Signatera Genome Predicted 93% of Post-Surgery Recurrences: What If the Test Is Positive?

Signatera Genome Predicted 93% of Post-Surgery Recurrences: What If the Test Is Positive?

Testing the blood after surgery for fragments of tumor DNA (ctDNA) is rapidly becoming a widely used way to reveal the risk of cancer returning before it shows up on imaging. We have previously covered an important study in this field, the GALAXY study, as well as approaches to de-escalating treatment with ctDNA. Now The Oncologist has published results from 386 patients with breast, lung, melanoma, kidney and colorectal cancer tested with Signatera Genome, the new whole-genome sequencing-based version of the Signatera test. The results are striking, but how to interpret them matters just as much as the results themselves.

Why does this matter?

In a substantial proportion of patients who undergo surgery for early-stage cancer, a number of cancer cells too small to be seen on imaging may remain in the body. This is called "molecular residual disease" (MRD), and it is the main source of recurrences. Today, decisions about adjuvant therapy after surgery rely largely on indirect markers such as stage, pathology and genomic risk tests. As a result, some patients receive treatment they do not need, while others are undertreated.

Circulating tumor DNA (ctDNA) testing attempts to measure residual disease directly. A positive test suggests that cancer cells remain, while a persistently negative test suggests a low risk of recurrence. This information could be used both to tailor adjuvant therapy to the individual and to detect recurrence earlier, perhaps at a still treatable stage. The approach can be thought of as extending the concept of biochemical recurrence (a rising PSA) in prostate cancer to other cancers.

Does a ctDNA test show that cancer will come back?

Largely yes, but not with certainty. In this study, patients who were ctDNA-positive after surgery had a very high risk of distant recurrence, while those who remained persistently negative had a very low risk. However, the test can miss some recurrences, particularly brain and locoregional recurrences. It has also not yet been proven that changing treatment based on a positive result prolongs survival.

How does the test work?

Signatera is a "tumor-informed" test: first, the tumor tissue removed at surgery and the patient's normal blood are sequenced, and mutations present only in the tumor are identified. A test customized for that patient is then designed, and these mutations are searched for at very high depth in subsequent blood samples. The original exome-based version examines only the roughly 1–2% of the genome that codes for proteins (the exome), whereas the Genome version looks at the tumor's whole genome using next-generation sequencing. This widens the pool of candidate mutations, allowing more and higher-quality alterations to be selected; the aim is to increase sensitivity in tumors that shed little DNA and at very early stages.

Bear in mind that this approach requires tumor tissue, that designing the initial test can take a few weeks, and that the test was developed not for cancer screening but for monitoring patients who have already been diagnosed and treated. Multi-cancer early detection tests that look for cancer in healthy people are a different technology answering a different question.

How was the study conducted?

Results

The first 3 months after surgery: who is at high risk?

Patients who were ctDNA-positive within the first three months after surgery had an approximately 24-fold higher risk of distant recurrence than those who were negative (HR 23.7). The absolute differences are even more telling: at 36 months, the proportion of patients alive without distant recurrence was 35.3% among ctDNA-positive patients and 93.9% among ctDNA-negative patients. In this early period, the test flagged 75% of recurrences in breast cancer and 71% in lung cancer in advance.

The finding on adjuvant therapy is also notable but must be interpreted with great caution. Among ctDNA-positive patients in the first 3 months, 12-month DRFS was 67.1% in the 23 patients who received adjuvant therapy and 33.3% in the 10 who did not. In ctDNA-negative patients, adjuvant therapy showed no clear benefit (91% and 94.4% at 24 months). This supports the hypothesis that "ctDNA-positive patients may benefit from treatment while ctDNA-negative patients may not." However, the groups were very small and not randomized, and it is unknown which patients were treated and why. The difference in the positive group was not statistically significant.

Post-treatment surveillance: how reassuring is remaining persistently negative?

During surveillance after completion of adjuvant therapy, patients who became ctDNA-positive at any time had an approximately 105-fold higher risk of distant recurrence than those who remained persistently negative (HR 104.7). At 36 months, 98.8% of persistently negative patients were free of distant recurrence, compared with only 14.7% of those who turned positive. The positivity rate increased with stage: 9.8% in stage I, 15.5% in stage II and 22.6% in stage III.

Note the numbers in parentheses: for example, the 100% sensitivity in kidney cancer is based on only 2 recurrences, and the 86% in colorectal cancer on 7. With numbers this small, percentages can easily shift. In 62% of positive results, the ctDNA level was below 100 molecules per million (0.01%), meaning the test's ultrasensitive range is genuinely being used. The authors also state that the Genome version detected recurrences a median of 3 months earlier than the exome version, but they do not present these data in the paper.

How should the results be read? Strengths and weaknesses

Strengths of the study

• 2,674 samples across five different cancers under real-world conditions, with strong results in the same direction for every cancer type.
• Nearly 99% DRFS in persistently negative patients: this has the potential to reduce the surveillance and treatment burden for some patients.
• The analysis used a statistical adjustment accounting for the different recurrence risks of different cancer types, and a landmark approach was chosen for the early period to avoid immortal time bias.

