Breast Cancer Research and Treatment

, Volume 130, Issue 3, pp 809–816

The prognostic role of circulating tumor cells (CTCs) detected by RT-PCR in breast cancer: a meta-analysis of published literature

Authors

  • Shu Zhao
    • Department of Internal MedicineThe Third Affiliated Hospital of Harbin Medical University
  • Yupeng Liu
    • Department of EpidemiologyPublic Health College of Harbin Medical University
    • Department of Internal MedicineThe Third Affiliated Hospital of Harbin Medical University
  • Hongbin Li
    • Department of Internal MedicineThe Third Affiliated Hospital of Harbin Medical University
  • Minghui Zhang
    • Department of Internal MedicineThe Third Affiliated Hospital of Harbin Medical University
  • Wenjie Ma
    • Department of Internal MedicineThe Third Affiliated Hospital of Harbin Medical University
  • Wenhui Zhao
    • Department of Internal MedicineThe Third Affiliated Hospital of Harbin Medical University
  • Jingxuan Wang
    • Department of Internal MedicineThe Third Affiliated Hospital of Harbin Medical University
  • Maopeng Yang
    • Department of Internal MedicineThe Third Affiliated Hospital of Harbin Medical University
Preclinical study

DOI: 10.1007/s10549-011-1379-4

Cite this article as:
Zhao, S., Liu, Y., Zhang, Q. et al. Breast Cancer Res Treat (2011) 130: 809. doi:10.1007/s10549-011-1379-4

Abstract

The prognostic significance of circulating tumor cells (CTCs) in patients with breast cancer is controversial. We performed a meta-analysis of published literature to assess whether the detection of CTCs in patients diagnosed with primary breast cancer can be used as a prognostic factor. We searched Medline, Science Citation Index, and Embase databases as well as reference lists of relevant articles (including review articles) for studies that assessed the prognostic relevance of tumor cell detection in the peripheral blood (PB). A total of 24 eligible studies with 4,013 cases and 1,333 controls were included. Meta-analyses were performed using a random-effects model, using the hazard ratio (HR) and 95% confidence intervals (95% CIs) as effect measures. The positive detection of CTCs in patients was significantly associated with poor overall survival (OS) (HR = 3.00 [95% CI 2.29–3.94], n = 17, P < 0.0001) and recurrence-free survival (RFS) (HR = 2.67 [95% CI 2.09–3.42], n = 22, P < 0.0001). CTC-positive breast cancers were significantly associated with high histological grade (HR = 1.21 [95% CI 1.09–1.35], n = 34, P < 0.0001), tumor size (>2 cm) (HR = 1.12 [95% CI 1.02–1.22], n = 31, P = 0.01). and nodal status (≥1) (HR = 1.10 [95% CI 1.00–1.21], n = 32, P = 0.037), but cytokeratin-19 (CK-19) mRNA-positive CTCs were not associated with these clinicopathological parameters of breast cancer. Furthermore, the presence of CTCs was not associated with estrogen receptor (ER) negativity, progesterone receptor (PR) negativity, or human epidermal growth factor receptor type 2 (HER2) positivity. Detection of CTCs in the PB indicates poor prognosis in patients with primary breast cancer. Larger clinical studies are required to further evaluate the role of these markers in clinical practice.

Keywords

Circulating tumor cellsBreast cancerPrognosisRT-PCR

Introduction

In 2009, breast cancer accounted for 27% (192,370) of all new cancer cases among women in the United States [1]. Breast cancer is still a major challenge for women’s health, and the outcome of breast cancer largely depends on the development of metastases during the course of the disease. The detection of circulating tumor cells (CTCs) is one field of research that focuses on a new method of detecting metastatic disease earlier, less invasively, and more reliably than currently available conventional methods, such as clinical presentation, radiographic evaluation, and serum tumor marker detection. CTCs are defined as tumor cells that are circulating in the peripheral blood of patients, which are shed from either the primary tumor or its metastases. CTCs can be released from the primary tumor into the bloodstream and may colonize distant organs to give rise to metastasis. The presence of CTCs in the blood was documented more than a century ago by T. R. Ashworth, an Austrian pathologist who first reported this type of cells [2]. Soon thereafter, various methods have been described for their detection, including PCR-based strategies that increase the chances of distinguishing CTCs from the cellular background of normal cells and are currently considered the most sensitive techniques for detecting CTCs [3]. Many authors [417] have demonstrated that a CTC-positive status is associated with poor clinical outcome and may, therefore, play an important role in the diagnosis and treatment of breast cancer patients. However, other studies [1822] failed to show such an association between the presence of CTCs and poor clinical outcome and prognosis. Therefore, the relationship between CTCs and clinicopathological features remains controversial. The limited availability of samples may have resulted in this controversy of the clinical significance of these different CTC studies. This insufficiency of clinical data makes the clinical application of markers of CTCs unacceptable. The present meta-analysis was conducted to determine the association between CTCs and commonly identified clinical and pathological features of breast cancer, which allows us to consider the putative role of CTCs in the prediction of outcome in breast cancer.

