Methods For Evaluating Clonal Tumor Mutational Burden Marks a Pause Worth Noticing

What changes in this design

A mutation count from a tumor sample is usually treated as a single number covering everything found in the sample, with no separation between mutations present throughout the tumor and those confined to a minority of its cells.

Here the count is deliberately restricted to the clonal fraction - mutations carried by the whole tumor population rather than by isolated pockets of cells. The methods cover both arriving at that narrower figure and applying it afterward.

Two tumors could total the same number of mutations while differing sharply in how those mutations are distributed among cells. Separating the shared ones gives a different picture of what a patient is carrying than a single lumped total does.

Ten mutations per megabase

A line drawn at 10 mutations per megabase is one I keep returning to. On June 16, 2020, the FDA gave accelerated approval to pembrolizumab for TMB-high solid tumors, using FoundationOne CDx and that threshold. Count enough DNA changes and a person might qualify for a drug, whatever organ the cancer began in.

An empty result isn't what I mean to suggest. In KEYNOTE-158, 102 of 790 patients were TMB-high; their objective response rate was 29%, versus 6% in TMB-low patients, and 57% of responses lasted at least 12 months. But this approval rests on total TMB. It counts mutations whether every cancer cell carries them or only a smaller offshoot does.

Shared tumor mutations

Clonal TMB is what I care about because "clonal" changes the question. It asks not merely how many DNA changes a sample holds, but which are shared across the tumor's cells. That is, I think, the right distinction to test if we want to predict an immune treatment's chance of working.

The 2016 work I read gives the idea a footing, not a verdict. Clonal neoantigen burden was associated with survival in lung adenocarcinoma, and tumors enriched for clonal neoantigens showed greater sensitivity to PD-1 and CTLA-4 blockade in lung cancer and melanoma. Subclonal neoantigens induced by chemotherapy were enriched in poor responders. The work gives a reason, I take it, to separate shared mutations from late-arriving ones, not a settled cutoff for every cancer.

Foundation Medicine's new claim

That is why US20260253667A1 caught my attention. Published August 27, 2026, it is assigned to Foundation Medicine, whose FoundationOne CDx was used for the FDA's 10 mut/Mb TMB-high threshold. Its title, "Methods For Evaluating Clonal Tumor Mutational Burden," puts a clonal refinement beside the count that established the familiar line.

Filed August 30, 2023, and claiming an August 31, 2022 provisional, the application has 20 claims. Claim 1 covers determining cTMB; later claims reach treatment selection, subject identification, reports, survival prediction, and a system. The filing frames cTMB, as I read it, for clinical decisions, not merely a laboratory result. A patent claim draws a legal boundary. It cannot prove better outcomes.

An ovarian cancer subset

The sharpest clinical signal I found is outside this filing. In the 91-person Phase 2b VITAL study, Gradalis reported a cTMB-high ovarian-cancer subgroup whose tumors were homologous-recombination-proficient, or HRP: median overall survival was 68 months with Vigil and 19 months with placebo. The reported hazard ratio was 0.23; seven-year overall survival was 45% versus 8%. Blinded post-hoc whole-exome analysis identified a subset of 11 treated people and 12 controls, a scale I kept in view.

In February 2025, the FDA granted Vigil Regenerative Medicine Advanced Therapy designation for a defined HRP, cTMB-high ovarian-cancer population in complete response after frontline treatment. Designation is something I don't treat as approval or proof of benefit. A clonal measure is moving into a regulatory development path before this publication, as I see it.

The harder half of the problem

The old cutoff has not stopped being fragile simply because a more focused measurement is being patented. Experts publicly disputed 10 mut/Mb after the 2020 approval. In one multi-laboratory comparison, a panel value of 10.2 recalibrated to 9.9 whole-exome-equivalent mutations per megabase, with an uncertainty interval from 5.9 to 16.5. That sample would not, I would say, be cleanly classifiable on either side of 10.

Nor does high TMB reliably rescue every immunotherapy idea. In a 2025 CTLA-4 blockade study, 21 people selected for high TMB missed the primary endpoint: six-month progression-free survival was 6%, median progression-free survival 1.9 months, and median overall survival 6.3 months. Two had partial responses, both above 20 mut/Mb. The study reads to me as a warning against treating a mutation tally as a universal passport.

This filing marks, I think, a sensible turn from sheer quantity toward shared targets. Prospective work must show that cTMB changes decisions and outcomes across assays and tumor types before I'll call it a better clinical test. The harder half is proving that the sharper count helps real people before we use it as a clinical guide.

Sources

Reporting for this piece drew on the published filing and the references below.

1. US20260253667 patent application record
2. FDA approval of pembrolizumab for TMB-H solid tumors
3. Clonal neoantigens and sensitivity to immune checkpoint blockade
4. TMB cutoff debate after tissue-agnostic approval
5. Cross-laboratory TMB calibration method
6. Phase 2 tremelimumab study in high-TMB advanced cancers
7. VITAL cTMB-H and HRP ovarian cancer analysis
8. Vigil receives FDA RMAT designation

The record, briefly. US20260253667A1, “Methods For Evaluating Clonal Tumor Mutational Burden” The publication describes methods for determining clonal tumor mutational burden (cTMB) in a tumor sample or tumor-derived sample from a subject, and methods for using the determined cTMB.

Identity details. The saved title-page transcription needs review, so names are not repeated here. Check the official published title page.

More angles on this

• Methods Of Identification And Diagnosis Of Lung Diseases Using Makes a Small Case for Better TimingAnother reading from the same G16B classification family, bioinformatics.
• Device And Method Of Virtual Reality Interaction Based On Bimanual Treats the Relationship as the PointAnother reading from the same G16B classification family, bioinformatics.
• CPC G16B: BioinformaticsThe official classification family this record was filed into, published in the USPTO Cooperative Patent Classification scheme.
• CPC C12Q: Measuring or testing processes involving enzymes, nucleic acids or microorganismsThe official classification family this record was filed into, published in the USPTO Cooperative Patent Classification scheme.
• Eight labels, two numbers, and the evidence a person gets before deciding anythingAnother reading from the neighbouring G16H classification family, healthcare informatics.

Read around this one:Public patent recordUSPTO published figure