From the Journals

ctDNA may predict relapse risk in early breast cancer

View on the News

Clinical utility still unknown

Although a strength of the study is the inclusion of all subtypes of breast cancer, Garcia-Murillas et al. found that the ability to detect circulating tumor DNA (ctDNA) was likely influenced by biologic factors, including receptor subtypes. The study had a median follow-up of 36.3 months (in the combined cohorts); however, because the risk of relapse for luminal estrogen receptor–positive breast cancers is known to persist for decades, these data cannot be applied to late recurrences, which are largely derived from luminal estrogen receptor–positive disease. Longer-term follow-up with serial sampling of ctDNA will be required to demonstrate validation for this patient population.

As addressed by the authors, the clinical utility for ctDNA detection in early-stage breast cancer is still unknown. Proof of clinical utility can be accomplished through prospective, multi-institutional trials randomizing ctDNA-positive patients to therapy versus control and demonstrating reductions in disease-free and overall survival. The use of real-time testing and rapid turnaround time may prove to be challenging if we are to implement ctDNA testing as an integral biomarker for clinical decision making. However, the study by Garcia-Murillas et al. is a major step forward in reaching this goal because the results suggest the feasibility and clinical validation of ctDNA for patients with early-stage disease.

Remarks from Swathi Karthikeyan, MS, of Johns Hopkins University, Baltimore, and Ben Ho Park, MD, PhD, of Johns Hopkins and Vanderbilt University, Nashville, Tenn., are condensed and adapted from an editorial accompanying the study by Garcia-Murillas et al. Dr. Park reported royalties from Horizon Discovery, serving as a scientific advisory board member for Loxo Oncology, having an ownership interest in Loxo Oncology, serving as a recent paid consultant for Foundation Medicine, Jackson Laboratories, H3 Biomedicine, Casdin Capital, Roche, Eli Lilly, and Astra Zeneca, and having research contracts with Abbvie, Foundation Medicine, and Pfizer. No other disclosures were reported.



Following therapy with curative intent for early-stage primary breast care, the presence of circulating tumor DNA may identify those patients at high risk for relapse, investigators reported.

Among 101 women treated for early-stage breast cancer and followed for a median of nearly 3 years, detection of circulating tumor DNA (ctDNA) during follow-up was associated with a 2,400% increased risk for relapse, and detection of ctDNA at diagnosis but before treatment was associated with a nearly 500% risk, wrote Isaac Garcia-Murillas, PhD, of the Institute of Cancer Research, London, and colleagues.

“Prospective clinical trials are now required to assess whether detection of ctDNA can improve outcomes in patients, and a phase 2 interventional trial in TNBC [triple-negative breast cancer] has been initiated. This trial may develop a new treatment paradigm for treating breast cancer, in which treatment is initiated at molecular relapse without waiting for symptomatic incurable metastatic disease to develop,” they wrote in JAMA Oncology.

The investigators conducted a prospective, multicenter validation study of samples collected from women with early-stage breast cancer irrespective of hormone-receptor or HER2 status. The patients were scheduled for neoadjuvant chemotherapy followed by surgery, or surgery followed by adjuvant therapy.

Of 170 women recruited, 101 had tumors with identified mutations and were included in the main cohort. The investigators also conducted secondary analyses with patients in this cohort plus an additional 43 women who had participated in a previous proof-of-principle study.

They first sequenced tumor DNA to identify somatic mutations in primary tumors that could then be tracked using a breast cancer driver gene panel. For each sample, a personalized digital polymerase chain reaction (dPCR) assay was created to identify the mutations in plasma samples.

The plasma samples were collected every 3 months for the first year of follow-up, then every 6 months thereafter.

In the main cohort, the median age was 54 years, and the median follow-up was 35.5 months. The investigators found that, for the primary endpoint of relapse-free survival, ctDNA was associated with a hazard ratio for relapse of 25.2 (P less than .001). Detection of ctDNA in samples taken at the time of diagnosis was also associated with worse relapse-free survival, with an HR of 5.8 (P = .01).

In a secondary analysis, ctDNA detection preceded clinical relapse by a median of 10.7 months, and was associated with relapse in all breast cancer subtypes.

Of 29 patients who experienced a relapse, 22 of 23 with extracranial distant metastatic relapse had prior ctDNA detection.

The remaining six patients experienced relapse without ctDNA detection either before or at the time of relapse. Each of these six patients had a relapse at a single site: in the brain in three patients (with no extracranial relapses), in the ovaries in one patient, and solitary locoregional relapses in two patients.

The investigators acknowledged that the results “demonstrate clinical validity for ctDNA mutation tracking with dPCR but do not demonstrate clinical utility. Without evidence that mutation tracking can improve patient outcome, our results should not be recommended yet for routine clinical practice.”

The study was funded by Breast Cancer Now, Le Cure, and National Institute for Health Research funding to the Biomedical Research Centre at the Royal Marsden Hospital and the Institute of Cancer Research. Dr. Garcia-Murillas had no disclosures. Multiple coauthors reported grants and/or fees from various pharmaceutical companies.

SOURCE: Garcia-Murillias I et al. JAMA Oncol. 2019 Aug 1. doi: 10.1001/jamaoncol.2019.1838.

Next Article: