Study: A new blood test may help predict early-stage breast cancer patients at highest risk for recurrence

IN-DEPTH REVIEW OF RESEARCH

Study background:

Ninety-five percent of women with early-stage breast cancer have no indication of metastases at the time of diagnosis. Many of these patients will not experience a relapse.  The 5-year survival rate for women with local breast cancer is 98% and it is 88% for women with regional breast cancer.

Relapse or recurrence is attributed to small groups of cancer cells that remain after treatment (called “molecular residual disease” or “MRD”). There are too few of these cells to observe easily. When these remaining cancer cells may migrate to other parts of the body they are called metastases. Early detection and treatment is optimal for patients who relapse.

But what is the best way to predict who will relapse and who will not? Those remaining tumor cells will sometimes lead to tumor DNA in the blood of patients, called circulating tumor DNA or ctDNA. Tumor cells in patients with ctDNA in their blood may subsequently grow until they are clinically detectable and/or causing symptoms—a recurrence of their primary cancer.

This study’s researchers wanted to know whether ctDNA detection accurately predicts the chance of recurrence in breast cancer patients. The key to using ctDNA is to find a mutation that has occurred only in the tumor during its growth (not an inherited mutation). To do this, DNA from the primary tumor is sequenced to find mutations that differ from the patient's non-tumor DNA. These mutations act as a signature of the tumor cells.

Researchers of this study wanted to know:

Whether ctDNA in a patient's blood indicates that they are at greater risk of relapse. The underlying question is whether ctDNA detection might be a useful tool clinically for identify patients at increased risk of relapse.

Populations looked at in this study:

This study included 170 women with a mean age of 54 years from 5 UK hospitals.

All patients had primary breast cancer with no evidence of metastases by standard monitoring methods.

Study design:

This was a prospective multicenter study in the United Kingdom from November 2011 to October 2016. All participants had early-stage breast cancer. In 140 participants, early-stage breast cancer was treated neoadjuvantly (before surgery); 30 other participants had surgery and adjuvant treatment (after surgery).

To determine if ctDNA was a useful tool to predict relapse, blood samples from each participant were taken before and after their initial treatment. At the time of surgery, each participant's primary tumor DNA was sequenced to identify mutations that uniquely identified the tumor cells. Blood samples from patients were then collected every 3 months for the first year and 6 months in subsequent years for up to 5 years. Each sample was tested for the presence of tumor-specific DNA (ctDNA), and data was collected about whether the patient had experienced a relapse.

To have sufficient numbers for a comparison of different cancer subtypes, data analysis of the group of 101 participants was combined with data from the initial proof-of-principle study of 43 patients, for a total group of 144 patients. Among these 144 patients, 29 relapsed during the study. 

Study findings:  

Of 170 patients participating in this study, 101 had a tumor-specific mutation:

• All 101 patients had at least one tumor-specific mutation:
  • 78 patients had 1 tumor-specific mutation.
  • 23 patients had multiple tumor-specific mutations.
• 69 patients had no tumor-specific mutations identified.

Blood samples and personal cancer history were collected over an average of 35 months from the 101 participants with an identified tumor mutation.

Key findings include:

1. Detection of ctDNA in the blood was a strong indicator of relapse among these participants. Patients with ctDNA that was detected after initial treatment were 25 times more likely to relapse than patients who did not have ctDNA detected after initial treatment.

Among the 101 participants in this study, the presence of ctDNA in the blood was associated with subsequent relapse.

• 16 of 101 (nearly 16%) patients had ctDNA detected in their blood
  • 9 relapsed during the study.
  • 7 had not relapsed at the time the study was published.
  • On average, there was 38 months between initial treatment/surgery and ctDNA detection.
• 12 (nearly 12%) patients relapsed during this study of 101 patients.
  • 9 had ctDNA detected prior to relapse (75%).    
  • 3 did not have ctDNA detected (25%).

2. ctDNA was similarly useful as a predictor of relapse for all breast cancer subtypes examined (HER-positive, HER-negative, triple-negative cancers).

To have sufficient numbers for comparison of different cancer subtypes, data from the group of 101 participants were combined with data from the initial proof-of-principle study of 43 patients for a total group of 144 patients.

Among these 144 patients:

• 29 of 144 patients relapsed.
  • 23 of 29 patients (79%) had ctDNA detected in their blood prior to their relapse.
  • 6 of 29 (21%) patients did not have ctDNA detected prior to their relapse.
    • Among the 6 patients without ctDNA detected, all had one site of relapse:
      • 3 at the brain only.
      • 1 at the ovary only.
      • 2 at one site near their primary breast cancer site.
• ctDNA detection occurred 10.7 months before clinical relapse for all breast cancer types.

3. Detection of ctDNA before surgery or other treatment was also correlated with relapse.

• ctDNA was detected in 41 of 80 (51%) patients in blood samples taken before treatment.
• Patients with ctDNA detected before treatment were nearly 6 times more likely to relapse than patients with no ctDNA detected before diagnosis.
• Detection at diagnosis was associated with larger tumor size and higher grade tumors.

4. ctDNA was more likely to be detected for metastases in locations in the body other than the brain compared to metastases in the brain. In this study, relapse sites that were only in the brain were not detected by ctDNA.

• ctDNA was detected in 22 of 23 patients (96%) who relapsed when the relapse site was not the brain.
• ctDNA was detected in 1 of 6 patients (17%) who relapsed when the relapse site was in the brain.

5. While ctDNA appears to be a good predictor of patients who eventually experienced a clinical relapse, some patients who did not have ctDNA detected relapsed during this study—lacking ctDNA does not mean that relapse will not happen.

• Among the group of 101 patients, 3 of 9 (33%) patients who relapsed had no ctDNA detected.
• Among the combined group of 144 patients, 6 of 29 (21%) patients who relapsed had no ctDNA detected.

Limitations:

Because this study was conducted in only one geographical area—the UK—results may or may not be generalizable to other regions and populations.

Tumor-specific mutations were identified in 101 of 170 participants. For 69 participants, the presence or absence of ctDNA were not tracked; in a clinical setting, one would want to identify mutations for a given patient's tumor.

Although this study indicates that ctDNA detection is a promising tool for predicting who will need more monitoring or treatment, the numbers of participants represented are small. Additionally, two groups of participants, those from this study and those from original proof-of-principle study, showed different rates of relapse. While combining these groups was planned, this may represent unknown biases in the two groups. A larger trial with a longer follow-up period is needed to assess whether this technique is truly useful in a clinical setting and whether it helps improve outcomes.

Among patients with brain metastases, none were detected by ctDNA in this study. This suggests that additional tools or more refined ctDNA detection will be required to monitor for relapse of this type.

Conclusions:

While promising, this tool is not yet validated for clinical use. If it becomes validated, the presence of ctDNA might be useful to determine earlier treatment approaches prior to symptoms of cancer relapse. However, as the study researchers stated, "Our results demonstrate clinical validity for ctDNA mutation tracking...but do not demonstrate clinical utility. Without evidence that mutation tracking can improve patient

outcome[s], our results should not be recommended yet for routine clinical practice."

A phase II clinical trial for triple-negative breast cancer is underway to assess whether use of ctDNA in a clinical setting will improve patient outcomes.

Share your thoughts on this XRAYS article by taking our brief survey.