Methods

Publication search

The following databases were systematically searched without time restrictions using the English language in November 2010: the electronic database PubMed (http://www.ncbi.nlm.nih.gov/pubmed/), Embase, and the Science Citation Index. “Circulating tumor cells” or “CTCs” and “breast cancer” were used as the key words. The citation lists associated with all the studies retrieved in the search were used to identify other potentially relevant publications. The reference list was checked for relevant articles, including review articles. The search results were then screened according to the following inclusion criteria: (1) any form of reverse transcription PCR (RT-PCR) used for the evaluation of the association between the putative markers of circulating tumor cells and either overall survival (OS), recurrence-free survival (RFS), or prognostic factors of breast cancer; (2) >20 analyzed patients and sufficient data to calculate a hazard ratio (HR) as a comparable effect estimate for RFS and/or OS or an odds ratio (OR) with a 95% confidence interval (95% CI); and (3) exclusion of letters to the editor, reviews, and articles published in non-English language books or papers.

Two reviewers (S Zhao and YP Liu) independently screened and retrieved the literature list and, in the case of potentially relevant references, obtained the full articles and extracted the following data from each study: the year of publication, the first author’s surname, the number of cases and controls, the number of different clinical and pathological parameters, and the assessment methods of survival expression. Disagreements were resolved by discussion and were checked by a third investigator, QY Zhang.

Statistical analysis

We used the maximally adjusted risk estimates with 95% confidence intervals in our meta-analyses, but when these data were not available, we used the minimal or crude data instead. Heterogeneity between studies was tested with the Q test and I2 statistic. If the I2 statistic was ≤50%, the fixed effect model was used to pool studies; otherwise, the random-effects model was used. We evaluated potential publication bias by a funnel plot, which was further examined by the Egger and Begg’s test. Statistical analyses were done using Comprehensive Meta-analysis software, version 2.0 (Biostat, Inc., USA). All tests were two-sided, and P values less than 0.05 were considered to be statistically significant.

Results

Description of studies

The systematic literature search yielded a total of 24 studies, which comprised 4,013 patients for a final analysis. All 24 studies analyzing peripheral blood applied a molecular detection method (PCR, RT-PCR, or RT followed by quantitative PCR) of circulating tumor cells. Fifteen studies based their detection of tumor cells on one tested gene/antigen [48, 10, 11, 13, 15, 16, 18, 19, 2123]. The remaining studies [3, 9, 12, 14, 17, 20, 2426] used two or more different markers in their assays. The baseline characteristics of the included studies are summarized in Table 1. In 15 studies, analysis of CTC detection as a prognostic factor was confirmed using multivariate analysis. Twenty-three studies included early breast cancer, and five studies included metastatic breast cancer.
Table 1

Baseline characteristics of included studies for the meta-analyses

References

Year

CTCs markers (n/N %)

Age (range), year

Tumor stage

Case

Control

Folllow-up, month

Outcomes measured

Grunewlad [24]

2000

hMAM (11/133 8.3) EGFR (13/133 9.8) CK19 (64/133 48.1)

NR

I–IV

133

91

NR

NR

Suchy [18]

2000

hMAM (11/98 11.2)

49 (30–69)

I–IV

98

33

49 (8–107)

RFS

Kahn [4]

2000

CK19 (48/109 44.0)

48 (34–89)

I–IV

109

45

NR

NR

Stathopoulou [5]

2002

CK19 (44/148 29.7)

54 (30–74)

I–II

148

113

28 (7–62)

RFS, OS

Xenidis [6]

2003

CK19 (44/161 27.3)

54 (30–74)

I–II

161

No

29 (7–55)

RFS, OS

Lin [19]

2003

hMAM (41/112 36.6)

47 (32–85)

I–IV

112

27

NR

NR

Jotsuka [7]

2004

CEA (30/100 30.0B) (14/100 14.0A)

NR

I–III

100

14

56 (42–70)

RFS

Ntoulia [8]

2006

hMAM (14/101 13.9)

53 (26–75)

I–II

101

70

24 (1–80)

RFS, OS

Chen [9]

