What Is Tumor-Agnostic Immunotherapy? How Biomarker-Based Cancer Treatment Works

What This Article Covers

For most of oncology's history, a drug's approval was tied to where a cancer started — a “lung cancer drug,” a “colon cancer drug.” Tumor-agnostic immunotherapy breaks that rule: it treats cancer based on a genetic or molecular signature found inside the tumor's cells, regardless of whether that tumor began in the colon, the uterus, the pancreas, or somewhere else entirely. This article walks through how that idea became real medicine — the two biomarkers (MSI-H/dMMR and TMB-High) that currently unlock FDA-approved immunotherapy across cancer types, the checkpoint inhibitors built around them (pembrolizumab and dostarlimab), the genuinely striking rectal cancer results that put this approach on the public's radar, and the real limits on who qualifies and who can even get tested.

The Old Rule, and Why It Started to Break

Cancer drugs were traditionally studied and approved organ by organ because that was the only practical way to run trials — enroll enough lung cancer patients, enough breast cancer patients, and test the drug in each group separately. But as genomic sequencing became routine, researchers noticed something inconvenient for that model: some tumors from wildly different organs shared the same underlying genetic defect, and drugs that targeted that defect worked similarly well regardless of the organ involved. Immune checkpoint inhibitors — drugs like pembrolizumab (Keytruda) that release the brakes on T cells — turned out to work especially well in tumors with one specific defect: a broken DNA “spell-checking” system called mismatch repair.

The Biomarkers That Matter: MSI-H/dMMR and TMB-High

When a tumor is “mismatch repair deficient” (dMMR), it can't fix routine copying errors in its DNA, so mutations pile up — a state detectable as “microsatellite instability–high” (MSI-H) on a lab test. A related but distinct biomarker, tumor mutational burden (TMB), simply counts the total number of mutations per megabase of tumor DNA, however they got there. Both biomarkers point to the same underlying reason immunotherapy tends to work: a heavily mutated tumor produces more abnormal proteins (neoantigens) that a reactivated immune system can recognize as foreign. dMMR/MSI-H tumors turn up in dozens of cancer types, from stomach to small intestine to rare cancers with no other approved options, which is exactly why regulators eventually approved treatment by biomarker rather than by organ.

The First Approval That Changed the Rules

In May 2017, the FDA granted pembrolizumab accelerated approval for any unresectable or metastatic MSI-H or dMMR solid tumor that had progressed after prior treatment — the agency's first-ever tissue/site-agnostic cancer drug approval. The decision rested on a pooled analysis of 149 patients across five trials and 15 cancer types, showing an objective response rate of 39.6% (11 complete responses, 48 partial responses), with 78% of responses lasting six months or longer. That is a genuinely strong result for previously treated, hard-to-treat cancers — and it was the proof of concept that a tumor's molecular fingerprint could stand in for its zip code. Pembrolizumab later received full (non-accelerated) FDA approval for this same MSI-H/dMMR population. In 2021, GSK's dostarlimab (Jemperli) followed with its own accelerated approval for adult dMMR recurrent or advanced solid tumors, giving patients and oncologists a second tumor-agnostic checkpoint inhibitor option.

TMB-High: A Second Path Into the Same Door

In June 2020, the FDA approved pembrolizumab for a second, broader biomarker: TMB-High solid tumors (≥10 mutations per megabase by an FDA-approved test), again regardless of tumor origin, for patients who had progressed after prior treatment with no satisfactory alternatives. The pivotal data, drawn from the KEYNOTE-158 trial, showed a 29% overall response rate (4% complete, 25% partial) in 102 TMB-High patients — meaningfully lower than the MSI-H numbers, and a reminder that “tumor-agnostic” is not one uniform effect but a family of related, imperfect signals. Response durability was still notable: 57% of responders remained in response at one year.

Dostarlimab and the Rectal Cancer Result That Got Everyone's Attention

The result that pulled tumor-agnostic immunotherapy into mainstream conversation came from a small Memorial Sloan Kettering trial testing dostarlimab, alone, in patients with dMMR locally advanced rectal cancer — before any surgery, chemotherapy, or radiation. In updated results reported in 2024, every one of the 42 evaluable patients (48 enrolled) achieved a clinical complete response, with a median follow-up of nearly 18 months and 24 patients holding that response for a year or more, none experiencing severe treatment-related side effects. It is a remarkable, closely watched result — and it deserves the caveats that go with any single-arm study of this size: there was no control group, it applies specifically to the roughly 5-10% of rectal cancers that are dMMR, and it has not itself led to an FDA approval. A global confirmatory trial (AZUR1) is underway to determine whether this becomes a new practice standard.

Who Actually Qualifies? The Caveats That Matter

Tumor-agnostic approval does not mean “works for everyone.” MSI-H/dMMR status shows up in only a minority of tumors overall — common in endometrial and colorectal cancers, present but rarer elsewhere — and finding it, or a TMB-High result, requires molecular testing (typically next-generation sequencing) that is not universally ordered or covered. Documented real-world disparities in who actually receives biomarker testing mean that access to these drugs is uneven well before efficacy even enters the picture. And even within an approved biomarker group, response rates of 29-40% mean most treated patients do not have a complete response — durable benefit for a meaningful minority, not a guarantee for all.

Bottom Line

Tumor-agnostic immunotherapy is one of the more genuinely important shifts in modern oncology: a tumor's molecular signature, not its address in the body, can now determine whether a checkpoint inhibitor is likely to help. The MSI-H/dMMR and TMB-High approvals for pembrolizumab, joined by dostarlimab for dMMR tumors, have given patients with rare or treatment-resistant cancers a real option they wouldn't otherwise have had — and the dostarlimab rectal cancer results show just how dramatic that benefit can be in the right biological context. The realistic picture still includes modest response rates for many patients, biomarkers that apply to a minority of tumors, and testing access that isn't equal for everyone who could benefit — caveats worth understanding fully before treating “tumor-agnostic” as a synonym for “works for any cancer.”