Ultra-sensitive circulating tumor DNA (ctDNA) monitoring is increasingly reframing how metastatic breast cancer clinical trials are designed and how treatment decisions may be timed, shifting the focus from reacting to radiographic progression toward intervention at the earliest molecular signs of resistance.1
Ultra-Sensitive ctDNA and CSF Liquid Biopsy Enable Earlier Molecular Intervention in Metastatic Breast Cancer
• Ultra-sensitive ctDNA and CSF-based liquid biopsy are shifting monitoring from radiographic response toward earlier, molecularly guided intervention.
• First-generation MRD assays may miss low-level ctDNA shedding—especially in ER-positive disease, raising the risk that “negative” results reflect assay sensitivity rather than true biologic clearance and supporting higher-sensitivity platforms for longitudinal trending.
• ctDNA-guided switching strategies have been validated in PADA-1 and SERENA-6, supporting escalation at emergent ESR1 mutation detection as a framework for adaptive, biology-driven trial design.
Pedram Razavi, MD, PhD, MPH, a breast medical oncologist and Director of Liquid Biopsy & Genomics at Memorial Sloan Kettering Cancer Center in New York, New York, and Milana V. Dolezal, MD, MSci, a clinical associate professor of medicine - oncology at Stanford Medicine Cancer Center in Emeryville, California, described how these tools are being integrated into metastatic disease monitoring strategies and next-generation trial platforms, particularly in settings where imaging may lag behind biologic change.
Razavi noted that first-generation minimal residual disease (MRD) assays may not be sufficiently sensitive for a meaningful subset of patients, especially those with estrogen receptor (ER)–positive disease, where ctDNA can fall below conventional detection thresholds at baseline and more frequently on treatment. In that context, a negative result may reflect assay limitations rather than true biologic clearance, constraining the reliability of ctDNA for longitudinal response interpretation. By contrast, ultra-sensitive platforms may reduce false-negative assessments, strengthen confidence in monitoring, and provide earlier readouts that could function as intermediate end points in clinical trial designs, an especially relevant consideration in adjuvant and neoadjuvant settings where disease-free survival (DFS) events require prolonged follow-up.
The clinical impact of this molecular-first approach is supported by ctDNA-guided trials such as the phase 3 PADA-1 (NCT03079011) and SERENA-6 (NCT04964934) studies, which tested whether therapy alteration at the emergence of an ESR1 mutation detection could improve outcomes compared with waiting for radiographic progression. Razavi emphasized that the broader implication extends beyond a single regimen, asserting that earlier molecular intervention may limit clonal expansion and downstream heterogeneity, establishing a framework for adaptive, biology-driven trial designs.
Dolezal also highlighted how this monitoring strategy intersects with biomarker development in HER2-positive disease, where distinguishing patients who experience early progression vs long-term responders remains a persistent clinical challenge. In parallel, Razavi pointed to cerebrospinal fluid (CSF)–based liquid biopsy as an emerging tool for high-risk central nervous system scenarios, particularly in patients with leptomeningeal disease, where CSF’s low background cell free DNA (cfDNA) may enable more sensitive detection than cytology and support quantitative response assessment across systemic, radiation, and intrathecal therapies.
How could ultra-sensitive ctDNA monitoring shift metastatic breast cancer trial designs toward earlier, molecularly guided interventions?
Technological advances position ctDNA not only as a mechanism-of-resistance test at progression, but as a foundation for adaptive trial paradigms that trigger intervention at the first molecular evidence of actionable resistance.
“We are seeing both in [the] early-stage setting and in [the] metastatic setting that such ultra-sensitive ranges of detection are actually essential for breast cancer. The current first-generation MRD assays that had [detection rates of approximately] 100 parts per million…are not sensitive enough, especially for ER-positive breast cancers,” Razavi expressed. “Even in [the] metastatic setting, more than 20% of the patients have ctDNA levels below the range [of first-generation MRD assays], even at baseline.”
Razavi underscored a practical limitation of many current monitoring strategies in ER-positive disease: when ctDNA shedding is low at baseline or becomes suppressed on therapy, the absence of a detectable signal may reflect assay insensitivity rather than true biologic clearance. In this context, ultra-sensitive platforms are positioned to reduce false-negative assessments, improve confidence in longitudinal trend interpretation, and enable earlier pharmacodynamic readouts that may be leveraged as intermediate end points in trial design, particularly in settings where imaging lags behind molecular evolution and DFS events require prolonged follow-up.