2006

PTTG1 (70/92 76.1) Survivin (77/92 83.7) UbcH10 (76/92 82.6) TK1 (78/92 84.8)

50

I–III

92

100

NR

NR

Xenidis [10]

2006

CK19 (36/167 21.6)

49 (30–80)

I–III

167

No

32 (3–88)

RFS, OS

Benoy [20]

2006

CK19 (22/148 14.9) hMAM (29/148 19.6)

53 (28–88)

I–IV

148

37

26 (16–29)

OS

Xenidis [11]

2007

CK19 (22/119 18.5)

50 (28–77)

I–II

119

No

65 (22–106)

RFS, OS

Ignatiadis [12]

2007

CK19 (63/185 34.1)HER2 (33/185 17.8)

55 (28–80)

I–III

185

No

65 (10–120)

RFS, OS

Ignatiadis [13]

2007

CK19 (181/444 40.8)

54 (26–78)

I–III

444

No

53.5

RFS, OS

Mikhitarian [25]

2008

CEA (51/215 23.7) hMAM (29/215 13.5) PSE (58/215 27.0) PIP (36/215 16.7) EpCAM (7/215 3.3) CK19 (0/215) Muc1 (0/215)

56 (29–84)

I–III

215

49

NR

NR

Marques [21]

2008

hMAM (52/278 18.7)

51 (22–81)

I–III

278

28

43 (3–74)

RFS, OS

Ignatiadis [14]

2008

CK19 (72/175 41.1) hMAM(14/175 8.0) HER2 (50/175 28.6)

54 (28–75)

I–III

175

No

NR

RFS, OS

Shen [26]

2008

hMAM (31/94 33.0) Survivin (34/94 36.2) Htert (56/94 59.6)

50

I–IV

94

100

NR

NR

Apostolaki [15]

2008

HER2 (53/216 24.5)

55 (28–79)

I–II

216

104

78 (5–108)

RFS, OS

Xenidis [16]

2009

CK19 (179/437 41.0B) (143/437 32.7A)

54 (26–78)

I–III

437

98

54 (10–106)

RFS, OS

Daskalaki [22]

2009

CK19 (91/165 55.2B) (79/162 48.8A)

54 (26–75)

I–II

165

No

59 (13–95)

RFS, OS

Auwera [3]

2009

CK19(20/76 26.3) hMAM(41/76 53.9)

62 (34-85)

IV

76

20

NR

NR

Chen [17]

2010

CK19 (20/50 40.0) hMAM(13/50 26.0) CEA(7/50 14.0)

50 (29–69)

I–III

50

74

36

RFS

Ferro [23]

2010

hMAM (18/190 9.5)

63 (33–93)

I–III

190

330

NR

NR

NR not reported, OS overall survival, RFS recurrence-free survival, B before chemotherapy, A after chemotherapy

Impact of circulating tumor cells on the survival of breast cancer patients

Using the methods described above, the overall survival (OS) and/or recurrence-free survival (RFS) of 2,894 patients in 15 studies were analyzed. The presence of CTCs was significantly associated with poor OS (HR = 3.00 [95% CI 2.29–3.94], n = 17, P < 0.0001) (Fig. 1a) and RFS (HR = 2.67 [95% CI 2.09–3.42], n = 22, P < 0.0001) (Fig. 1b). We also stratified for the use of CK-19 (OS: HR = 3.59 [95% CI 2.52–5.11], n = 12, P < 0.0001 and RFS: HR = 2.94 [95% CI 2.17–3.98], n = 13, P < 0.0001) and hMAM (OS: HR = 1.94 [95% CI 1.03–3.67], n = 3, P = 0.04) (supplementary data) as the most frequently used CTC markers. These subgroup analyses suggested an association between poor prognosis and tumor cell detection; however, the hMAM mRNA-positive CTCs did not emerge as an independent prognostic factor for RFS (HR = 1.62 [95% CI 0.95–2.74], n = 5, P = 0.075). In addition, the presence of HER2 mRNA-positive CTCs was correlated with poor OS (HR = 2.27 [95% CI 1.34–3.85], n = 3, P = 0.002) and RFS (HR = 2.84 [95% CI 1.71–4.62], n = 3, P < 0.0001) (supplementary data), although not many studies were available for this analysis.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-011-1379-4/MediaObjects/10549_2011_1379_Fig1_HTML.gif
Fig. 1

Forest plot of HR for OS (a) and RFS (b) among the included studies. The plot shows the combined HR, which is calculated by a fixed-effects model, and demonstrates that the CTCs can be used as prognostic factors of OS and RFS in breast cancer patients. (Favor A prolonged OS and RFS, Favor B shorted OS and RFS)

