Meta-Analysis of the Prognostic Value of Circulating Tumor Cells … · Imaging, Diagnosis, Prognosis Meta-Analysis of the Prognostic Value of Circulating Tumor Cells in Breast Cancer

Imaging, Diagnosis, Prognosis

Meta-Analysis of the Prognostic Value of Circulating TumorCells in Breast Cancer

Liling Zhang1, Sabine Riethdorf2, Gang Wu1, Tao Wang1, Kunyu Yang1, Gang Peng1, Junli Liu1, andKlaus Pantel2

AbstractPurpose: The prognostic value of circulating tumor cells (CTC) detected in breast cancer patients is

currently under debate. Different time points of blood collections and various CTC assays have been used in

the past decades. Here, we conducted the first comprehensive meta-analysis of published literature on the

prognostic relevance of CTC, including patients with early and advanced disease.

Experimental Design:A comprehensive search for articles published between January 1990 and January

2012 was conducted; reviews of each study were conducted and data were extracted. The main outcomes

analyzed were overall survival (OS) and disease-free survival (DFS) in early-stage breast cancer patients, as

well as progression-free survival (PFS) andOS inmetastatic breast cancer patients. Pooled hazard ratio (HR)

and 95% confidence intervals (CIs) were calculated using the random and the fixed-effects models.

Subgroup and sensitivity analyses were also conducted.

Results: Forty-nine eligible studies enrolling 6,825 patients were identified. The presence of CTC was

significantly associated with shorter survival in the total population. The prognostic value of CTC was

significant in both early (DFS: HR, 2.86; 95% CI, 2.19–3.75; OS: HR, 2.78; 95% CI, 2.22–3.48) and

metastatic breast cancer (PFS: HR, 1.78; 95% CI, 1.52–2.09; OS: HR, 2.33; 95% CI, 2.09–2.60). Further

subgroup analyses showed that our results were stable irrespective of the CTC detection method and time

point of blood withdrawal.

Conclusion: Our present meta-analysis indicates that the detection of CTC is a stable prognosticator in

patients with early-stage and metastatic breast cancer. Further studies are required to explore the clinical

utility of CTC in breast cancer. Clin Cancer Res; 18(20); 5701–10. �2012 AACR.

IntroductionMetastasis is the main cause of cancer-related death.

However, even by currently available high-resolution imag-ing technologies,micrometastasis cannot be detected. Thus,in recent years much work has been entered into thedetection and characterization of disseminated tumor cells(DTC) and circulating tumor cells (CTC). Through a pooledanalysis accounting for 4,703 patients, Braun and collea-gues (1) in 2005 reported that the presence of DTC in breastcancer patients is an independent predicator of poor prog-nosis. However, bone marrow sampling is an invasive

procedure not easily accepted in the management of breastcancer. Thus, the focus in recent years has shifted to thedetection of CTC in peripheral blood collections.

CTC are tumor cells detectable in blood that are releasedby primary tumors, recurrences, or metastases and thatpossess antigenic and genetic tumor-specific characteristics.The first report on tumor cells inbloodstreamwas attributedto Ashworth in 1869 (2). As then, various methods havebeen developed and optimized for their detection. Immu-nocytochemistry (ICC) and reverse transcriptase polymer-ase chain reaction (RT-PCR) are the 2 main approachescurrently. Among various ICC methods, the CellSearchsystem is the only assay cleared by the U.S. Food and DrugAdministration (FDA) for clinical use. This system usesferrofluids coated with an antibody against the epithelialcell adhesion molecule (EpCAM) to magnetically enrichepithelial cells from whole blood. However, EpCAM-pos-itive CTC are missed by CellSearch device and various newCTC assays have been developed in the recent years. Nev-ertheless, it is still unclear whether the existing CTC assaysmay detect irrelevant "bystander" cells and miss the metas-tasis-initiating cells.

