Why a $7 Drug Might Beat a $20,000 Chemotherapy Infusion — And Why Nobody Wants You to Know

A beginner's guide to what we're building at CrisPRO.ai

We didn't set out to challenge the pharmaceutical industry.

We set out to build a better cancer decision engine. But when you follow the science honestly — no agenda, no funding from anyone who sells drugs — you end up in uncomfortable territory very quickly.

This is the story of how we got there.

Where It Started: Debunking a Distraction

A few months ago, we ran Ivermectin through our computational engine. You've probably heard the hype — social media convinced millions of people that this antiparasitic drug was a miracle cancer treatment.

Our engine ran the numbers. Ivermectin failed. The IC50 scores — the concentration needed to actually kill cancer cells — were unimpressive. The published late-phase clinical evidence was thin. We called it what it was: a distraction.

But here's what nobody tells you about debunking something. When you clear away the noise, sometimes you find a signal underneath.

And we found one.

The Signal We Weren't Looking For: Metformin

While our engine was running the Ivermectin analysis, it flagged something unexpected sitting right next to it in the data.

Metformin.

You probably know Metformin as a diabetes drug. Costs $7 to $15 a month at any pharmacy. Been around since the 1950s. About as glamorous as a glass of water.

But our engine kept ranking it unusually high for ovarian cancer. The IC50 scores were credible. The mechanism was real. And when we dug into the published literature, we found something that genuinely surprised us.

The scientific community already knew about this.

There are published studies — peer-reviewed, credible research — showing that Metformin doesn't poison cancer cells the way traditional chemotherapy does. Instead, it does something more elegant. It cuts off the tumor's energy supply. Cancer cells are energy-hungry machines. They hijack your body's insulin and growth hormone signals to fuel their rapid multiplication. Metformin severs those fuel lines. It activates a cellular alarm system called AMPK that essentially tells cancer cells: you don't have enough energy to divide right now.

Then it blocks mTOR — the master switch that tells cells to grow, multiply, and build new proteins. Turn that switch off in a cancer cell and the cell starves.

The biology is not controversial. Multiple research institutions have confirmed it.

So why aren't oncologists prescribing Metformin for cancer?

The Uncomfortable Answer: Follow the Money

This is where the story gets harder to tell — not because it's complicated, but because it sounds like a conspiracy theory until you look at the numbers.

A standard chemotherapy infusion costs $20,000 or more per treatment. Patients receive multiple cycles. The global chemotherapy market generates hundreds of billions of dollars every year. The companies that manufacture these drugs have market capitalizations larger than many countries' economies.

Metformin costs four cents a pill. It is off-patent. Nobody owns it. Nobody can charge $20,000 for it.

To prove definitively that Metformin treats cancer, someone needs to fund a proper Phase III clinical trial — the gold standard of medical evidence. These trials cost between $50 million and $200 million dollars. They require years of work and thousands of patients.

Here is the simple question nobody in the establishment will answer directly: Who is going to spend $100 million to prove that a drug they can sell for four cents works?

The answer is: nobody. Not willingly.

So the trials don't get funded. The guidelines don't get updated. Oncologists keep prescribing $20,000 infusions. And patients who might have responded to a $7 drug never find out.

What We Found Next: A Hidden Accomplice

While investigating Metformin, our engine surfaced another compound that operates on a completely different mechanism but attacks the same problem from a different angle.

Dupilumab — sold as Dupixent by Sanofi and Regeneron for allergies and eczema — generates $14 billion per year in revenue. It was never designed for cancer. But it blocks two proteins called IL-4 and IL-13, and those proteins do something inside tumors that the allergy world never paid much attention to.

They recruit the tumor's personal army.

Solid tumors — the kind you find in the ovary, lung, and kidney — are not just cancer cells. They are ecosystems. And one of the most important members of that ecosystem is a type of immune cell called an M2 macrophage. These cells are supposed to fight infections. Inside tumors, they get hijacked. They switch sides. IL-4 and IL-13 are the signals that flip them from protectors into tumor bodyguards — suppressing the immune system's ability to attack the cancer and building a physical shield around the tumor.

Dupilumab blocks those signals. And when you do that in laboratory models, something remarkable happens. The tumor loses its bodyguards. The immune system gets back in. The T-cells — the cancer killers — can suddenly reach their target.

This research is published. It is real. A paper in Cancer Immunology Research in early 2026 showed Dupilumab directly reduced malignant T-cell proliferation and promoted anti-tumor immunity.

But Sanofi and Regeneron make their $14 billion selling it for eczema. Running a cancer trial risks cannibalizing their existing revenue streams. And if it succeeds, it threatens the entire $25-billion-per-year checkpoint inhibitor market that companies like Merck have built around Keytruda.

The Triple Hypothesis

Here is where CrisPRO.ai's computational engine started generating something genuinely exciting.

What if you combined all three?

• Metformin to starve the tumor's energy supply and block mTOR

• Dupilumab to strip away the tumor's immune shield

• A checkpoint inhibitor to activate the immune system's T-cells against the now-exposed tumor

Three drugs. Three different mechanisms. Three angles of attack simultaneously.

The individual science for each exists in peer-reviewed literature. Nobody has computationally modeled the three-way combination in a real ovarian cancer patient cohort — until now.

What We Built and Why It Matters

This is what CrisPRO.ai's engine actually does, in plain language.

We take publicly available genomic data from real cancer patients — data collected by the National Cancer Institute and sitting in open databases. We run it through our engine, which scores each patient's tumor for three things simultaneously:

1. Is this tumor's immune microenvironment dominated by M2 macrophages driven by IL-4/IL-13? (Dupilumab target)

2. Is this tumor's mTOR pathway hyperactive with suppressed AMPK and high insulin signaling? (Metformin target)

3. Does this tumor have CCNE1 amplification or MBD4 silencing? (The genetic vulnerability layer)

When all three conditions overlap in a single patient's tumor, our engine flags them as a candidate for the triple combination approach.

We ran this on 427 real ovarian cancer patients from a large TCGA genomic database.

The pipeline is running. The results are coming in.

Why We Have a Good Feeling

We are being careful. Science requires honesty above ambition. We are not claiming we have cured ovarian cancer. We are not telling patients to go home and take Metformin instead of chemotherapy.

What we are saying is this:

There is a subgroup of ovarian cancer patients — identifiable from their tumor's genomic profile — who may respond dramatically to a combination of inexpensive, well-tolerated drugs that the establishment has no financial incentive to test together.

We can find those patients computationally, for essentially zero cost, using publicly available data.

The major pharmaceutical companies know pieces of this story. GSK filed patents covering MBD4 and PARP combinations in 2018. Sanofi knows Dupilumab affects the tumor microenvironment. Multiple academic labs have published on Metformin's anti-cancer mechanisms.

But nobody has connected all the dots at once. Nobody has asked: what happens when you find the patient whose tumor is simultaneously vulnerable to all three interventions?

That is the question CrisPRO.ai was built to answer.

We started by debunking a distraction. We followed the science where it led. We found signals the establishment had financial reasons to ignore. And we built a computational engine to find the patients who are falling through the cracks of a system designed around profit, not biology.

The data is almost ready. The pipeline is running.

We're seeking partners, collaborators, and anyone sitting on clinical datasets who wants to find out what the science actually says — regardless of what it costs per pill.