Correlation of circulating tumor cells with clinicopathological parameters

CTC-positive breast cancers were associated with biologically aggressive phenotypes, such as high histological grade (HR = 1.21 [95% CI 1.09–1.35], n = 34, P < 0.0001; the overall result of the analysis was not altered by the exclusion of the studies by Mikhitarian as they contained a very large HR = 113.7) (Fig. 2a), tumor size (>2 cm) (HR = 1.12 [95% CI 1.02–1.22], n = 31, P = 0.01) (Fig. 2b), and axillary lymph node positivity (HR = 1.10 [95% CI 1.00–1.21], n = 32, P = 0.037) (Fig. 2c). However, CK-19 mRNA-positive CTCs were not associated with high histological grade (HR = 1.08 [95% CI 0.94–1.23], n = 12, P = 0.272), tumor size (>2 cm) (HR = 1.07 [95% CI 0.96–1.20], n = 11, P = 0.191), or axillary lymph node positivity (HR = 1.01 [95% CI 0.90–1.14], n = 11, P = 0.819). Furthermore, hMAM mRNA-positive CTCs were not associated with high histological grade (HR = 1.33 [95% CI 1.03–1.72], n = 10, P = 0.027) (supplementary data).
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-011-1379-4/MediaObjects/10549_2011_1379_Fig2a-b_HTML.gif
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-011-1379-4/MediaObjects/10549_2011_1379_Fig2c_HTML.gif
Fig. 2

Forest plot of HR was assessed for an association between CTCs and clinicopathological features, such as tumor grade (a) (Favor A CTCs had no relationship with high histological grade, Favor B CTCs were associated with high histological grade); tumor size (b) (Favor A CTCs were not associated with tumor size >2 cm, Favor B CTCs were associated with tumor size >2 cm); and lymph node status (c) (Favor A CTCs were not associated with axillary lymph node positivity, Favor B CTCs were associated with axillary lymph node positivity)

CTC positivity was not associated with clinical parameters such as estrogen receptor (ER) negativity (HR = 1.09 [95% CI 0.97–1.21], n = 29, P = 0.141), progesterone receptor (PR) negativity (HR = 0.98 [95% CI 0.89–1.08], n = 29, P = 0.732), or human epidermal growth factor receptor type 2 (HER2) positivity (HR = 1.13 [95% CI 0.94–1.36], n = 20, P = 0.207) (supplementary data). Interestingly, the HER2 mRNA-positive CTCs were not associated with HER2 status in some studies (immunohistochemistry ++−+++) (HR = 0.89 [95% CI 0.54–1.44], n = 3, P = 0.626) (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-011-1379-4/MediaObjects/10549_2011_1379_Fig3_HTML.gif
Fig. 3

Forest plot of HR for HER2 mRNA-positive CTCs and their association with HER2. The plot shows that the HER2 mRNA-positive CTCs were not associated with HER2 (immunohistochemistry ++−+++)

Sensitivity analyses and publication bias

Sensitivity analyses using fixed-effects model suggested that the results did not change substantially. There was no obvious publication bias in this meta-analysis (Fig. 4).
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-011-1379-4/MediaObjects/10549_2011_1379_Fig4_HTML.gif
Fig. 4

Funnel plot of publication bias. The plot shows that there was no obvious indication of publication bias for the outcome of OS and RFS

Discussion

Through a meta-analysis study in 2005, Braun et al. reported that the presence of micrometastases in the bone marrow of patients with breast cancer is an independent predictor of poor prognosis [27]; however, there are no pooled studies reporting the relationship between CTCs and clinical outcome in breast cancer patients. Our present meta-analysis is the first study to systematically estimate the association between CTC markers and breast cancer survival as well as clinicopathological parameters. The present results indicate that CTC markers were significantly associated with high histological grade, axillary lymph node positivity, and tumor sizes >2 cm, as well as OS and RFS. This phenomenon indicates that with more advanced stages (TNM) of cancer, CTCs are more easily detected because, in patients with advanced breast cancer, tumor cells can easily shed from the primary tumor and enter the blood circulation. This association was especially strong in the analysis of the CK-19 (OS and RFS) subgroup, which suggests that this marker could be developed for clinical applications. In the present study, the relationship between CTCs (including all markers) and ER, PR and HER2 were not obtained; thus, the molecular and immunophenotypic characterization of the mRNA expression of markers of CTC-positive patients with ER-negative (basal-like CTCs) or ER-positive (luminal-like CTCs) cancer requires further confirmation. Recently, Theodoropoulos et al. reported the existence of a subpopulation of CTCs with putative stem cell progenitor phenotypes in patients with metastatic breast cancer [28]. It would be interesting to examine whether there are different subpopulations of micrometastatic cells with stem cell/progenitor properties that are responsible for the development of metastases in ER-negative and ER-positive patients.

