Circulating tumor cells as early predictors of metastatic spread in breast cancer patients with limited metastatic dissemination

Breast cancer mortality has decreased considerably over the past two decades as a result of earlier diagnosis and major treatment advances in the adjuvant and metastatic settings [1]. Despite this progress, metastatic disease is still largely considered an incurable condition, and 5-year survival rates are <25% [1]. However, metastatic breast cancer is a heterogeneous disease, and long-term patient outcomes can be influenced by various biological features, as well as by the extent and site of metastatic disease. On the one hand, widespread visceral disease is typically associated with symptoms leading to deterioration of performance status and short survival [2]. On the other hand, limited metastatic dissemination, primarily confined to nonvisceral tissues, is more frequently associated with an indolent disease course and prolonged survival [3]-[5]. Nevertheless, the current standard assessment of metastatic disease by morphological and functional imaging does not provide adequate information on tumor biology and the presence of micrometastases, limiting the possibility to predict tumor metastatic potential [3].

During the past decade, several techniques capable of detecting and quantifying circulating tumor cells (CTCs) in cancer patients have been developed [4],[5]. It has been proposed that subpopulations of CTCs with tumor-initiating potential act as a central mediator of metastatic dissemination, giving rise to the formation of distant micrometastases, which subsequently generate overt detectable and frequently measurable lesions [6]. In support of this theory, multiple studies have shown that ≥5 CTCs/7.5 ml of blood, counted using the CellSearch System (Janssen Diagnostics, Raritan, NJ, USA) and evaluated before starting systemic treatment, is associated with poor outcome in patients with metastatic breast cancer [7]-[10]. In addition, high CTC counts are associated with greater metastatic tumor burden, expressed as the number of metastatic sites [8],[11]. Importantly, despite this association, the prognostic value of CTCs is independent from the initial number of metastatic sites [8],[11]. This may suggest that the negative prognostic impact of high CTC counts is not merely expression of the overt tumor burden, but also may reflect higher biological aggressiveness and presence of undetectable micrometastatic disease, and ultimately may predict a greater tendency to metastatic spread.

Therefore, we hypothesized that CTC counts, when evaluated prior to starting systemic treatment, represent an early marker of metastatic spread and are useful primarily in patients with limited metastatic dissemination and potentially eligible for locoregional treatments with a curative intent [12]-[14]. To test our hypothesis, we analyzed the patterns of recurring metastatic dissemination in patients with advanced breast cancer who had a CTC count before starting a new line of systemic treatment.

Systemic metastatic spread represents the main cause of breast cancer-related morbidity and death. Thus, for patients in the advanced stages of disease, it is critical to improve treatment strategies that may affect not only proliferation but also migration and invasion, important features of the metastatic process. The traditional factors currently used for prognostic stratification and treatment decisions in the advanced stages of breast cancer, including hormone receptor and HER2 status, site and extent of metastatic burden, and length of disease-free interval, often do not adequately predict treatment response and disease evolution. In addition, treatment aggressiveness is still debated in the presence of disease confined to nonvisceral organs or asymptomatic visceral metastases. In this scenario, the use of blood-based disease monitoring, such as CTC assessment, may represent a complementary and informative strategy [19]-[23]. Recently, the results of a prospective clinical trial confirmed the strong prognostic value of CTCs in metastatic breast cancer patients [10],[24]. Moreover, a large pooled analysis of individual patient data showed that adding CTC count status to clinicopathological predictive models significantly improved survival prognostication in advanced breast cancer [24]. The results of our present study suggest a novel role for CTCs as predictors of disease evolution in patients with advanced breast cancer, particularly those with limited metastatic dissemination. We found that pretreatment CTC counts ≥5/7.5 ml correlated with greater metastatic dissemination at the time of treatment failure, owing to more frequent appearance of new metastatic lesions and sites. Also, the extent of the newly developed metastatic burden was greater in patients with ≥5 CTCs/7.5 ml compared with those with <5 CTCs/7.5 ml [6]. We can hypothesize that elevated CTC numbers, when observed prior to starting systemic treatment, may reveal greater propensity to metastatic seeding and more extensive micrometastatic disease and thus may function as an early predictor of overt metastatic spread. This further emphasizes the need to investigate the molecular features of these cells in addition to their prognostic value. Expectedly, CTC counts ≥5/7.5 ml were also associated with lower frequency of disease confined to nonhematogenous metastatic sites, such as lymph nodes and soft tissues. Importantly, among the patients with nonvisceral metastases in our sample, ≥5 CTCs/7.5 ml predicted higher risk of visceral progression. This finding could have important clinical implications. Indeed, it is generally recommended that aggressive treatment strategies, such as chemotherapy-based regimens, are needed in cases of widespread visceral metastases and rapidly progressive disease [25],[26]. On the basis of our results, it can be proposed that pretreatment CTC counts may identify in advance patients who have higher risk of developing widespread visceral disease and consequently might benefit from early administration of more aggressive systemic treatments. Nevertheless, the retrospective nature of our analysis suggests the need to prospectively validate the use of CTCs as surrogate markers of metastatic potential, together with the investigation of new therapeutic strategies aimed at targeting biological properties of CTCs, with the goal of preventing or delaying PD and possibly improving patient outcomes [27],[28].