“Having such assays for monitoring disease, both in early and metastatic disease, is critical for future clinical trials. If you have a sensitive assay, you can monitor the response better, both in early- and late-stage settings.” Razavi added. “ctDNA can potentially become a proxy for progression-free survival [PFS] and DFS, especially for DFS [where we could] can cut the time that is needed for adjuvant and neoadjuvant clinical trials to see [responses].”
How is CSF-based liquid biopsy being positioned to improve detection of leptomeningeal disease in metastatic breast cancer?
At the 2025 San Antonio Breast Cancer Symposium, Razavi highlighted work presented by Alexandra M. Miller, MD, of NYU Langone Health, emphasizing CSF as an increasingly practical substrate for liquid biopsy in metastatic breast cancer. Razavi characterized CSF as a comparatively “clean” analyte associated with a lower background cfDNA signal than plasma, which can improve the feasibility of detecting tumor-derived biomarkers in patients with CNS involvement.
“CSF analysis provides a quantitative measurement of response to treatment, both radiation and systemic therapy, and also [intrathecal] therapies,” Razavi expressed. “It’s much more sensitive [than] cytology, and [Miller] showed a lot of data [demonstrating] that it’s very sensitive in identifying leptomeningeal disease. In addition to that, cfDNA analysis and looking for ctDNA in CSF also is highly sensitive. [In fact,] its sensitivity can be higher than a circulating tumor cell analysis. These are some of our patients with worse outcomes, and we need to treat these patients more efficiently. The first step is to detect [response and progression] early.”
How do PADA-1 and SERENA-6 data support early intervention with ctDNA-guided therapy switching?
Despite increasing use of ctDNA for resistance detection, treatment changes in clinical practice routinely deferred until radiographic progression, allowing resistant clones to expand and accumulate additional alterations.² This delay may ultimately compromise the effectiveness of subsequent therapies.
PADA-1and SERENA-6 were designed to test an alternative strategy: switching therapy at the time of molecular resistance detection rather than waiting for clinical progression. At the 2025 ASCO Annual Meeting, data from SERENA-6 demonstrated that this molecularly guided approach translated into a clinically meaningful benefit.³ In patients with a detected ESR1 mutation prior to radiographic progression, switching to camizestrant plus continued CDK4/6 inhibition significantly improved investigator-assessed PFS compared with continued AI plus CDK4/6 inhibition (HR, 0.44; 95% CI, 0.31–0.60; P < .00001). The median PFS was 16.0 months (95% CI, 12.7-18.2) in the camizestrant arm (n = 157) vs 9.2 months (95% CI, 7.2-9.5) in the control arm (n = 158). The 12- and 24-month PFS rates in the camizestrant arm were 60.7% and 29.7%, respectively, compared with respective rates of 33.4% and 5.4% in the control arm.
“The data for SERENA-6 are clear. It shows improvement in PFS when the patients were identified to have ESR1 mutation [and] switched from the backbone of AI to [camizestrant], which is much more effective for ESR1-mutated tumors,” Razavi said. “There was also a trend for time to second progression benefit as well, but the data were not mature. The concept of [switching therapy based on ctDNA prior to clinical progression] is important, and I think we should do this. This is the very first step.”
Beyond ESR1-guided switching, can ctDNA-defined biomarkers such as PIK3CA alterations help distinguish fast progressors from long-term responders in HER2-positive disease and refine treatment sequencing on THP?
In clinical practice, one of the most persistent challenges in HER2-positive metastatic breast cancer is distinguishing patients who will experience early progression from those who derive prolonged benefit from first-line therapy. As Dolezal noted, clinicians routinely observe a striking divergence in outcome with some patients experiencing rapid progression despite standard treatment, while others remain stable for many years on trastuzumab (Herceptin)- and pertuzumab (Perjeta)-based maintenance.
