In partnership with
Wake up to better business news
Some business news reads like a lullaby.
Morning Brew is the opposite.
A free daily newsletter that breaks down what’s happening in business and culture — clearly, quickly, and with enough personality to keep things interesting.
Each morning brings a sharp, easy-to-read rundown of what matters, why it matters, and what it means to you. Plus, there’s daily brain games everyone’s playing.
Business news, minus the snooze. Read by over 4 million people every morning.
Hi!
Welcome to AIMedily.
Today, I’m writing this newsletter next to my son, who’s feeling a bit under the weather. I’m hoping he wakes up much better tomorrow.
Here’s what stood out this week.
🤖 AIBytes
Researchers studied whether using AI as part of everyday hospital care could improve how sepsis is detected, documented, and treated. And whether this could improve patient outcomes.
🔬 Methods
Hospital: Lausanne University Hospital (CHUV), Switzerland
Patients:
97,559 hospital stays in intervention wards
25,851 hospital stays in matched control wards
Hospitals introduced a Sepsis Learning Health System that included:
A standard sepsis care pathway
A centralized sepsis registry
HERACLES, an AI model that continuously monitors patients
Every 6 hours, it classified patients as no sepsis, possible sepsis, or confirmed sepsis
📊Results
Sepsis detection:
AI identified sepsis in about 9–10% of hospital stays
Routine ICD-10 coding identified only 2–4%
AI accuracy (confirmed sepsis):
AUROC: 0.88
Precision: 0.76
Recall: 0.62
Documentation:
Sepsis coding improved significantly in AI-supported wards.
No improvement was seen in control wards.
Mortality:
In-hospital mortality decreased in AI-supported wards.
90-day mortality also decreased.
No similar improvement occurred in control wards.
Care delivery:
Patients were more likely to receive antibiotics within 1 hour when the sepsis pathway was followed.
🔑 Key Takeaways
AI worked best when used as part of a system, not as a stand-alone alert.
Traditional coding misses many sepsis cases.
AI-supported care was linked to lower mortality.
Continuous monitoring and clinician involvement were critical.
🔗 Despraz J, Matusiak R, Nektarijevic S, et al. An artificial intelligence-powered learning health system to improve sepsis detection and quality of care. npj Digital Medicine. 2026;9:106. doi:10.1038/s41746-025-02180-2
This review looks at how AI and machine learning are being used to study microRNAs (miRNAs) in cancer for diagnosis, prognosis, and biomarker development.
🔬 Methods
Data discussed:
Large public databases (e.g., TCGA, GEO)
Clinical cohorts from multiple cancer types
AI approaches:
Traditional Machine Learning
Deep learning models
Newer foundation and transformer-based models
Clinical use cases:
Cancer detection
Cancer subtype classification
Risk prediction and prognosis
📊 Results
Single miRNA biomarkers have not been accurate enough for routine clinical use.
Better results are seen when multiple miRNAs are combined into panels.
Some AI-based miRNA panels achieved AUC values above 0.90 in studies across several cancer types.
A serum miRNome model predicted the tissue of origin for 13 solid tumors with about 90% accuracy in early-stage disease.
AI models perform better than traditional statistics when combining miRNA data with clinical information.
miRNA-based therapies are still experimental:
Some trials were stopped due to immune-related toxicity, including inflammatory reactions.
Challenges remain with delivery, safety, and regulation.
🔑 Key Takeaways
AI has improved how miRNAs are studied in cancer, especially for diagnostic panels.
Results are promising, but most tools are not ready for routine clinical use.
miRNA-based treatments face important safety and delivery barriers.
🔗 Jurj, A., Dragomir, M.P., Li, Z. et al. MicroRNAs in oncology: a translational perspective in the era of AI. Nat Rev Clin Oncol (2026). https://doi.org/10.1038/s41571-025-01114-x
🦾TechTools
Autonomous AI diagnostic that detects more than mild diabetic retinopathy in adults with diabetes—no specialist image interpretation required.
Provides an immediate, point-of-care result to guide referrals.
FDA De Novo–cleared, designed to expand access to diabetic eye screening in primary care settings.
Ambient AI platform that captures clinical conversations and generates structured, specialty-specific notes directly inside the EHR.
Extracts key clinical elements to support coding and revenue cycle optimization, going beyond basic scribing.
Workflow automation (scheduling, reminders, admin support) to reduce documentation and operational burden.
Helps you quickly see the full picture by pulling together EHR data, claims, labs, and notes.
It’s used mainly in primary care and value-based care to highlight care gaps and improve documentation.
It works inside the EHR, with use shaped by clinical priorities and team workflows.
🧬AIMedily Snaps
ACCESS (Advancing Chronic Care with Effective, Scalable Solutions) Medicare Model (Link).
Can Medical AI Lie? Large Study Maps How LLMs Handle Health Misinformation (Link).
Price Explores Challenges of Medicine, AI, and the Need for a Doctor ‘In the Loop’ (Link).
Op-ed: Experience-centered AI is the future of healthcare innovation (Link).
Invisible Text Injection and Peer Review by AI Models (Link).
🧪Research Signals
AI succeeds in diagnosing rare diseases (Link).
Assessment of Short-Answer Questions by ChatGPT in a Medical School Course (Link).
AI tool predicts over 1,000 diseases years before they happen — and more are on the way (Link).
A scoping review of silent trials for medical artificial intelligence (Link).
The Missing Dimension in Clinical AI: Making Hidden Values Visible (Link).
Bridging AI and Clinical Reasoning (Link).
🧩TriviaRX
Which life-saving medical breakthrough was discovered in 1928 after a scientist returned from vacation and noticed mold growing in a petri dish?
A. Insulin
B. Penicillin
C. The polio vaccine
D. Cortisone
Now, the answer from last’s weel TriviaTX:
✅ B. The smallpox vaccine
In 1796, Edward Jenner inoculated 13-year-old James Phipps with material from a cowpox lesion, demonstrating protection against smallpox.
That’s all for this week.
As always, thank you for taking the time to read.
You’re already ahead of the curve in medical AI — don’t keep it to yourself. Forward AIMedily to your colleagues who’d appreciate the insights.
Until next Wednesday.
Itzel Fer, MD PM&R
Forwarded this email? Sign up here
P.S. Enjoying AIMedily? 👉 Write a review here (it takes less than a minute).