The link between micrometastatic disease and CTCs, as well as their capability to predict overt metastatic dissemination, may have important implications in the clinical management of patients with single-site metastatic disease. Indeed, among patients with this specific condition, pretreatment CTC counts ≥5/7.5 ml were correlated with higher risk of developing new metastatic sites and lesions at the time of treatment failure. This may imply limited value of combined therapeutic modalities that include local therapies in patients with high CTC counts. However, 43% of patients with single-site disease and <5 CTCs/7.5 ml experienced dimensional increases in the preexisting lesions when treatment resistance occurred. Hence, pretreatment CTC counts <5/7.5 ml may identify a subgroup of patients with single-organ disease who have higher probability of maintaining this condition for a longer time upon systemic treatment and for whom locoregional ablative procedures, if feasible, can contribute to symptom control and prolong survival [12]-[14].

The retrospective nature of the analyses and the lack of CTC molecular profiling represent the main limitations of this study, principally because molecular heterogeneity of CTCs may influence their role in metastatic seeding. In particular, specific subpopulations of CTCs undergoing molecular reprogramming known as the epithelial-to-mesenchymal transition (EMT) [29],[30] process lose their epithelial differentiation and acquire a mesenchymal phenotype, with increased invasion capabilities and stemness properties [30],[31]. The results of several studies suggest that CTCs with EMT features are the major effectors of metastatic seeding and are responsible for tumor progression [32]-[35]. Importantly, in the presence of the EMT process and loss of epithelial markers, EpCAM-based isolation methods, such as the CellSearch System, do not perform adequately. Thus, sensitive new strategies for CTC isolation and analysis are needed in order to thoroughly use these cells for prediction of disease evolution [36],[37].

Despite these limitations, our study is clinically relevant because we propose, for the first time to our knowledge, a potential marker of disease spread in breast cancer patients with limited metastatic dissemination.

Our findings suggest a new potential role for pretreatment CTC counts as early predictors of metastatic potential. This type of evaluation may be useful for improved risk stratification in patients with nonvisceral disease and single-organ metastatic involvement. Considering the limited number of patients evaluated in our study, our results warrant larger studies based on CTC enumeration and molecular profiling to develop novel treatment decision tools for breast cancer patients with limited metastatic dissemination.

MG and AG conceived of and designed the study, collected clinical data, performed data analysis and interpretation, and drafted the manuscript. SJ performed the CellSearch CTC analysis of all patients and contributed to data interpretation and manuscript drafting and revision. UDG and MM collected clinical data and contributed to data interpretation and manuscript drafting and revision. EC and HG performed data analysis and contributed to data interpretation and manuscript drafting and revision. SA performed data analysis and contributed to manuscript drafting and revision. BCH performed pathological review of CTC images in the Clinical Laboratory Improvement Amendments-certified pathology laboratory and contributed to data interpretation and manuscript revision. NTU, RHA, VV and GNH provided clinical assessment of patients and contributed to analysis, study conception and manuscript revision. SDP performed data analysis and contributed to analysis, study conception and manuscript revision. JMR and MC conceived of and designed the study, performed data analysis and interpretation, and contributed to manuscript drafting and revision. All authors substantially participated in manuscript revision. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All authors read and approved the final manuscript.

The authors’ affiliations at the time of study design, data collection and analysis were as follows. MG, UDG, MM and ENC: Department of Hematopathology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA. MC: Breast Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.