Despite the technical advancements inCTCdetection, theprognostic relevance of CTC in breast cancer remains

Authors' Affiliations: 1Cancer Center, Union Hospital, Tongji MedicalCollege, Huazhong University of Science and Technology, Wuhan, China;and 2Department of Tumor Biology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Corresponding Author: Professor Klaus Pantel, Department of TumorBiology, UniversityMedical Center, Hamburg-Eppendorf, Martinistr. 52, D-20246Hamburg, Germany. Phone: 49-40-74105-3503; Fax: 49-40-74105-5379; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-12-1587

�2012 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org 5701

on April 11, 2021. © 2012 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst August 20, 2012; DOI: 10.1158/1078-0432.CCR-12-1587

http://clincancerres.aacrjournals.org/

controversial. In fact, the recently released American Societyof Clinical Oncology Tumor Marker Guidelines stated thatmeasurement of CTC should not be used for diagnosis ortreatment modification in patients with breast cancer (3).Some studies (4–13) showed that the presence of CTC wassignificantly associated with shorter survival in breast can-cer. In contrast, other studies (14–20) failed to showsuch anassociation between the presence of CTC and worse prog-nosis. This discrepancy may result from the rather smallsample size of patients enrolled in these studies as well asdifferences in the time point of blood collection and CTCassays used.

With the aim to gain a better insight into the prognosticvalue of CTC in patients with breast cancer, we conductedthe first comprehensive meta-analysis of published litera-ture on this topic, including patients with early andadvanced disease. In particular, we evaluated the prognosticvalue of CTC status (presence vs. absence) on disease-freesurvival (DFS) and overall survival (OS) in early-stagepatients (M0), as well as on progression-free survival (PFS)and OS in metastatic patients (M1). Furthermore, we madesubgroup analyses to evaluate whether the detection meth-od and time point of blood collection influence the prog-nostic value of CTC.

Materials and MethodsSearch strategy

A comprehensive search of Medline and ISI Web ofKnowledge was done between January 1990 and January2012 as themajor reports delivered before 1990 focused onthe technical issues. The search strategy included the fol-lowing keywords variably combined: "circulating tumorcell (s)," "breast cancer," "breast neoplasm," "micrometas-tasis," and "prognos�." Furthermore, reference lists ofretrieved articles and reviewarticleswere reviewedmanuallyto implement our search. Only studies published in peer-reviewed journals were included, data from letters andmeetings abstracts were not eligible.

Eligibility criteriaThe studies were included in the meta-analysis if they

reported survival data in breast cancer patients stratified byCTC status (presence/positive and absence/negative), pro-vided sufficient data for determining an estimate of HR anda 95% confidence interval (CI), and enrolled more than 30patients, and study patients did not overlap with patients inother included studies. Although no language restrictionswere imposed initially, for the final analysis only articleswritten in English were included.

When more than 1 blood sample per patient was with-drawn, the results were excluded if survival data were notstratified by CTC status at each time point (e.g., persistent,increased, and decreased). We also excluded the studies inwhich histologic type of breast malignancies was inflam-matory breast cancer or sarcoma.

We did not assign each study a quality score, because nosuch score has received general agreement for use in aprognostic meta-analysis, especially of observational stud-ies. Instead,we conducted subgroup and sensitivity analysesbecause they are widely recommended.

Data extraction and outcomesWe recorded the following information about each eli-

gible trial: author’s names, journal and year of publication,number of patients analyzed, tumor stage, volume andtiming of blood withdrawal, method of CTC detection, andcutoff value of CTC.We also recorded PFS,DFS,OS, survivalcurves, HR, and 95% CI if available. The HR was measuredby comparing the CTC positive and CTC negative arms. Forone study (21), the reportedHR referred to theCTCnegativearm rather than the CTC positive arm and was recalculatedby taking reciprocal of HR to keep consistency with othertrials.

When more than 1 blood sample per patient was with-drawn at different time points, we recorded results of eachtime point and classified these time points as "baseline,""mid-therapy," and "posttherapy."

When more than 1 marker was used to detect CTC, andtheHR for survival or the survival curve was reported for thedifferent markers, we recorded all of these results as inde-pendent data sets.

The endpoints of clinical outcome analyzed were OS,PFS, and DFS. In patients with stage M0, the effect of CTCdetection on DFS was evaluated, whereas in patients withstage M1 the respective endpoint was PFS. In both groups,the effect of CTC detection on OS was analyzed. Additionalsubgroup analyseswere conductedwith regard to the type ofthe detection method and the time point of blood with-drawal in relation to systemic therapy.

