AI Based on ECG Detects Hidden Liver Cirrhosis in Routine Practice for the First Time
A clinical study published in Nature Medicine showed that a machine learning algorithm on standard ECG increases the detection of advanced liver fibrosis by more than 4 times. In a cluster-randomized trial involving 15,596 patients, the intervention increased the rate of new cirrhosis diagnoses from 0.5% to 1.0%.
This is not news for cardiologists. It's an obituary for oral statins and weekly injections.
VERVE-102 data, published on May 25 in NEJM, went unnoticed by the general public. But that's a mistake. Because in these 35 patients and 18 months of follow-up lies the answer to the question: "Where will $10 billion of Amgen and Novartis's annual revenue go by 2032?"
I'm writing this analysis for those who understand: gene therapy is no longer the domain of rare diseases. Welcome to the era of mass genome editing.
[The Essence]: What's Really Happening
The official version: "VERVE-102 is a single infusion that reduces LDL by 62% in patients with familial hypercholesterolemia. Safety is acceptable."
The unofficial version: Eli Lilly just showed how to sink a competitor's Phase 3 without even starting it.
The key word here is GalNAc-LNP. The predecessor VERVE-101 failed due to hepatotoxicity. In October 2025, competitors at Intellia had a clinical hold for the same reason. That's when the market decided: "The liver is the Achilles' heel of editing."
VERVE-102 uses a fundamentally different lipid nanoparticle with attached N-acetylgalactosamine (GalNAc). It's a sugar "pass" that binds to asialoglycoprotein receptors (ASGPR) on hepatocytes.
What does this provide?
- Dual targeting: The particle enters via both LDLR (classic) and ASGPR.
- Lower dose: Less inflammatory burden on the liver.
- No transaminitis hell: Unlike VERVE-101, where one patient had a grade 3 ALT elevation, VERVE-102 is clean.
Insider view: Lilly paid $1.3 billion in June 2025 not for the cholesterol data, but for the patent on this delivery technology. Because it's applicable to ANGPTL3 (VERVE-201 — triglyceride reduction), Lp(a), and any other gene in the liver.
Timeline and Context
The devil is in the details. Here's the timeline no one is discussing:
- 2023–2024: VERVE-101 shows LDL reduction up to 73% in one patient, but another experiences grade 3 thrombocytopenia + ALT elevation. Verve makes a Solomon-like decision: discontinue VERVE-101, launch VERVE-102 with a new LNP.
- April 2025: Verve shows data from 14 patients. Reduction 53–69%. Safe.
- June 2025: Eli Lilly writes a check for $1 billion (plus $300 million CVR for Phase 3 start). This wasn't an acquisition; it was a technology evacuation from a turbulence zone.
- May 25, 2026 (NEJM): Data from 35 patients. LDL reduction 62% at 1.0 mg/kg dose, PCSK9 reduction 88%. Effect lasts 18 months.
Now Lilly is preparing to start Phase 2 by the end of 2026. Note: no pauses for restructuring. This means the internal audit of the NEJM data was flawless.
Who Wins and Who Loses
Loser #1 (obvious to everyone): Amgen (Repatha).
The PCSK9 inhibitor market is $4 billion. These are injections every 2 or 4 weeks. VERVE-102 is "inject and forget." Yes, the risk of long-term oncological effects? Yes, unknown. But for young HeFH patients in their 30s, the prospect of getting an infusion once in a lifetime vs. injecting 260 times — the choice is clear.
Loser #2 (not obvious): Treating cardiologists.
Their monetization model is collapsing. Currently, a patient comes for cholesterol check-ups every 3–6 months. That means prescriptions, repeat visits, patient retention. VERVE-102 is a transaction. Come once, pay once, go and don't come back. Cardiovascular medicine is turning into surgery: come — fix — leave. Who will pay the salaries of mid-level staff in 10 years?
Winner (total): Eli Lilly.
Lilly didn't buy a therapy. Lilly bought a liver delivery platform. VERVE-201 targeting ANGPTL3 (HoFH) is already in Phase 1b. Imagine what can be done with the same GalNAc-LNP for treating glycogen storage disease, hemophilia B (factor IX is synthesized in the liver), or even metabolic liver diseases. That's a $50 billion market.
What the Media Isn't Saying
The main lie is the claim of "one-time treatment."
Yes, the gene is edited forever. But hepatocytes divide (the liver regenerates). The problem: during cell division, edited alleles can be "diluted." We don't know how the percentage of edited cells will change after 5 years.
The official NEJM data shows LDL reduction of 62% at 18 months. What about month 30? If the percentage drops to 40% — that's still a therapeutic effect, but not "complete remission."
Second, what's being kept quiet: The study excluded patients with active hepatitis or cirrhosis. In real life, 30% of hypercholesterolemia patients have NAFLD (non-alcoholic fatty liver disease). How will GalNAc-LNP behave in a diseased liver? No one knows. Clinical trials in that population haven't been conducted yet.
Third, most important for insiders: VERVE-102 is an adenine base editor. It doesn't cut DNA (like Cas9 in competitors), but replaces one letter with another. This radically reduces the risk of large chromosomal rearrangements. But a point mutation is still a mutation. If the base editor works in the wrong place (off-target), consequences may appear in 10 years. Regulators are turning a blind eye for now. But if one patient develops hepatocellular carcinoma (liver cancer) in 5 years, the project will be shut down overnight.
Forecast: Next 30 Days and 90 Days
Next 30 days:
Stocks of second-tier base editing companies (Beam Therapeutics) will get a short-term boost. But don't fall for it. Beam doesn't have such a delivery system (GalNAc-LNP) in a clinically proven form for cardiology. But Lilly does.
Next 90 days:
Lilly will officially announce the Phase 2 design. Pay attention to endpoints. If they choose a surrogate endpoint (LDL reduction) rather than MACE (cardiovascular events), Phase 3 will last not 5 years but 18 months. That means FDA approval as early as 2029–2030, not 2032.
Main signal for the market: Lilly will confirm or deny plans for VERVE-201 (ANGPTL3). If they launch Phase 1b for VERVE-201 in the coming quarters — know this: they are preparing to wipe out not only cholesterol but also triglycerides.
My conclusion: We are entering an era where chronic disease ceases to be chronic. The first (in 2029) will be HeFH. Then homozygous hypercholesterolemia. And then ordinary hypertension (the AGT gene is already in the crosshairs of other players). Big Pharma's business model is shifting from "selling a patient drugs for 20 years" to "charging them $200,000 once and giving a warranty."
Watch the ALT. And don't say you weren't warned.
— Editorial Team
