Circulating tumor DNA shows prognostic power in early breast cancer, but clinical utility remains unproven

Circulating tumor DNA, or ctDNA, can flag breast cancer recurrence months before scans or symptoms, yet there is still no proof that acting on those results helps patients live longer, according to a review published May 28 in JAMA Oncology.

The analysis, led by Ilana Schlam, M.D., M.P.H., from the Dana-Farber Cancer Institute and Harvard Medical School, surveyed the fast-growing evidence base for ctDNA-based testing in early-stage breast cancer. Schlam and her colleagues drew a sharp line between two concepts that payers and clinicians often conflate: prognostic validity, which these assays have clearly established, and clinical utility, which they define as improving patient outcomes. 

Schlam and her colleagues explain that the term “minimal residual disease (MRD)” refers to the presence of microscopic disease that can’t be detected by traditional imaging. They say that the ctDNA assays are the most promising way to detect MRD in patients with solid tumors.

The prognostic signal was strong and consistent for the ctDNA tests, they say. After curative-intent treatment, detectable ctDNA was strongly associated with future distant recurrence, and its emergence during surveillance reliably preceded clinically apparent metastatic disease. In one study cited in the review, patients who were ctDNA-positive one year after surgery faced a recurrence hazard roughly 21 times that of ctDNA-negative patients, with a median lead time to recurrence of nearly 19 months. Negative predictive values for staying recurrence-free often exceeded 95% in several cohorts.

The review distinguished two assay designs. Tumor-informed, or bespoke, tests sequence a patient's tumor first to build a personalized panel, maximizing specificity and achieving limits of detection in the one- to three-parts-per-million range. Tumor-agnostic tests skip the tissue step, offering faster turnaround and lower cost but lower sensitivity and specificity. The authors noted the bespoke approach has tradeoffs that matter operationally. Roughly 20% of patients may lack sufficient tissue for analysis, and building a custom assay lengthens turnaround time.

For health plans, the reimbursement picture is uneven. The review's assay table showed several tests, including Natera's Signatera, Personalis' NeXT Personal, NeoGenomics' RaDaR and SAGA's Pathlight, carrying specific Medicare coverage indications tied to stage, subtype and clinical setting, while others such as Guardant Reveal, FoundationOne Tracker and Exact Sciences' Oncodetect had no listed early-breast-cancer coverage as of mid-March. Coverage varies with stage, hormone-receptor status and whether testing occurs in the neoadjuvant, adjuvant or surveillance window, complicating any uniform utilization-management approach.

The assays are not interchangeable, and head-to-head comparisons are lacking. The optimal timing and frequency of testing also remain undefined, leaving open questions about how often to draw blood and at what cost

The decisive evidence was still being generated. The review cataloged numerous ongoing interventional trials, including phase 3 efforts such as SURVIVE, TREAT-ctDNA and DARE, testing whether ctDNA-guided escalation or de-escalation of therapy actually changes outcomes. Early experience underscored the difficulty such that some trials struggled to enroll because ctDNA-positive patients proved scarcer than expected, and one de-escalation effort was discontinued early.

The authors' bottom line is cautionary for benefit designers weighing coverage expansion. Until trials confirm that ctDNA-guided treatment improves survival rather than simply predicting it, these tests remain prognostic tools in search of proven utility. The review urges clinicians to weigh the strengths, limitations and patient preferences before incorporating ctDNA into routine care.

“MRD status is a strong and consistent prognostic marker, but clinical utility must be established before routine use,” they wrote in their JAMA Oncology. Existing data support trials testing ctDNA-guided escalation or de-escalation strategies, they said. After surgery, MRD positivity identifies a population at high risk of recurrence and represents a rational target for escalation trials, they said. But they caution that although the measurement of MRD in the surveillance setting holds the promise of early detection of “molecular relapse” that won’t be discovered by traditional imaging, there are problems, such as the “high numbers needed to screen, uncertain lead-time benefit, and unanswered questions about optimal timing, overlap with scan-detectable metastatic disease, frequency, and the potential harm of early intervention.”