The present meta-analysis encompasses many different CTCs markers, and CK-19 and hMAM have been studied most extensively. CK-19, a general marker of epithelial cells, is strongly expressed in epithelial but not in mesenchymal cells and has been extensively studied as a potential marker for the detection of CTCs in breast cancer [36, 1014, 16, 17, 2022, 24, 25]. hMAM, a member of the secretoglobin-uteroglobin family, was originally identified from a primary human breast cancer by Watson and Fleming [29]. It is restricted to the adult mammary gland and to breast cancer cell lines. The expression of hMAM has been shown to have diagnostic and prognostic value as a specific molecular marker in breast cancer [8, 14, 2325]. In some studies, however, hMAM can also be found in patients with benign breast disease [17] and other cancers such as ovarian cancer [18]. In the present subgroup analysis, the CK-19 mRNA-positive CTCs were not associated with clinicopathological parameters (e.g., grade, tumor size, and lymph nodes) of breast cancer. A possible reason for this may be that breast cancer cells may disseminate from the primary tumor to axillary lymph nodes and distant metastatic sites by different routes. These separate routes may depend on the different chemokine receptor-ligand axis [30, 31]. Interestingly, the HER2 mRNA-positive CTCs were not associated with HER2 status (immunohistochemistry ++−+++) (P = 0.635) or other important clinical or pathological tumor parameters (P > 0.05). This finding is in agreement with the previously described heterogeneity of circulating occult tumor cells in breast cancer [32, 33].

There are certain limitations in the present meta-analysis. First, we noticed a considerable degree of interstudy heterogeneity. Differences in detection methods, types and numbers of target genes, sampling site and time, and the demographic or clinicopathological data of the included patients should be considered as potential sources of heterogeneity. Our study, however, indicates that the detection of CK-19 mRNA-positive CTCs both before and after chemotherapy is correlated with poor outcome in patients with breast cancer, especially at the early stage of the disease. We addressed the heterogeneity by a rigorous methodological approach that uses a random-effects model for more conservative estimates. We required a minimum of three studies to carry out pooled analyses, and only included studies with a minimum sample size of 20 patients. In addition, we applied several subgroup and sensitivity analyses to assess potential sources of bias and the observed interstudy heterogeneity. Second, although approaches based on RT-PCR have a high sensitivity for the detection of CTCs in general, an important limitation is that these methods cannot quantify the number of CTCs, and no morphological evaluation of cells can be obtained.

In summary, available evidence supports the notion of a strong prognostic value of CTCs in the peripheral blood. It may be speculated that these patients may potentially benefit from adjuvant therapy. Therefore, monitoring of CTCs during the administration of adjuvant treatment could permit the tailoring of treatment according to the prognosis of each individual patient. In addition, because blood can be easily obtained at different times during the follow-up sessions, the simultaneous detection of CTC markers might serve as a tool to assess the “real-time” CTC status of patients. Thus, detection of CTCs during patient follow-ups would offer the opportunity for early intervention. This may make the eradication of these circulating tumor cells more feasible, especially if the patient tumor burden is still low and the appearance of clinically overt metastases is undetected. Some studies have shown that trastuzumab can effectively eliminate CTCs overexpressing HER2 regardless of the HER2 status of the primary tumor in patients with breast cancer [34, 35]. Similarly, we can speculate that the presence of CK-19+ and/or hMAM mRNA+ tumor cells in the peripheral blood of breast cancer patients could be used to select for those people that could derive potential benefit from vaccine strategies targeting CK-19- and/or hMAM-expressing cells. Moreover, clinical trials are required to prospectively assess the value of postoperative CTC detection as a simple method to monitor response to systemic therapy. However, we note that these studies should apply standardized detection techniques to reduce interstudy heterogeneity. These hypotheses should be further tested in well-designed, adequately powered, prospective, and randomized clinical studies.

Acknowledgments

This experiment was finished in the oncobiology key lab of the Heilongjiang common institution of higher learning. This work was supported by the National Natural Science Foundation of China [grant number 81071889].

Supplementary material

10549_2011_1379_MOESM1_ESM.doc (130 kb)
Supplementary material 1 (DOC 130 kb)

Copyright information

© Springer Science+Business Media, LLC. 2011