Points to keep in mind

• The very high hazard ratios partly reflect the design: Being "positive at any time" during surveillance includes turning positive very shortly before recurrence. In such cases, the test is partly detecting an imminent recurrence early rather than predicting it. A value such as HR 104.7 is therefore not surprising and does not by itself imply clinical benefit.
• Some recurrences were not counted: Brain metastases and isolated locoregional recurrences were excluded from the sensitivity calculation. Of 229 persistently negative patients, 7 (3%) developed a locoregional or brain recurrence or a new primary cancer. Because brain metastases are common in lung cancer, melanoma and breast cancer, this is an important limitation.
• Short follow-up: Median 22.5 months. In diseases that can recur many years later, such as hormone receptor-positive breast cancer and kidney cancer, late recurrences have not yet been captured.
• Small numbers: There were only 2 distant recurrences in kidney cancer and 7 in colorectal cancer. For some cancer types, the results have wide confidence intervals.
• Selection bias and conflicts of interest: Patients were those who had undergone commercial testing and whose clinical data were in the company database. Most authors are Natera employees; the paper states "funding: none," but the analyses were performed using company data and by company employees.
• Minor inconsistencies: The abstract reports an "approximately 30% improvement at 24 months" in ctDNA-positive patients who received adjuvant therapy, whereas the results section gives the same comparison using 12-month data. The "approximately 33% improvement" mentioned in the text is actually an absolute difference of 34 percentage points.

The real question: what do we do if the result is positive?

A test predicting recurrence and acting on that test benefiting the patient are two different things. The clearest data from randomized trials that guided treatment decisions by ctDNA come from colon cancer:

• DYNAMIC (stage II colon cancer, 2022): In the ctDNA-guided group, the proportion of patients receiving chemotherapy fell from 28% to 15%, while two-year recurrence-free survival remained similar (93.5% and 92.4%). In other words, de-escalating treatment in ctDNA-negative patients appeared safe.
• DYNAMIC-III (stage III colon cancer, 2025): In ctDNA-negative patients, de-escalating treatment was not shown to be as good as standard treatment in the overall group (85.3% and 88.1% at 3 years); it looks promising only in the low-risk subgroup. In ctDNA-positive patients, escalating chemotherapy did not improve recurrence-free survival.

For now, therefore, ctDNA testing is a very valuable tool, particularly for risk stratification and clinical trials; but evidence supporting its use as "if positive, give this treatment; if negative, stop treatment" exists only in a few specific situations. We also discussed the debate over shortening adjuvant therapy in colon cancer in our article on 3 months of oxaliplatin.

Who might benefit from this test?

• Colon cancer: This is where the evidence for using ctDNA in treatment decisions is strongest. It can be considered to support the adjuvant chemotherapy decision in patients with stage II disease.
• Patients with high-risk breast cancer, lung cancer and melanoma: In these cancers, where adjuvant treatment options are expanding (for example, abemaciclib in breast cancer, perioperative immunotherapy in lung cancer and immunotherapy in melanoma), the test may provide additional risk information; however, decisions should still follow guidelines.
• Patients who wish to join a clinical trial: Many trials are testing new treatments based on ctDNA positivity. For these patients, the test can be a gateway to trial enrollment.
• Who may not be a good candidate: Patients without available tumor tissue, those in whom brain metastasis risk is the main concern, and those whose treatment would not change even with a positive result. In patients with very low tolerance for uncertainty, the anxiety caused by a positive result that cannot be acted on should also be taken into account.

When a result is positive, the first step is usually to look for disease with imaging; methods such as PET-CT become valuable at this point. If imaging is negative, treatment decisions should always be made by a multidisciplinary team and, whenever possible, within a clinical trial.

Frequently asked questions

What is a ctDNA test?

ctDNA consists of small fragments of DNA that tumor cells release into the bloodstream. Tumor-informed ctDNA tests look for these fragments in the blood after surgery to determine whether cancer remains in the body at a level too low to be seen on imaging, in other words, molecular residual disease.

What is the difference between Signatera Genome and the exome-based Signatera test?

The exome-based Signatera test examines only the protein-coding regions of the tumor (the exome) and tracks 16 mutations. Signatera Genome instead sequences the tumor's whole genome and selects from a larger pool of mutations; the aim is to increase sensitivity in tumors that shed little DNA and at early stages.

My ctDNA test was negative. Does that mean my cancer will not come back?

The risk is markedly lower but not zero. In this study, about 99% of persistently negative patients were free of distant recurrence at 36 months; however, brain and locoregional recurrences and new cancers were not counted, and follow-up was short. Routine follow-up visits should continue.

My ctDNA test was positive. What should I do?

A positive result indicates a high risk of recurrence but does not by itself require a change in treatment. Imaging is usually performed first; treatment decisions should be made with your oncologist, in line with guidelines and, whenever possible, within a clinical trial.

Glossary

ctDNA (circulating tumor DNA)

Fragments of DNA released into the blood by tumor cells that carry the tumor's mutations.

Molecular residual disease (MRD)

Cancer cells that remain in the body after treatment and cannot be detected by imaging but whose traces can be found with molecular tests.

Tumor-informed test

A test that first sequences the patient's tumor tissue to identify mutations specific to that patient and then searches the blood only for those mutations.

Exome and whole-genome sequencing

Exome sequencing reads the roughly 1–2% of the genome that codes for proteins, whereas whole-genome sequencing reads the entire genome.

Landmark analysis

An analytical method that groups patients by their test result at a fixed time point (here, 3 months after surgery) and compares what happens afterward, reducing time-related bias.

Sensitivity and specificity

Sensitivity shows what proportion of patients who have a recurrence test positive; specificity shows what proportion of patients without recurrence test negative.