Can ctDNA Testing in Breast Cancer Detect Recurrence Before It Is Evident Pathologically?

A Blood Test Instead of the Pathology Report? Reading Residual Disease After Breast Cancer Surgery With ctDNA

In triple-negative and HER2-positive breast cancer, complete disappearance of the tumor after treatment given before surgery (neoadjuvant therapy), known as pathological complete response, is considered the strongest sign of a good outcome. Yet a substantial share of patients who do not achieve a complete response also remain disease-free for many years. The PREDICT-DNA study, published in the Journal of Clinical Oncology in March 2026, showed that a test able to measure circulating tumor DNA (ctDNA) at very low levels separates these patients better than the pathology report. Nearly all patients with a positive ctDNA result within the first year after surgery had a recurrence, whereas 94% of triple-negative patients with a negative result were disease-free at 5 years. We have previously covered post-surgery ctDNA testing and ctDNA-guided treatment in metastatic breast cancer.

Why does this matter?

In stage II–III triple-negative and HER2-positive breast cancer, the standard approach is chemotherapy first (plus HER2-targeted therapy in HER2-positive disease, and now immunotherapy in triple-negative disease), followed by surgery. If the tumor has completely disappeared at surgery (pathological complete response), the long-term outlook is very good. If not, most patients receive additional treatment after surgery: in triple-negative disease (where treatment sequencing is debated even for small tumors) capecitabine or, in BRCA carriers, olaparib; in HER2-positive disease, T-DM1.

The problem is this: some patients without a complete response may in fact have no cancer cells left in the body, and they experience the side effects of additional treatment for no benefit. Conversely, the disease returns in some patients who did achieve a complete response. The pathology report shows only the removed breast tissue and the axillary lymph nodes; it cannot see microscopic disease left elsewhere in the body (minimal residual disease). Circulating tumor DNA aims to trace exactly this invisible disease.

What is a ctDNA test used for in breast cancer?

A ctDNA test measures tumor-derived DNA fragments circulating in the blood to detect microscopic cancer remaining in the body. According to PREDICT-DNA, in triple-negative and HER2-positive breast cancer, post-treatment ctDNA positivity signals a high risk of recurrence, while a negative result signals a very good outcome. However, the benefit of changing treatment based on the result has not yet been proven.

How was PREDICT-DNA conducted?

Results

1. ctDNA clearance does not mean complete response

Before treatment, the ultrasensitive assay detected ctDNA in 93% of patients (99% in triple-negative, 86% in HER2-positive disease). Pathological complete response was achieved in 41% of patients. All patients with a complete response were ctDNA-negative before surgery; however, 40 of 100 ctDNA-negative patients still had residual tumor at surgery. The negative predictive value was only 60%, falling short of the 90% primary endpoint. This clearly shows that one cannot say "if no tumor DNA remains in the blood, surgery is unnecessary."

2. Preoperative ctDNA is a better risk marker than pathological response

In the prespecified secondary endpoint, preoperative ctDNA positivity significantly increased the risk of recurrence in triple-negative patients. By contrast, separation by whether or not a pathological complete response was achieved was not statistically significant.

A striking finding: half of the triple-negative patients without a complete response were ctDNA-negative before surgery, and their 3-year recurrence-free survival (94%) was very close to that of patients with a complete response (96%). Moreover, none of these patients had received postoperative pembrolizumab. In a model evaluating pCR and ctDNA together, only ctDNA remained an independent risk marker (HR 12). A similar trend was seen in HER2-positive patients (3-year recurrence-free survival 62% with positive vs 93% with negative ctDNA), but this analysis was not prespecified and the number of recurrences was small.

3. Postoperative ctDNA: near-complete separation

Of the 151 patients who provided a blood sample within the first year after surgery, 90% were ctDNA-negative, and 97% of them were recurrence-free at 5 years. All 13 of the 15 ctDNA-positive patients who were followed for at least 3 years had a recurrence. ctDNA positivity was detected a median of 7.9 months before clinical recurrence in triple-negative disease and 5.9 months before in HER2-positive disease. About two-thirds of these positive results were at levels detectable only with an ultrasensitive assay (<100 ppm).

When preoperative and postoperative measurements were evaluated together in triple-negative patients: of 12 patients who were positive before surgery and turned negative after surgery, only 2 had a recurrence; all 6 patients who were positive at both time points had a recurrence (median 15.8 months).

How should the results be interpreted?

Strengths

• Prospective, multicenter design (93% of tumor samples from core needle biopsies); those performing the ctDNA analysis were blinded to clinical data.
• Ultrasensitive assay; tumor DNA was detectable in 93% of patients before treatment, and a substantial share of positive results were at levels that less sensitive assays would have missed.
• Approximately 5 years of follow-up and repeated measurements at different time points.
• Preliminary results from the French SCANDARE study, which used the same assay, are similar.