Statistic analysisHR of each study was either directly collected from the

original article or calculated as suggested by Parmar andcolleagues (22). The pooled HR for survival was calculatedby fixed and random-effects models. And if heterogeneityamong the studies was observed, only random-effects esti-mates were presented.

Translational RelevanceCTC can be detected in peripheral blood from early-

stage andmetastatic breast cancer patients, and differenttime points of blood collections and various CTC assayshave been used in the past decades. Although withencouraging results from smaller single-center studies,the prognostic value of CTC in breast cancer patientsremains uncertain because single trials may report con-flicting results. A comprehensivemeta-analysismay helpresolve this controversy and givemore accurate estimatesof the average effect. The results of our meta-analysisshowed that the presence of CTC was significantly asso-ciated with poorer survival in both early and metastaticbreast cancer, which may help provide more convincingevidence to using CTC detection in future clinicalpractice.

Zhang et al.

Clin Cancer Res; 18(20) October 15, 2012 Clinical Cancer Research5702




Heterogeneity between studies was evaluated with theCochran’s Q test as well as the I2 index. Potential causes ofheterogeneity were explored by meta-regression analyses.Publicationbiaswas evaluated using the funnel plot and theEgger’s and Begg’s test. The effect of publication bias on thepooled effect was assessed by the "trim-and-fill" method.To evaluate the stability of the results, a 1-way sensitivity

analysis was conducted. The scope of this analysis was toevaluate the influence of individual studies by estimatingthe average HR in the absence of each study.All statistical tests were conducted with the software

STATA version 10.0.

ResultsCharacteristics of identified studiesSeven hundred thirty-three records were identified by the

primary computerized literature search. However, afterscreening of the titles and abstracts, 672 studies wereexcluded because they were either duplicate, review articles,laboratory studies, written in non-English, or irrelevant tothe current study. Sixty-one articles were further reviewed indetail. Twelve studies were further excluded because ofmultiple publications, no survival data, or small samplesize. In refs. 23 and 24, although the patients came fromthe samepopulation, the detectionofCTCwas conducted at"posttherapy" and "baseline," respectively. Therefore, bothstudies were included, but the number of patients wasonly extracted from the larger one. Finally, 49 studies

(6–21, 23–55) were identified as eligible for inclusion inthe meta-analysis (Fig. 1).

The included 49 studies encompassed 6,825breast cancerpatients. Early stage (M0, n¼ 2,993; refs. 11–13, 17–19, 23–25, 28–31, 34, 35, 37, 39, 43, 51) and metastatic (M1, n ¼3,069) patients (6–10, 20, 21, 27, 33, 36, 40–42, 44, 45, 47–50, 52–54) were enrolled in 19 (38.8%) and 22 (44.9%)studies, respectively. The remaining 8 studies (16.3%) hadpooled patients with I–IV stages (refs. 14–16, 26, 32, 38,46, 55; M0, M1; n ¼ 763). The main features of the eligiblestudies are summarized in Table 1.

Global Analysis of CTC Effects on SurvivalHRs for PFS/DFS were available in 35 studies (6, 9–

14, 16–20, 23–32, 34, 35, 37, 39, 41, 43, 44, 48–51, 53,54) accounting for 4,978 patients. In 12 studies (6, 13, 16,17, 19, 32, 34, 35, 37, 39, 41, 48), more than 1 HR wasextracted from each study because multimarkers or multi-time points were used to detect CTC in these studies andHRs were reported separately.

The estimated pooled HR for all studies showed a signif-icantly increased risk of disease progression in patients withCTC positivity (HR, 2.07; 95%CI, 1.80–2.39; P¼ 0.00; df¼48; random effects). As the heterogeneity among studieswas significant (P < 0.001, I2 ¼ 85.5%), random-effectsmodel was applied. To explore potential sources of hetero-geneity, we conducted meta-regression considering follow-ing covariates: publication year, sample size, detection

Figure 1. Flowchart of the strategyused for the selection of reports usedin our analysis. CTC, circulatingtumor cells.

