Circulating tumor cells in HER2-positive metastatic breast cancer patients: a valuable prognostic and predictive biomarker

Human epidermal growth factor receptor 2 (HER2) is a 185 kDa transmembrane tyrosine kinase receptor encoded by the HER2 gene on chromosome 17q21. HER2 over-expression or amplification occurs in approximately 20% of all breast cancer patients and is associated with aggressive growth, short survival and poor prognosis [1‐3]. HER2 positivity correlates with the clinical outcome of treatment with anti-HER2 agents such as trastuzumab (Herceptin, Genentech, South San Francisco, CA, USA) and lapatinib (Tykerb, GSK, Philadelphia, PA, USA) [4‐7]. Therefore, HER2 is considered to be a vital prognostic and predictive factor, and treatment of HER2-positive patients remains one of the great therapeutic challenges in metastatic breast cancer (MBC).

Despite therapeutic advances over the past decades, individually tailored therapeutic regimens for HER2-positve patients remain far from satisfactory. For example, the benefit of single-agent anti-HER2 therapy, in the form of either trastuzumab or lapatinib, is only in the range of ~25% [8]. There are three possible explanations for this phenomenon. First, HER2-positive MBC patients may be divided into subgroups with different prognoses. Second, the initial assessment of HER2-positivity may be inaccurate due to the inherent limitations of traditional methods, including tumor heterogeneity, subjectivity in the interpretation of results or technical limitations such as variability in tissue processing and reagents [9]. Third, previous studies have demonstrated that there are inconsistencies in HER2 expression between primary tumors and their metastases [10, 11]. Such inconsistencies indicate that tumor cells are under constant evolvement or clonal selection and that the detected HER2 status may not necessarily reflect the patients’ real-time phenotypes. Therefore, it is critical to discover more precise prognostic marker and real-time methods for HER2 testing to optimize individualized therapeutic regimens for HER2-positive MBC patients.

Circulating tumor cells (CTCs) are cells that shed from the tumor and enter the circulation, a process that is required for cancer metastasis. Considerable efforts have been made to develop technologies for CTC detection and characterization; among these technologies, the CellSearch system (Veridex LLC, Raritan, NJ, USA) is the only one approved by US Food and Drug Administration (FDA) for clinical use in treating MBC [12, 13]. CellSearch has a CTC detection cutoff of ≥5 cells/7.5 ml blood, and it has been demonstrated to be an independent prognostic factor for the prediction of progression-free survival (PFS) and the overall survival (OS) of MBC patients [14‐17]. In addition to detection, the molecular characterization of CTCs is now recognized as a valuable tool that can provide real-time information to distinguish subgroups of patients who can benefit from certain types of therapy [18, 19]. Unfortunately, previous studies failed to illustrate the prognostic value of CTCs in HER2-positive MBC patients using CellSearch [20]. Although recent studies have made great efforts to compare HER2 statuses between tumor tissue and CTCs [18, 19, 21‐29], a satisfactory CTC HER2-positive criterion has not yet been established.

In the present study, we demonstrated that CTC enumeration using a modified cutoff has superior prognostic value for HER2-positive MBC. Furthermore, we proposed that HER2 positivity should be determined by both the HER2 intensity of individual CTCs and the percentage of CTCs with that intensity. Using the criterion defined as >30% of CTCs over-expressing HER2, we found that HER2 expression in CTCs was different from that in tumor tissues, and this expression could significantly improve the response prediction for anti-HER2 therapy.

In this study, we found that CTC enumeration with a cut-off of ≥1 but not ≥5 CTCs could serve as a useful prognostic factor for HER2-positive MBC patients. CTC enumeration using CellSearch (with a cut-off of ≥5 CTCs) is widely accepted as a prognostic factor for MBC patients [14‐17]; however, its prognostic power for the HER2-positive subgroup seems to be inadequate [20]. More studies are needed to determine an appropriate cut-off for all subgroups of MBC patients. We found that HER2-positive patients had relatively lower CTC counts than ER-positive/HER2-negative patients. Our results were consistent with the report of Giordano et al., in which a larger proportion of HR-positive/HER2-negative patients had ≥5 CTCs than those with other tumor subtypes (P = 0.024) [20]. In line with these findings, Punnoose et al. found that the CTCs in a population of HR-positive/HER2-negative patients displayed higher levels of EpCAM, a CTC enrichment marker used in the CellSearch system [19]. It is possible that the ≥5-CTCs cut-off is an unsuitable prognostic indicator for a subgroup of HER2-positive patients. We tried other possible cut-offs and eventually found that the cut-off of ≥1 CTC yielded significant differences in PFS. Our results indicated that the underlying molecular subtype and gene expression patterns might not be same for different subtype of patients and that an adapted cut-off should be considered to make prognosis judgments for various subgroups of patients.

