Blood Tests, Brain Cells, and Billion-Dollar Bets: Health Tech's Biggest Week
From blood tests predicting heart disease 15 years out to a $1.5B cancer diagnostics deal, this week's health breakthroughs reshape medicine.
By Dr. Asher Knippel
AI Blood Test Predicts Six Cardiovascular Diseases Up to 15 Years Early
Researchers at the University of Hong Kong have developed CardiOmicScore, a blood-based test that analyzes thousands of proteins and metabolites to predict a person's risk of heart attack, stroke, heart failure, and three other cardiovascular diseases — as many as 15 years before symptoms appear. The study was published in Nature Communications.
What sets CardiOmicScore apart from existing genetic risk scores is that it reflects a patient's current biological state, not just inherited predispositions. Because it reads real-time signals — how proteins and metabolites are behaving in the body right now — it can capture the effects of lifestyle, diet, stress, and other factors that static DNA-based tests miss entirely.
Cardiovascular disease remains the world's leading cause of death. A test that flags elevated risk a decade or more in advance could give clinicians a meaningful window to intervene with lifestyle changes or medication before irreversible damage occurs. The team is working toward broader clinical validation.
A Banner Week for Alzheimer's Research at AAIC 2026
The Alzheimer's Association International Conference (AAIC 2026), held in London this week, produced a cluster of findings that together signal a significant shift in how the disease may be caught and managed.
The standout result involves the p-tau217 blood test, a biomarker that can detect early signs of Alzheimer's-related brain changes 5 to 10 years before cognitive symptoms emerge. Critically, researchers showed it performs well enough to be deployed by family doctors — not just specialists — dramatically expanding the potential reach of early screening. Current diagnostic pathways typically rely on expensive PET scans or invasive spinal taps; a reliable blood test changes the math entirely.
The conference also reported that structured lifestyle interventions — combining exercise, diet, cognitive engagement, and cardiovascular risk management — reduced cognitive decline across diverse global populations, reinforcing that risk is not fixed and that behavioral modification can make a meaningful difference.
A third finding came from a Phase 2 clinical trial of a cannabis-derived drug that reduced agitation in patients with advanced Alzheimer's. Agitation is one of the most distressing symptoms for patients and caregivers alike and currently has limited treatment options. The results are early-stage, but the signal was promising enough to warrant continued investigation.
Brain Immune Cells Are Stealing Sleep from Alzheimer's Patients — and Suppressing Them Helps
A parallel thread in Alzheimer's research this week came from the University of Kentucky, where scientists investigating the disease's notorious disruption of sleep identified a previously underappreciated culprit: microglia, the brain's resident immune cells.
Published in Alzheimer's & Dementia, the study found that microglia become chronically overactive as Alzheimer's progresses, and that this hyperactivation — not amyloid plaques — is the primary driver of sleep disturbance in their animal models. When researchers temporarily depleted microglia in mice, the animals recovered more than two hours of sleep per day.
The finding matters because sleep disruption in Alzheimer's is a major source of suffering and is linked to faster cognitive decline. If microglial overactivation is causing the problem, drugs that modulate microglial activity could become a new therapeutic avenue — one that's independent of the amyloid-targeting approaches that have dominated the field. The team is clear that this is early animal-model work and that translating it to humans will require considerable additional research.
Tempus AI Buys Personalis for $1.5 Billion to Expand Cancer Monitoring
On the commercial side of health tech, AI precision medicine company Tempus AI agreed to acquire Personalis — a leading developer of molecular residual disease (MRD) blood testing — for $1.5 billion.
MRD testing is an increasingly important tool in oncology. After a cancer patient undergoes surgery or treatment, the central question becomes: did we get it all? MRD tests look for traces of tumor DNA circulating in the blood — sometimes just a few fragments among billions of normal DNA molecules — to detect recurrence far earlier than imaging alone can.
Personalis has built a platform specifically designed for tumor-informed MRD monitoring, meaning its tests are personalized to each patient's unique cancer mutations rather than relying on generic cancer markers. Integrating that capability into Tempus's existing multimodal AI platform — which already covers genomic profiling and treatment selection — creates a more complete loop: from diagnosis to treatment choice to ongoing surveillance.
The deal reflects a broader industry conviction that catching cancer recurrence early, when the disease is still localized and more treatable, is one of the highest-value interventions in all of medicine.
Bristol Myers Squibb Puts NVIDIA's Most Powerful AI to Work on Drug Discovery
Bristol Myers Squibb has become the first pharmaceutical company to deploy NVIDIA's DGX SuperPOD system powered by Vera Rubin NVL72 accelerators — currently the most advanced AI computing platform available commercially.
BMS says the system delivers roughly 10 times greater AI performance per megawatt than its predecessor, an efficiency gain that matters enormously when running the large-scale molecular simulations and generative models that underpin modern drug discovery. The company's stated goal is to cut overall drug development timelines by 20 to 30 percent, with a longer-term target of 50 percent reduction.
Drug development is notoriously slow and expensive — the average new drug takes over a decade and costs billions of dollars to bring to market, with most candidates failing somewhere along the way. AI-accelerated discovery doesn't just speed things up; in principle it expands the search space of molecules that can be evaluated at all, potentially surfacing candidates that would never have been identified through conventional screening. Whether BMS's investment delivers on those promises will play out over years, but the bet signals where the industry's biggest players think leverage lies.