Diagnostic utility of ESR1 mutation detection in liquid biopsy of metastatic breast cancer patients

About 20–30% of hormone receptor positive breast cancers (BCs) show local or distant recurrence after endocrine treatment [1]. Many of these metastatic BC cases, treated with tamoxifen or aromatase inhibitors (AI), acquire a secondary resistance against hormone blockade [2]. Ligand-independent, constitutive estrogen receptor α (ER) activity, caused by the acquisition of ESR1 gene mutations in the ligand-binding domain, represents a resistance mechanism of tumor cells against anti-estrogenic therapy [3]. In case of early ESR1 mutation detection, patients can benefit from switching the treatment to fulvestrant. Progression-free survival in a recent phase 3 trial was 11.9 months in the fulvestrant and palbociclib group versus 5.7 months in the AI and palbociclib group [4]. Furthermore, with elacestrant, a new oral ER degrader is available. In heavily pre-treated patients, 12-month PFS of ESR1 mutated cases was 22.3% with elacestrant versus 9.4% in the standard-of-care group [5]. A noninvasive alternative to tissue biopsies of metastatic sites is the mutation detection in circulating free DNA (cfDNA) from tumor cells extracted from blood plasma samples (liquid biopsy). Here we present real-world data from ESR1 mutation testing in external liquid biopsies in molecular pathology routine.

We analyzed liquid biopsy samples from in total 162 patients with metastatic BC to identify mutations in ESR1, PIK3CA, ERBB2, and TP53. The mean cfDNA amount extracted from 5 ml blood plasma was 124.6 ng; the median was 74.0 ng (details for all samples in supplementary Table 2). There was a positive correlation of mutation calls with a higher amount of cfDNA (supplementary Figure 1). In 88 patients (54.3%), we detected a total of 162 gene mutations (Fig. 1, supplementary Table 2). In 30 patients (18.5%), we detected co-mutations in two or three of the investigated genes (Fig. 2). ESR1 mutations could be detected in 42 cases (25.9%). The lowest and the highest detected allele frequency (AF) of an ESR1 hotspot mutation was 0.2% and 70.9%, respectively (both p.E380Q, case no. 61 and 75 in supplementary Table 2). Thirteen cases showed multiple ESR1 mutations, indicating sub-clonal diversity (Fig. 3). The recently described ESR1 p.F404L/I/V fulvestrant resistance mutations [8] are not present in our cohort. PIK3CA mutations occurred most frequently (56 cases, 34.6%), among them 16 cases with a combined PIK3CA/ESR1 mutation. In six cases (3.7%), ERBB2 mutations in exon 8, 12, 19, or 20 were detectable. TP53 mutations were present in 17 cases (10.5%), mostly co-mutated with ERBB2, ESR1, or PIK3CA. The mean turnaround time in the molecular laboratory was 5.9 working days (range 3 to 11 days), which is acceptable for NGS panel sequencing including data analysis.

Bild vergrößern

Bild vergrößern

Bild vergrößern

Information about the location of metastatic sites was available for 130 of 162 patients (80.2%, supplementary Table 2), but only for 25 patients FFPE material from metastases was available for sequencing (supplementary Table 3). For 15 patients, we detected the same mutations in liquid biopsies and metastases. For five patients, we did not detect any mutations in the liquid biopsy while there were eight detectable mutations in the metastatic site DNA, among them two ESR1 hotspot mutations. Conversely, also for five patients, we detected seven mutations in the liquid biopsy that were not detectable in the metastatic site DNA. Of note, all of these were ESR1 hotspot mutations. Two of these five patients had several sub-clonal ESR1 mutations in the liquid biopsy from which one was also detectable in the metastatic site DNA. Three of these five patients had a positive ESR1 mutation call exclusively in the liquid biopsy.

Since market-entering of the oral ER degrader elacestrant at the end of 2023, liquid biopsy testing of metastatic BC increased immensely. In the laboratory-developed NGS panel employed in this study, the four disease relevant genes ESR1, PIK3CA, ERRB2, and TP53 were tested. Detection of an ESR1 mutation in metastatic BC should be followed by changing anti-estrogenic therapy to fulvestrant or elacestrant [4, 5]. PIK3CA mutations are frequent in BC and therefore a useful biomarker to ensure that cfDNA originating from tumor cells is present. Furthermore, PIK3CA mutations are possible targets for alpelisib [9] or capivasertib [10]. ERBB2 mutations are present in bone marrow metastases from lobular BC cases, and are possibly targetable with anti-HER2 therapy [6]. TP53 mutations are associated with primary endocrine resistance and might not be adequately treated with endocrine therapy alone [11].

In a recent study on tissue biopsies from breast cancer metastases, we detected ESR1 hotspot mutations in 78 of 521 (15.0%) samples [12]. In the current study, using liquid biopsy, we detected ESR1 mutations in 25.9% of cases. In a recent meta-analysis of 16 studies with in total n = 2744 patients, the overall incidence of ESR1 mutations was 23% (ranging from 11 to 55%) [13]. Differences in the mutation frequency are possibly not caused by the testing material (tissue or liquid), but by the selection of samples in the studies. For example, in n = 194 ER-positive bone marrow BC metastasis, we found 14% ESR1 mutation frequency [14], whereas lung and liver metastasis showed 26.3% and 31.3% ESR1 mutation frequency, respectively [12]. Nearly 97% of the patient samples analyzed in the current study were collected after the EMA approval of elacestrant. This strong increase reflects the urgent clinical need of new therapeutic approaches in anti-endocrine therapy in metastatic BC.

Our comparison of cfDNA and metastatic site DNA from 25 patients shows that ESR1 hotspot mutations may be missed in both sample materials. In three patients, an ESR1 mutation was detected only in the liquid biopsy cfDNA, whereas two patients had the positive mutation call only in metastatic site DNA. False-negative results in liquid biopsy analysis can be caused by contamination with white blood cell DNA or absence of tumor DNA at time of blood collection. Further, the sample transfer from external clinics and oncology practice units to the molecular pathology laboratory is prone to errors, e.g., transfer time too long. Clear instructions and regular consultation are necessary to reduce handling errors which can lead to false-negative results [15]. Nevertheless, especially in patients with multiple metastatic sites, liquid biopsies are the preferred material for molecular analysis, because the genetic heterogeneity of metastatic sub-clones is potentially represented in the cfDNA. However, in case of negative results from a liquid biopsy, the analysis of a most recent metastasis specimen should be considered if possible.

In conclusion, our results demonstrate that ESR1 detection in liquid biopsy is feasible under routine conditions identifying important therapy-changing predictive biomarkers in a substantial portion of metastatic BC patients. To the best of our knowledge, this is the first study which demonstrates the feasibility and diagnostic gain of ESR1 mutation detection in real-world liquid biopsy outpatient samples since the approval of elacestrant.