Signatera Test: How ctDNA Blood Tests Detect Recurrence

Signatera is a blood test that looks for tiny traces of a patient’s specific tumor DNA circulating in the bloodstream after cancer surgery. Made by Natera, it belongs to a class of tests designed to detect molecular residual disease, meaning cancer cells too few to show up on a scan but numerous enough to shed detectable DNA fragments into the blood. The test has earned FDA breakthrough device designation for use across multiple solid tumors, and a growing body of evidence suggests it can flag recurrence months before imaging catches it.¹ How it works, where the evidence is strongest, and what the test can and cannot tell you are more nuanced than most marketing materials suggest.

How the Test Works

Signatera is what’s known as a tumor-informed assay. That distinction matters. Rather than scanning the blood for a fixed set of common cancer mutations, the test starts with your actual tumor tissue, removed during surgery or biopsy. The lab sequences a large portion of the tumor’s genome to identify mutations unique to that specific cancer. From those, a custom panel of targets is built for each patient. When a blood sample is drawn later, the lab uses ultradeep sequencing to search specifically for those pre-identified mutations in the cell-free DNA floating in the bloodstream.²

This personalized approach gives the test a significant advantage over methods that look for a generic set of mutations. Because the panel is bespoke, the assay can distinguish true tumor-derived DNA from background noise, including mutations that arise naturally in blood-forming cells as people age. That age-related phenomenon, called clonal hematopoiesis, is a well-known source of false positives for tests that aren’t tumor-informed. By tracking only variants already confirmed in the patient’s tumor, Signatera sidesteps most of that interference.³

The trade-off is that the test needs adequate tumor tissue to begin with, and the custom panel takes time to design. In a real-world study of patients with liver and pancreatic cancers, ctDNA was successfully evaluable in about 85% of cases, but success rates varied by tumor type, dropping to roughly half for one subtype of bile duct cancer.⁴ So while the concept is elegant, the practical yield depends on having enough quality tissue and a tumor that sheds enough DNA into the blood.

Colorectal Cancer Has the Deepest Evidence

If there’s one cancer type where the data behind ctDNA-guided care is most mature, it’s colorectal cancer. Multiple studies now converge on the same finding: patients who test positive for ctDNA after surgery are at dramatically higher risk of recurrence compared to those who test negative.

In a large interim analysis of the BESPOKE CRC study, patients with detectable ctDNA after surgery (called MRD-positive) had roughly 20 times the risk of disease recurrence compared to MRD-negative patients across stages II and III combined. The effect was consistent regardless of stage. Among those who were MRD-positive, receiving adjuvant chemotherapy extended median disease-free survival from about 7 months to nearly 19 months. Among MRD-negative patients, chemotherapy provided no measurable benefit at all.⁵

A separate real-world study of resected stage I through III colon cancer reinforced those results. Patients positive for ctDNA in the post-surgical window had nearly ten times the recurrence risk, and that ratio climbed even higher during ongoing surveillance. Chemotherapy helped the ctDNA-positive group but made no difference for ctDNA-negative patients.⁶ A pooled analysis of three colorectal cancer cohorts found that roughly 12% of patients had detectable ctDNA at the first post-surgical blood draw, taken four to ten weeks after surgery, and that detection rate stayed stable regardless of whether the draw was at four weeks or ten.⁷

The clinical implication is striking: if roughly one in eight patients tests positive after surgery and those patients are the ones who actually benefit from chemotherapy, the test could spare the other seven from months of toxic treatment that wouldn’t have helped them anyway.

What About Other Cancers

The evidence is expanding beyond colorectal cancer, though the data in other tumor types is not as deep.

In breast cancer, a study of patients with early-stage disease found that those with detectable ctDNA after surgery had about five and a half times the risk of developing distant recurrence compared to those without it. Perhaps most usefully, ctDNA detection preceded clinical relapse by a median of 16 months, giving a substantial window in which treatment could theoretically be started earlier.⁸ Longer-term follow-up in a separate breast cancer cohort confirmed that serial blood draws over time, rather than a single snapshot, provide the strongest prognostic signal.⁹

In melanoma, longitudinal ctDNA tracking showed that rising ctDNA levels preceded radiologic progression by a median of about 11 months, providing an even longer early-warning window than most imaging schedules would catch.¹⁰ In bladder cancer, a retrospective study found that patients whose ctDNA levels dropped or converted from positive to negative during treatment had better outcomes, while persistent positivity was associated with high mortality.¹¹

What all these cancer types share is the same basic finding: detectable ctDNA after treatment is a strong negative prognostic signal, and trends over time matter more than any single blood draw. The strength of the association varies, and sample sizes in bladder and liver cancers are still small, but the direction of the evidence is consistent.