Based on clinical outcomes, our results indicated that a CTC HER2-positive criterion defined as >30% of CTCs over-expressing HER2 could improve the response prediction in anti-HER2 therapy. In recent years, great efforts have been undertaken to compare the HER2 status of tumor tissue and CTCs and determine whether anti-HER2 therapy would be beneficial. The studies summarized in Additional file 4: Table S3 indicated that the HER2 status of the CTCs was totally different from that of the tumor tissue. The overall discrepancy rate between the two sample sources ranged from 15% to 61%. More importantly, the HER2 detection methods used in these studies varied, and there is no current consensus on how HER2 positivity should be determined in CTCs.

For the IF-based HER2 staining method used in the CellSearch system, two HER2 positivity criteria were proposed. According to Pestrin et al., CTCs can be defined as HER2-positive if at least 50% of them were HER2-positive by IF [27]. Riethdorf et al. [28] and Ignatiadis et al. [29] noted that the intensity of HER2 staining using the CellSearch system was variable, ranging from absent or weak to intermediate and sometimes bright. They proposed a model in which HER2 expression in CTCs was scored as 0, 1+, 2+, or 3+ according to the staining intensity of HER2 in 6 types of breast cancer cell lines with known HER2 statuses [28, 29]. CTCs were categorized as HER2-positive if at least one CTC showed strong HER2 staining intensity; however, due to CTC heterogeneity, the intensity of an individual CTC might not represent the actual HER2 status of the patient.

We postulated that a reasonable CTC HER2-positive criterion should seek experience from IHC, which detect HER2 protein expression with a similar technological principle as IF. Most importantly, the criterion should be validated by clinical evidence. The HER2 positive criterion using IHC was defined as uniform and intense membrane staining of >30% of invasive tumor cells membrane staining (the original threshold was >10%) [30]. Accordingly, we proposed and tested two criteria for HER2 positivity: >30% or >10% of CTCs over-expressing HER2. Conceivably, such criteria that combine qualitative and quantitative aspects encompassed a comprehensive evaluation of the entire pool of the isolated CTCs. Based on the patients’ clinical outcomes, we found that only the 30% threshold could give more precise instruction for anti-HER2 therapy.

Using this threshold, we found that, surprisingly, only patients who have both HER2-positive tumor tissue and CTCs could substantially benefit from anti-HER2 therapy. Conversely, up to 52% (14/27) of the histologically HER2-positive patients had actually HER2-negative CTCs, and these patients may not benefit from anti-HER2 therapy. Our results are consistent with the recent work of Niikura et al., who reported that patients with HER2-positive primary breast tumors could not benefit from trastuzumab therapy due to loss of HER2 in the metastases [11].

Our data underscore the importance and urgency of HER2 testing in CTCs, which is a real-time and dynamic procedure compared with HER2 testing on metastatic tumors. Through CTC characterization, patients with HER2-positive tumors and CTCs are strongly recommended to undergo anti-HER2 therapy. Furthermore, patients who have HER2-positive tumors but HER2-negative CTCs could avoid overtreatment with anti-HER2 agents.

Even though our study may help select patients for anti-HER2 therapy, it was an exploratory single-center study, and the number of the enrolled patients was not adequate for powerful statistical analysis. To obtain more robust evidence, large cohort, multi-center and prospective clinical trials should be designed in the near future, in which therapeutic decisions are based on HER2 analyses of both tumor tissue and CTCs.

Our data demonstrate that CTC enumeration with a modified cut-off is a valuable prognostic tool for HER2-positive MBC patients. The HER2 status of CTCs may be different from that of tumor tissues and can predict responses to anti-HER2 therapy. Our findings underscore the necessity of a comprehensive CTC analysis (regarding both number and HER2 status), which may be a valuable prognostic and predictive tool for optimizing individually tailored therapies for HER2-positive MBC patients.