Blood Test Predicts Breast Cancer Relapse
CHICAGO — Relapse after the treatment of primary breast cancer can be reliably predicted well in advance of clinical indications with a circulating free DNA (cfDNA) test, according to a new study.
Results were presented here at the 2014 Annual Meeting of the American Society of Clinical Oncology®.
Tumor-specific mutations were present 6 months after primary treatment in 4 of 5 patients who relapsed after surgery, but in 0 of the 26 patients who did not relapse. The relapses occurred at a median of 8.1 months after surgery.
During a Highlights of the Day session, Michael Gant, MD, from Medical University of Vienna in Austria, described the study as “pivotal”.
“If I had to choose only 1 abstract, I would choose this one,” said Dr. Gant, who was not involved in the study.
“We have shown that tracking tumor-specific mutations in plasma can predict early relapse following the treatment of primary breast cancer,” said study lead author Nicholas C. Turner, MD, academic consultant medical oncologist at the Breakthrough Breast Cancer Research Centre, Institute of Cancer Research, London, United Kingdom.
He noted that the study provides “proof of principle” of the technique.
“Rising mutation abundance in the postsurgical period is a sensitive and specific predictor of relapse, in contrast to a single baseline or postsurgical sample. Patients with detectable mutations or rising abundance may be suitable for therapeutic trials of novel agents,” he added.
Patients with detectable mutations in cfDNA after primary treatment are at high risk for relapse, Dr. Turner explained.
He pointed out that most previous research on cfDNA involved metastatic disease, where as the patients who participated in this study had earlier-stage breast cancer.
Of the 31 patients with primary early-stage breast cancer, 35% had estrogen-receptor-positive, HER2-negative disease. Median follow-up was 17 months. All patients received neoadjuvant chemotherapy prior to surgery, most of which was epirubicin and cyclophosphamide plus accelerated paclitaxel with or without trastuzumab.
Plasma samples were taken before neoadjuvant chemotherapy, immediately after surgery, and every 6 months thereafter.
To identify tumor-specific mutations, the researchers performed targeted next-generation sequencing on the baseline tumor biopsy. At least 1 of the known breast-cancer driver mutations was detectable at baseline in 75% of the patients.
To track the mutations in cfDNA, mutation-specific digital polymerase chain reaction (PCR) assays were developed for each individual tumor mutation. The digital PCR quantified tumor-specific DNA down to the level of a single molecule and was both accurate and reproducible, Dr. Turner explained.
At baseline, there was no difference in tumor-specific mutations between the 5 patients who relapsed during the follow-up period and the 26 who did not.
Immediately after surgery, 4 of the 5 patients who went on to relapse had detectable mutations in cfDNA; there were no cfDNA mutations in the patients who did not relapse.
Over time, the number of cfDNA mutations became amplified in the patients who relapsed.
Dr. Turner described the cases of 2 breast cancer patients with the same PI3CA mutation. At baseline, one patient had 12 copies of cfDNA/mL plasma and the other had 16 copies/mL. After surgery, the patient who did not go on to relapse had no cfDNA, whereas the patient who went on to relapse had 7 copies/mL, indicating a failure to clear micrometastatic disease.
Six months after surgery, the patient who went on to relapse had 198 copies/mL, and relapsed at 8 months. There were steady increases in tumor cfDNA in the plasma prior to relapse.
The other patient continued to have 0 copies/mL and remains disease-free 23.2 months after surgery, Dr. Turner reported.
The fifth patient who relapsed had an isolated relapse in the brain on trastuzumab after achieving a pathologic complete response in the primary breast tumor, and did not have detectable cfDNA mutations at relapse.
“That coincides with recent data showing that cfDNA is difficult to detect in glioma cases,” Dr. Turner said.
He pointed out that the comparison of cfDNA counts in patients who had not relapsed was limited by the short postsurgical follow-up period, and that the ability to detect later relapses was not addressed in the study.
There is a need for some type of “liquid biopsy” to help guide cancer treatment, because repeated tumor biopsies are often not feasible, said Minetta C. Liu, MD, from the Department of Laboratory Medicine and Pathology at the Mayo Clinic in Rochester, Minnesota.
“This study provides preliminary evidence that tracking mutations over time can help predict relapse,” she said. “But will acting on what we find make a difference for our patients?”
Dr. Liu said that an important next stage will be to standardize the processes for monitoring cfDNA.
2014 Annual Meeting of the American Society of Clinical Oncology (ASCO): Abstract 511. Presented May 31, 2014