The Lead Time Advantage Over Imaging

One of the most compelling selling points of ctDNA testing is the lead time it provides over conventional scans. Cancer recurrence usually has to reach a certain physical size before a CT scan or MRI can detect it. A few hundred or thousand cells growing in a lymph node or organ won’t light up on imaging. But those cells can shed DNA fragments into the bloodstream that a sensitive assay picks up.

Across the cancers studied so far, this lead time ranges from roughly a year to well over a year. The breast cancer data showed a median lead of 16 months before clinical relapse.¹² The melanoma data showed about 11 months.¹³ Whether detecting recurrence earlier actually translates to better survival is the critical unanswered question. Lead time is only useful if earlier intervention changes outcomes. The clinical trials currently underway are designed to answer exactly that.

Tumor-Informed Versus Tumor-Agnostic Tests

Signatera is not the only ctDNA test on the market. Other commercial options include tests from Guardant Health, Foundation Medicine, and others. Some of these are tumor-agnostic, meaning they look for a fixed panel of known cancer-associated mutations in the blood without first profiling the patient’s individual tumor. Both approaches have their place, but the performance gap widens depending on how the test is used.

A systematic review and meta-analysis comparing the two approaches in colorectal cancer found that when blood is drawn at a single fixed timepoint after surgery, both tumor-informed and tumor-agnostic assays performed similarly. But when blood was drawn repeatedly over time, tumor-informed assays like Signatera significantly outperformed tumor-agnostic ones, with a sensitivity of 88% versus about 60% and no meaningful difference in false-positive rates.¹⁴ A second meta-analysis reached nearly identical numbers, finding 88% versus 62% sensitivity for serial sampling, and demonstrated stronger prognostic discrimination for tumor-informed assays overall.¹⁵

The practical takeaway: if you’re only going to get a single blood draw, the choice of assay matters less. But for ongoing surveillance, where the whole point is to catch a trend before it becomes a visible mass, the tumor-informed approach appears to have a clear edge. The caveat, of course, is that tumor-informed tests require tumor tissue and a longer setup time, so in some clinical situations a tumor-agnostic test may be the only practical option.

Clinical Trials Using ctDNA to Guide Treatment Decisions

The evidence described so far is mostly observational: ctDNA status predicts who will relapse, but the treatment decisions were still made the old-fashioned way. The next generation of trials is interventional, meaning the test result itself determines what treatment a patient gets.

The DYNAMIC-III trial enrolled patients with stage III colon cancer and randomized them to either standard treatment decisions or ctDNA-guided decisions. In the ctDNA-guided arm, patients who tested negative had their therapy scaled back, while those who tested positive received escalated therapy.¹⁶ The PEGASUS trial took a similar approach, de-escalating treatment for ctDNA-negative patients and escalating it for ctDNA-positive patients, with an additional step up if ctDNA remained positive after the first round of chemotherapy.¹⁷

These trials represent a fundamental shift in how post-surgical cancer care could work. Instead of giving everyone the same chemotherapy regimen based on staging alone and accepting that most patients either didn’t need it or it didn’t help them, the idea is to match the intensity of treatment to the actual biological risk measured in the blood. It’s personalized not by genomic subtype or demographic profile, but by whether there’s detectable evidence of disease still present in the body.

Cost and Coverage

Signatera is not cheap. A single test typically costs several thousand dollars, and surveillance protocols call for repeated draws over months or years. This naturally raises questions about whether routine use would be cost-effective for healthcare systems.

A cost-effectiveness analysis focused on early-stage colorectal cancer found that adding ctDNA-based MRD testing to traditional clinical risk factors for guiding chemotherapy decisions led to an estimated increase of about 0.7 quality-adjusted life years per patient and a cost savings of roughly $9,800 per patient compared to using clinical factors alone.¹⁸ The savings come largely from avoiding unnecessary chemotherapy in the majority of patients who are MRD-negative. Chemotherapy is expensive, causes side effects that generate their own medical costs, and reduces quality of life. Skipping it when there’s no detectable disease to treat saves money and suffering.