Potentially relevant publications

identified and screened for retrieval

n = 733

Potentially relevant publications

435 publications excluded

323 not relevant to CTC or breast cancer

112 duplicates

Title reading

retrieved for detailed evaluation

n = 298

Abstract reading


77 review articles, 60 methodologies

22 CTC characteristics,

36 no survival data,1 book, 2 letter

11 commentaries,

Potentially relevant studies included

in the meta-analysis

n = 61

15 cells in bone marrow

12 not English,1 case report


Studies eligible for inclusion in

meta-analysis

Full text reading7 multiple publications,

3 no survival data,

2 small sample size

n = 49

Meta-Analysis of Circulating Breast Cancer Cells

www.aacrjournals.org Clin Cancer Res; 18(20) October 15, 2012 5703




Table 1. Characteristics of studies included in the meta-analysis

First author ofstudy (ref.), y

NO. ofpatients

Samplingtime Stage

Detectionmethod

Detectionrate, % Cutoff of CTCþ Outcome HR estimation

Stathopoulou (25), 2002 148 Baseline M0 RT-PCR 30 — DFS, OS Reported in textGaforio (26), 2003 92 Baseline M0, M1 Other ICC 62 1 CTC/10 mL PFS, OS Data extrapolatedWeigelt (27), 2003 94 Mid-therapy M1 RT-PCR 31 — PFS, OS Data extrapolatedXenidis (28), 2003 161 Posttherapy M0 RT-PCR 27 — DFS Data extrapolatedGiatromanolaki (29), 2004 100 Baseline M0 RT-PCR 33 — DFS Reported in textPierga (14), 2004 114 Baseline M0, M1 Other ICC 25 1 CTC/3 � 106 MNC DFS, OS Data extrapolatedMasuda (30), 2005 98 Baseline M0 RT-PCR 18 — DFS Reported in textBenoy (15), 2006 148 Baseline M0, M1 RT-PCR 15, 20a — OS Reported in textBudd (7), 2006 138 Mid-therapy M1 Cellsearch 25 5 CTC/7.5 mL OS Reported in textHayes (6), 2006 177 Baseline M1 Cellsearch 49 5 CTC/7.5 mL PFS, OS Reported in text

Mid-therapy 31Ntoulia (18), 2006 101 Baseline M0 RT-PCR 14 — DFS, OS Reported in textWiedswang (31), 2006 318 Posttherapy M0 Other ICC 10 1 CTC/1 � 107 MNC DFS, OS Reported in textWong (32), 2006 123 Mid-therapy M0, M1 Other ICC 34 4 CTC/sample

13 CTC/samplebPFS, OS Reported in text

Data extrapolatedc

Wulfing (11), 2006 35 Baseline M0 Other ICC 49 1 CTC/50 mL DFS, OS Data extrapolatedXenidis (12), 2006 167 Baseline M0 RT-PCR 22 — DFS, OS Reported in textApostolaki (23), 2007 214 Posttherapy M0 RT-PCR 21 — DFS, OS Reported in textCristofanilli (33), 2007 151 Baseline M1 Cellsearch 44 5 CTC/7.5 mL OS Reported in textIgnatiadis (34), 2007 185 Baseline M0 RT-PCR 34, 52d — DFS, OS Reported in textXenidis (35), 2007 119 Posttherapy M0 RT-PCR 18 — DFS, OS Reported in textBidard (8), 2008 37 Baseline M1 Other ICC 41 1 CTC/3 � 106 MNC OS Data extrapolatedDawood (36), 2008 185 Baseline M1 Cellsearch 38 5 CTC/7.5 mL OS Reported in textIgnatiadis (37), 2008 175 Baseline M0 RT-PCR 41, 8, 29e — DFS, OS Reported in textTkaczuk (38), 2008 105 Baseline M0, M1 Other ICC 56 1 CTC/15–20 mL OS Data extrapolatedYagata (9), 2008 38 Baseline M1 Cellsearch 29 5 CTC/7.5 mL PFS, OS Data extrapolatedApostolaki (24), 2009 216 Baseline M0 RT-PCR 25 — DFS, OS Reported in textBotteri (10), 2009 80 Baseline M1 Cellsearch 61 5 CTC/7.5 mL PFS, OS Data extrapolatedDaskalaki (39), 2009 165 Baseline M0 RT-PCR 55 — DFS, OS Reported in text

