It Can Find Colorectal Cancer in Blood. Someone Else Sells That Test Alone.

In July 2026, the U.S. Food and Drug Administration approved a blood test. SimpleScreen CRC. One blood draw, to screen for colorectal cancer. Not a colonoscopy, not a stool sample: a screening option that runs on blood. The company that developed it is Freenome. About 120 million people in the United States are eligible for colorectal cancer screening, and 40 million to 50 million of them go unscreened, the filing says.

But the company that sells this test in the United States as a stand-alone colorectal cancer test is not Freenome.
Abbott has the exclusive right to commercialize it, and what comes back to Freenome is an upfront payment, milestones and tiered royalties.
The royalty is capped at 10%, and the cap is reached when the gross margin hits 20%. That is what the body of the filing says. The notes to the financial statements put it differently.

Freenome has taken in about $1.6 billion since it was founded, and by the end of June 2026 it had piled up an accumulated deficit of $1.5 billion. Cash, cash equivalents and short-term marketable securities at that same date came to $102.0 million.

And the S-1 filed on August 18, 2026 is not a document for raising new money. It registers the resale of shares held by existing securityholders. When those shares sell, the proceeds do not come to the company.

Ten questions. We lay out only what the S-1 says.

Question 1

What is this company, really?

In one line: a company that reads the signs of cancer out of blood.

Cancer is harder to act on the later it is found. The filing puts that state of things in numbers. Every day, cancer claims more than 1,600 lives in the United States alone. Only about 14% of cancers are detected by screening, and roughly half are not detected until an advanced stage.

What Freenome is trying to do is make the entrance to that a single blood draw. The company writes out its own mission this way.

“

Our mission is to detect cancer and disease at earlier, more treatable stages by making screening easy and accessible.

FREENOME S-1 (AUG 2026) / PROSPECTUS SUMMARY

The technology has a name: multiomics. Rather than look only at the DNA in blood, it reads several kinds of molecules at once, RNA and proteins among them, and combines them with machine learning. It is built to pick up both the signal a tumor puts out and the signal that comes from elsewhere, the immune system for instance.

Why look at more than one? Because some cancers shed a great deal of their own DNA into the blood and others shed almost none.

A figure in the filing lays the thinking out. Sensitivity from DNA alone is 43% in prostate cancer and 74% in pancreatic cancer. Combine several kinds of molecules and the numbers become 91% and 93%.

That figure carries a hard caveat, though: “Values shown are conceptual examples and do not represent clinical performance of any Freenome product candidate.” Case counts are written into the figure, and the company itself disowns the numbers all the same.

The first product approved in the United States is SimpleScreen CRC. A test for colorectal cancer, for adults 45 and older who are at average risk for the disease. It received FDA approval in July 2026.

The study behind it is PREEMPT CRC. It ran at 201 sites in the United States and the United Arab Emirates and met all primary endpoints, the filing says. More than 48,000 people were enrolled; about 27,000 were evaluated. Freenome funded and ran it itself.

The filing calls the approval “a blood-based screening option.” A screening option, not a diagnosis.

One thing should be clear here. This article does not recommend any test. Every sensitivity and specificity figure below comes from Freenome’s filing. The prostate and pancreatic numbers just cited are the ones the filing itself places as conceptual examples. And the approval is a U.S. FDA approval; it does not mean approval or availability anywhere else.

A company that bet on one thing: reading the signs of disease out of blood.

Question 2

What does it compete on?

Three fronts: how it reads, how it accumulates, and how rare the approval is.

Freenome is not the only company trying to find cancer in blood. The filing names as competitors Caris Life Sciences, Exact Sciences, GRAIL, Guardant Health, LabCorp, Natera and Quest Diagnostics.

Refusing “one size fits all”

The company’s technical position is explicit. The filing writes: “We fundamentally believe a ‘one size fits all’ technological approach is insufficient to detect every cancer across stages and subtypes.”

Concretely, it uses a non-bisulfite epigenetic assay with base-level resolution. A way of reading the chemical modifications on DNA at the level of the individual base rather than the fragment, as the company explains it.

On top of that it keeps deep learning models of its own. The structure learns a representation of each fragment from base-level features and rolls those up into a prediction for the sample. The protein assays were developed in-house, the filing says, and the autoantibodies came in with the 2023 acquisition of Oncimmune.

In v2, still in development, these improvements are expected to more than triple the number of distinct cell-free DNA molecules the assay can count. It is also targeting approximately 95% full automation in the lab workflow.

Note that the filing attaches a caveat to the tables and figures in this section: these are internal development metrics, not clinical performance claims.