Insurance coverage, though, remains uneven. A review of private payer and Medicare coverage policies found significant variability in which ctDNA tests are covered and for what indications. The FDA’s breakthrough device designation for Signatera has helped with some payers, but coverage often lags behind the clinical evidence, especially for newer indications beyond colorectal cancer.¹⁹ If you’re considering this test, checking with your insurance before the blood draw is worth the phone call.

Limitations You Should Know About

The enthusiasm around ctDNA testing is warranted, but it comes with real caveats that are easy to overlook.

A negative result does not guarantee you’re cancer-free. Some tumors shed very little DNA into the blood. Brain tumors, for instance, are shielded by the blood-brain barrier and tend to release minimal ctDNA into peripheral circulation. Certain early-stage cancers simply don’t produce enough cell-free DNA to be reliably detected. The ESMO Precision Medicine Working Group has noted that tissue-based testing remains preferred in many situations because ctDNA assays still have limitations in detecting certain types of genomic changes, including structural rearrangements and gene copy number changes.²⁰

Tumor heterogeneity can also complicate matters. A study in clear cell kidney cancer compared custom panels built from different parts of the same patient’s tumor, including both the primary tumor and metastatic sites. In only three out of five patients were the MRD results fully concordant across all panel designs. In the two discordant cases, a panel built from one metastatic site gave a different MRD call than panels built from other sites.²¹ If the mutations chosen for tracking aren’t shared across all parts of the cancer, the test could miss residual disease harbored in a subclone not represented on the panel.

Then there’s the practical issue of assay failure. As noted in the hepatopancreatobiliary cancer study, success rates vary by tumor type, and in some cases the test simply can’t be completed, whether due to insufficient tissue, low tumor mutation burden, or technical issues with panel design.²²

What Changes in the Blood Over Time

One of the more useful aspects of ctDNA monitoring isn’t just whether the test is positive or negative at a single point, but the direction of the trend. Researchers are increasingly finding that the trajectory of ctDNA levels carries as much or more prognostic information as the binary positive/negative call.

In patients with metastatic cancer receiving immunotherapy, changes in ctDNA concentration over the course of treatment were significantly associated with progression-free survival. The association was numerically stronger when ctDNA was measured in mean tumor molecules per milliliter of blood rather than as a simple allele frequency.²³ In practical terms, a patient whose ctDNA levels drop sharply after starting treatment is getting a real-time signal that the therapy is working, weeks or months before a scan might confirm tumor shrinkage. Conversely, rising levels can signal resistance or progression early enough to consider switching strategies.

This trend-based monitoring also explains why serial sampling outperforms single timepoint testing. A single negative draw could reflect a momentary dip in shedding. But three consecutive negative results, or a steady downward trend from a positive starting point, carries much stronger prognostic weight. The melanoma and breast cancer studies both underscore this: longitudinal patterns were far more informative than any isolated result.²⁴

The Colorectal Cancer Experience as a Template

How ctDNA testing has evolved in colorectal cancer is likely a preview of what other tumor types will go through over the next several years. The progression has followed a recognizable pattern: first, retrospective studies showing ctDNA positivity predicts relapse; then prospective validation confirming the same; then cost-effectiveness modeling; then interventional trials using the result to change care; and finally, guideline incorporation.

Colorectal cancer is furthest along this path, with both the BESPOKE and DYNAMIC studies generating the kind of data that professional guidelines committees look for. The colorectal data also established an important principle that will carry over to other cancers: the biggest value of ctDNA testing may not be in catching recurrence earlier but in identifying who can safely skip chemotherapy. In the BESPOKE CRC interim analysis, MRD-negative patients had excellent outcomes regardless of whether they received chemotherapy, with no detectable survival benefit from treatment.²⁵ That finding, if it holds up in larger randomized trials, could change the default for hundreds of thousands of cancer patients a year from “treat unless there’s a reason not to” to “don’t treat unless the blood says there’s something to treat.”

For patients navigating a new cancer diagnosis, the question to ask your oncologist isn’t just whether Signatera or another ctDNA test is available, but whether your specific tumor type and stage have enough evidence behind ctDNA-guided decisions to make the result actionable. In colorectal cancer, the answer is increasingly yes. In breast cancer and melanoma, the evidence is strong enough to be clinically informative even if treatment algorithms haven’t fully caught up yet. In rarer cancers, the test may provide peace of mind or early warning, but the evidence base for acting on results is still thin.