Posttherapy 52De Giorgi (40), 2009 102 Baseline M1 Cellsearch 50 5 CTC/7.5 mL OS Reported in text

Mid-therapy 21Liu (41), 2009 72 Baseline M1 Cellsearch 35 5 CTC/7.5 mL PFS Data extrapolated

Mid-therapy 28Marques (17), 2009 321 Baseline M0 RT-PCR 19 — DFS Reported in textSerrano (16), 2009 71 Mid-therapy M0, M1 Other ICC 66 6 CTC/10 mL PFS, OS Reported in text

Posttherapy 54Tewes (42), 2009 32 Mid-therapy M1 RT-PCR 52 — OS Data extrapolatedXenidis (13), 2009 437 Baseline M0 RT-PCR 41 — DFS, OS Reported in text

Posttherapy 33Bidard (43), 2010 115 Baseline M0 Cellsearch 23 1 CTC/7.5 mL DFS, OS Reported in textBidard (44), 2010 67 Baseline M1 CellSearch 52 5 CTC/7.5 mL PFS Data extrapolatedChen (19), 2010 50 Baseline M0 RT-PCR 54 — DFS Reported in textDe Giorgi (45), 2010 195 Baseline M1 Cellsearch 47 5 CTC/7.5 mL OS Reported in textHu (46), 2010 45 Baseline M0, M1 Other ICC 40 — OS Reported in textNakamura (47), 2010 107 Baseline M1 CellSearch 37 5 CTC/7.5 mL OS Reported in textGradilone (20), 2011 42 Baseline M1 RT-PCR 67 5 CTC/7.5 mL PFS Data extrapolatedConsoli (48), 2011 93 Baseline M1 CellSearch 47 5 CTC/7.5 mL PFS, OS Reported in text

Mid-therapy 44Giordano (49), 2011 517 Baseline M1 CellSearch 31 5 CTC/7.5 mL PFS, OS Reported in textHayashi (50), 2011 52 Baseline M1 CellSearch 60 5 CTC/7.5 mL PFS, OS Reported in textMego (21), 2011 292 Baseline M1 CellSearch 41 1 CTC/7.5 mL OS Data extrapolatedMolloy (51), 2011 82 Baseline M0 RT-PCR 20 — DFS Reported in textMuller (52), 2011 245 Baseline M1 CellSearch 50 5 CTC/7.5 mL OS Reported in text

(Continued on the following page)

Zhang et al.





method, time point of blood withdrawal, and tumor stage(M0 vs. M1). At univariate analysis, the explanatory variablethat influenced HR estimates was the tumor stage only (P¼0.001; Table 2).The HRs for OS were available in 40 studies (6–

16, 18, 21, 23–27, 31–40, 42, 43, 45–50, 52–55) account-ing for 5,832 patients. More than 1HR for OS was extractedin 9 studies (6, 13, 15, 16, 32, 37, 39, 45, 48) for the samereason as mentioned above.The pooled HR showed a significantly increased risk of

mortality in patients withCTCpositivity (HR, 2.45; 95%CI,2.07–2.90; P ¼ 0.00, df ¼ 50; random effects). The hetero-geneity among studies was significant (P < 0.001, I2 ¼70.8%). The results ofmeta-regression considering the samecovariates analyzed for OS showed that the only explana-tory variable that influenced HR estimates was also thetumor stage (P ¼ 0.01; Table 2).

CTC impact on survival in early-stage breast cancerThe HRs for DFS were available in 19 studies with early-

stage breast cancer. The estimated pooled HR showed that

the presence of CTC is a significantly increased risk ofdisease recurrence in early-stage patients (HR, 2.86; 95%CI, 2.19–3.75; P ¼ 0.00; df ¼ 28; random effects; Fig. 2A).The heterogeneity among studies was significant (P¼ 0.00,I2 ¼ 65.2%), and publication bias existed (PBegg ¼ 0.03;PEgger ¼ 0.03; Supplementary Fig. S1A). The trim-and-fillanalysis revealed that 6 studies might be missing and that ifthese were published, the adjusted HR would be 2.28 (95%CI, 1.71–3.03; P ¼ 0.00; df ¼ 34; random effects).