“Clinically, we want to have good biomarkers to understand: who are these fast progressors at the beginning, vs [which patients] are out here at the long tail of these [survival] curves? In our practices, we all have patients with stage IV disease who’ve 10 years, which is pretty remarkable in terms of just staying on an trastuzumab/pertuzumab maintenance regimen,” Dolezal explained. “So how do we know who these patients are?”
In the phase 3 CLEOPATRA study (NCT00567190), exploratory biomarker analyses suggested that PIK3CA mutations were enriched among patients with less favorable outcomes, raising the possibility that PI3K pathway alteration status may serve as a marker of more aggressive biology in a subset treated with a taxane plus trastuzumab and pertuzumab (THP).4 For clinicians, this heterogeneity has direct practice implications, particularly as some long-term responders remain stable for years on trastuzumab- and pertuzumab-based maintenance prompting pragmatic questions about whether and when treatment holidays are appropriate, whether indefinite HER2 maintenance is necessary for all patients, and how molecular tools such as ctDNA might ultimately help identify candidates for safe de-escalation.¹
“Some of the [patients with the earliest progression on the CLEOPATRA study] did seem to have PIK3CA mutations as a marker of early progression on THP. In terms of our [historical] standard of care, with CLEOPATRA as being the first-line regimen, and then in terms of these long-term responders, can we give treatment holidays? I certainly do that clinically,” Dolezal said. “Do patients need to be on some sort of HER2-directed maintenance forever? How can we pick these patients out?”
In contrast to CLEOPATRA, where PI3K pathway alterations were explored as a potential marker of earlier progression on THP, more contemporary evidence has suggested that this signal may not translate across newer first-line strategies. In the phase 3 DESTINY-Breast09 (NCT04784715), investigator-assessed PFS was reported to be similar in patients in the experimental arm treated with fam-trastuzumab deruxtecan-nxki (T-DXd; Enhertu) plus trastuzumab, irrespective of PIK3CA mutation status. In the T-DXd/pertuxumab arm, the investigator-assessed median PFS was 40.7 months (95% CI, 38.0-not calculable [NC]) for those without PIK3CA mutations (n = 266) vs 36.0 months (95% CI, 29.7-NC) in those harboring PIK3CA mutations (n = 116).
“Different than in CLEOPATRA, where we thought [we] saw PIK3CA mutations maybe being a biomarker for early progression, when DESTINY-Breast09 evaluated this, the PFS was [similar with T-DXd plus pertuzumab]. That’s really an important differentiator to understand in terms of patient selection,” Dolezal explained.
References
- Bidard FC, Mayer EL, Park YH, et al. First-Line Camizestrant for Emerging ESR1-Mutated Advanced Breast Cancer. N Engl J Med. 2025;393(6):569-580. doi:10.1056/NEJMoa2502929
- Miller AM. Utilizing liquid biopsy for detection and monitoring CNS metastases. Presented at: 2025 San Antonio Breast Cancer Symposium; December 9-12, 2025; San Antonio, TX.
- Turner N, Mayer E, Park YH, et al. Camizestrant + CDK4/6 inhibitor (CDK4/6i) for the treatment of emergent ESR1 mutations during first-line (1L) endocrine-based therapy (ET) and ahead of disease progression in patients (pts) with HR+/HER2– advanced breast cancer (ABC): Phase 3, double-blind ctDNA-guided SERENA-6 trial. J Clin Oncol. 2025;43(suppl 17):LBA4. doi:10.1200/JCO.2025.43.17_suppl.LBA4
- Baselga J, Cortés J, Im SA, et al. Biomarker analyses in CLEOPATRA: a phase III, placebo-controlled study of pertuzumab in HER2-positive, first-line metastatic breast cancer. J Clin Oncol. 2014;32(33):3753-3761. doi:10.1200/JCO.2013.54.5384
- Tolaney SM, Jiang Z, Zhang Q, et al. Trastuzumab deruxtecan (T-DXd) + pertuzumab vs taxane + trastuzumab + pertuzumab (THP) for first-line (1L) treatment of patients (pts) with human epidermal growth factor receptor 2–positive (HER2+) advanced/metastatic breast cancer: interim results from DESTINY-Breast09. J Clin Oncol. 2025;43(suppl 17):LBA1008. doi:10.1200/JCO.2025.43.17_suppl.LBA1008
