Circulating Tumor DNA Markers for Early Progression on Fulvestrant With or Without Palbociclib in ER+ Advanced Breast Cancer

Ben O'Leary, MBBS, PhD; Rosalind J. Cutts, PhD; Xin Huang, PhD; Sarah Hrebien, BSc; Yuan Liu , PhD; Fabrice André, MD, PhD; Sibylle Loibl, MD, PhD; Sherene Loi, MBBS, PhD; Isaac Garcia-Murillas, PhD; Massimo Cristofanilli, MD; Cynthia Huang Bartlett, MD, PhD; Nicholas C. Turner, MBBS, PhD


J Natl Cancer Inst. 2021;113(3):309-317. 

In This Article

Abstract and Introduction


Background: There are no established molecular biomarkers for patients with breast cancer receiving combination endocrine and CDK4/6 inhibitor (CDK4/6i). We aimed to determine whether genomic markers in circulating tumor DNA (ctDNA) can identify patients at higher risk of early progression on fulvestrant therapy with or without palbociclib, a CDK4/6i.

Methods: PALOMA-3 was a phase III, multicenter, double-blind randomized controlled trial of palbociclib plus fulvestrant (n = 347) vs placebo plus fulvestrant (n = 174) in patients with endocrine-pretreated estrogen receptor–positive (ER+) breast cancer. Pretreatment plasma samples from 459 patients were analyzed for mutations in 17 genes, copy number in 14 genes, and circulating tumor fraction. Progression-free survival (PFS) was compared in patients with circulating tumor fraction above or below a prespecified cutoff of 10% and with or without a specific genomic alteration. All statistical tests were 2-sided.

Results: Patients with high ctDNA fraction had worse PFS on both palbociclib plus fulvestrant (hazard ratio [HR] = 1.62, 95% confidence interval [CI] = 1.17 to 2.24; P = .004) and placebo plus fulvestrant (HR = 1.77, 95% CI = 1.21 to 2.59; P = .004). In multivariable analysis, high-circulating tumor fraction was associated with worse PFS (HR = 1.20 per 10% increase in tumor fraction, 95% CI = 1.09 to 1.32; P < .001), as was TP53 mutation (HR = 1.84, 95% CI = 1.27 to 2.65; P = .001) and FGFR1 amplification (HR = 2.91, 95% CI = 1.61 to 5.25; P < .001). No interaction with treatment randomization was observed.

Conclusions: Pretreatment ctDNA identified a group of high-risk patients with poor clinical outcome despite the addition of CDK4/6 inhibition. These patients might benefit from inclusion in future trials of escalating treatment, with therapies that may be active in these genomic contexts.


CDK4/6 inhibitors (CDK4/6i) now play a key role in the treatment of advanced, estrogen receptor–positive (ER+) breast cancers,[1] with established efficacy in combination with endocrine therapy in both first- and second-line treatment.[2–8] However, a substantial proportion of patients progress early on treatment, and there is a clinical need to identify patients at risk of early progression.

There are a number of established molecular markers associated with poor outcome in early ER+ breast cancer, most notably the risk classifiers based on gene expression assessed in tumor biopsies, which are now routinely used to augment clinical decision making.[9] Genomic markers other than HER2 amplification associated with poorer outcome in primary disease include mutations in TP53,[10,11] amplifications in FGFR1,[12] which may contribute to endocrine therapy resistance,[13] and amplification of MYC.[14] Less is known of the associations between common genomic aberrations in advanced ER+ breast cancer and clinical outcome, particularly in the updated therapeutic landscape that includes combination CDK4/6i treatments.

Recent work has identified a number of potential genomic mechanisms of resistance to CDK4/6i, notably amplification of CCNE1, mutations in FAT1, CDK6 overexpression, and loss of RB1,[15,16] with emerging data for immune signatures and other oncogenic signalling.[17,18] Of these, clinical data support acquisition of RB1 mutations in a minority of cancers progressing on CDK4/6i,[19,20] with preexisting loss of functional RB1 associated with poor prognosis on CDK4/6i therapy. Loss of FAT1 was also associated with poor outcome on CDK4/6i therapy,[21] although inactivating mutations in FAT1 are rare in advanced ER+ breast cancer. We have shown previously that mutations in PIK3CA and ESR1 in advanced ER+ breast cancer previously treated with endocrine therapy do not predict response to palbociclib.[22]

Circulating tumor DNA (ctDNA) is found in the plasma of a substantial majority of patients with advanced cancer and presents a source of cancer DNA for noninvasive analysis of tumor somatic genetic features. In addition, circulating tumor fraction, the fraction of plasma DNA that is derived from the tumor, may be a biological marker that reports on both tumor bulk and tumor aggressiveness[23] and is associated with poorer clinical outcome in triple-negative breast cancer.[24]

In conducting this analysis, we hypothesized that genomic aberrations identified at baseline, including mutations, copy number, and circulating tumor fraction, could be predictive or prognostic of clinical outcome for patients with advanced ER+ breast cancer receiving fulvestrant with or without palbociclib. We investigated this using a multimodal ctDNA sequencing analysis of plasma DNA from the PALOMA-3 trial.