-overload biology written directly into the myocardium. This is where CMR becomes extremely powerful Because hypertension is not simply: ❌ “a blood pressure disease” It is: 👉 a myocardial remodeling disease And blood pressure alone explains surprisingly little about who progresses badly. The paper even states the correlation between BP and LV mass is only modest. Translation Two patients can have: 👉 similar BP 👉 similar medications 👉 similar clinic visits …but completely different myocardial biology. One adaptive. One maladaptive. One fibrotic. One progressing toward heart failure. And standard echocardiographic labels may not be enough anymore. Because CMR reveals: ⚡ fibrosis (T1) ⚡ strain abnormalities ⚡ chamber remodeling ⚡ phenotype-specific trajectories before overt systolic collapse. My take We still treat hypertension largely as: 👉 a hemodynamic number When advanced imaging increasingly shows it is: 👉 a structural and biological myocardial disease. And PCCT can provide basically the same information as CMR in this context. The real future Not: ❌ “treat the pressure” But: 👉 phenotype the heart. Because ultimately patients do not develop events from: 👉 systolic BP values They develop events from: 👉 myocardial remodeling biology. ⚡ The next frontier of hypertension management may not be stricter cuffs. It may be: imaging-guided myocardial phenotyping. #CMR #Hypertension #LVH #HeartFailure #CardiacMRI #PrecisionMedicine #Cardiology #MyocardialFibrosis