Caveats

• Primary endpoint not met: The main question was whether ctDNA predicts pathological complete response; the answer was "no." The strongest findings come from secondary and exploratory analyses.
• Small groups and wide confidence intervals: The confidence interval for the HR of 128 ranges from 15 to 1,083. Because the number of recurrences was small, the magnitude of risk is imprecise; the direction is clear, the size uncertain.
• Does not reflect current treatment: Most patients received treatment without immunotherapy. Whether the results would be the same with the KEYNOTE-522 regimen, now standard in triple-negative breast cancer, is unknown.
• Clinical utility not shown: The study shows that the test stratifies risk accurately (clinical validity), but it has not been shown that changing treatment based on ctDNA improves survival (clinical utility). The journal's editorialist, Kathy Miller, also cautioned that stopping effective treatment in a ctDNA-negative patient, or giving ineffective but toxic treatment to a positive patient, could cause harm.
• Conflict of interest (what to watch for when reading research): The test manufacturer (Personalis) partly funded the study, and company employees are among the authors.
• Residual tumor burden not measured: Because the residual cancer burden (RCB) score was not determined in all patients on pathology, it could not be compared with ctDNA.

What does this mean in practice?

• ctDNA does not replace surgery: A negative ctDNA result after neoadjuvant therapy does not reliably show that the tumor has completely disappeared. Surgery remains the standard.
• Risk stratification is very strong: A positive ctDNA result after surgery signals disease that is not yet visible on imaging but most likely remains in the body. A negative result, especially in triple-negative patients, points to a very good outcome.
• Too early to change treatment: It has not been shown that giving additional treatment to ctDNA-positive patients or de-escalating treatment in negative patients improves survival. ctDNA-positive patients are suitable candidates for clinical trials targeting molecular relapse (for example, ASPRIA, NCT04434040).
• The benefit of detecting recurrence early is uncertain: ctDNA is far more sensitive than classic tumor markers such as CA 15-3 and can reveal recurrence months earlier; however, it has not been proven in breast cancer that treating metastatic disease earlier prolongs survival.
• A tool for future trials: De-escalation trials in ctDNA-negative patients and escalation trials in positive patients can be designed more efficiently, with fewer patients.

Practical advice for patients and families

• If you are receiving treatment before surgery, discuss with your oncologist how the pathology result at surgery (whether or not there is a complete response) will affect the rest of your treatment.
• A ctDNA test is not yet a routine test; if you are considering one, ask beforehand whether the result would change your treatment.
• Not achieving a pathological complete response does not by itself mean a poor outcome; the need for treatment after surgery varies from person to person, and in this study a substantial share of patients without a complete response did very well.
• Do not stop recommended postoperative treatments (such as capecitabine, T-DM1 or immunotherapy) on your own based on a ctDNA result; manage immunotherapy side effects together with your care team.
• If needed, get a second opinion; exercise and rehabilitation are also important during treatment.

Frequently asked questions

What is a ctDNA test in breast cancer?

It is a blood test that measures tumor-derived DNA fragments circulating in the blood. In tumor-informed assays, mutations in the tumor tissue are identified first, and these mutations are then tracked in the blood. The aim is to detect residual disease that cannot be seen on imaging.

If ctDNA is negative, is surgery still needed?

Yes. In PREDICT-DNA, 40% of patients with a negative preoperative ctDNA result still had residual tumor at surgery. ctDNA cannot replace surgery.

What happens if ctDNA is positive after surgery?

In this study, all patients who were ctDNA-positive after surgery and followed for at least 3 years had a recurrence. However, it has not yet been shown that giving these patients additional treatment changes the outcome; clinical trials are the most appropriate option.

Is the ctDNA test available in Türkiye, and does SGK cover it?

Tumor-informed ultrasensitive ctDNA assays are not routinely performed in Türkiye; samples are usually sent abroad, and the tests are not covered by SGK.

Glossary

ctDNA (circulating tumor DNA)

DNA fragments released from tumor cells into the bloodstream.

Minimal residual disease (MRD)

Microscopic cancer that remains in the body after treatment and cannot be seen on imaging.

Neoadjuvant therapy

Drug treatment given before surgery.

Pathological complete response (pCR)

Absence of viable cancer cells in the breast tissue and axillary lymph nodes removed at surgery.

Negative predictive value

The probability that a person with a negative test result truly does not have the disease.

ppm (parts per million)

The number of tumor DNA fragments per one million normal DNA fragments in the blood.

Invasive disease-free survival

The time elapsed without invasive breast cancer recurrence, a new cancer or loss of life.