The HRs for OS were available in 13 studies with early-stage breast cancer. The estimated pooled HR showed thatCTC positivity was associated with a significantly increasedrisk of death (HR, 2.78; 95% CI, 2.22–3.48; P ¼ 0.00; df ¼17; fixed effects; Fig. 2B). The heterogeneity among studieswas absent (P ¼ 0.68, I2 ¼ 0.0%). Publication bias existed(PBegg ¼ 0.06; PEgger ¼ 0.03; Supplementary Fig. S1B). Thetrim-and-fill analysis revealed that 2 studies might be miss-ing and that if these were published, the adjustedHRwouldbe 2.66 (95%CI, 2.13–3.32; P¼ 0.00; df¼ 19; fixed effects).

One-way sensitivity analysis confirmed the stability ofour results (Supplementary Fig. S2A and B).

Table 1. Characteristics of studies included in the meta-analysis (Cont'd )

First author ofstudy (ref.), y

NO. ofpatients

Samplingtime Stage

Detectionmethod

Detectionrate, % Cutoff of CTCþ Outcome HR estimation

Pierga (53), 2011 267 Baseline M1 CellSearch 44 5 CTC/7.5 mL PFS, OS Reported in textReinholz (54), 2011 86 Baseline M1 RT-PCR 56–75; 23–38f — PFS, OS Reported in textSerrano (55), 2011 65 Posttherapy M0, M1 Other ICC 63, 46, 43, 31g 5 CTC/10 mL OS Data extrapolated

Abbreviations: ref., reference; CTC, circulating tumor cells;MNC,mononuclear cells; RT-PCR, reverse transcriptase polymerase chainreaction; ICC, immunocytochemistry; OS, overall survival; PFS, progression-free survival; DFS, disease-free survival.aDetection rates were for markers CK-19 and mammaglobin, respectively.bThe cutoff value was 4 CTC/sample in early-stage patients, whereas the cutoff value was 13 CTC/sample in metastatic patients.cHR for OS in early breast cancer was extrapolated from survival curves.dDetection rates were for markers CK-19 and HER2, respectively.eDetection rates were for markers CK-19, mammaglobin and HER2, respectively.fDetection rates were for markers CK-19 and mammaglobin at baseline, respectively.gDetection rateswere for the study time intervals of 1 to 5months,more than 5 to 12months,more than 12 to 24months, andmore than24 to 50 months, respectively.

Table 2. Results of meta-regression analysis exploring source of heterogeneity with progression-free/disease-free and overall survival.

PFS/DFS OSUnivariate analysis Univariate analysis

Covariates Coefficient SE P Coefficient SE P

Detection method 0.15 0.10 0.15 �0.02 0.08 0.83Time point of sampling �0.21 0.12 0.08 �0.13 0.09 0.15Stage of breast cancer �0.41 0.12 0.001 �0.25 0.10 0.01Sample size �0.001 0.001 0.25 0.0003 0.0006 0.64Publication year �0.07 0.04 0.07 �0.03 0.03 0.27

NOTE:Thedependent variable is the lnHR forPFS/DFSorOS fromeachstudy.Weights havebeenassignedaccording to the estimatedvariance of the lnHR. SE, standard error of the coefficient.






CTC impact on survival in metastatic breast cancerHRs for PFS were available in 12 studies with metastatic

breast cancer. The estimated pooled HR showed CTC pos-itivity was associated with a significantly increased risk ofprogression (HR, 1.78; 95% CI, 1.52–2.09; P ¼ 0.00; df ¼15; random effects; Fig. 2C). The heterogeneity amongstudies was moderate (P ¼ 0.03, I2 ¼ 43.8%). Publicationbias existed (PBegg¼ 0.82; PEgger¼ 0.02; Supplementary Fig.S1C). The trim-and-fill analysis revealed that 7 studiesmight be missing and that if these were published, theadjusted HR would be 1.48 (95% CI, 1.25–1.74; P ¼0.00; df ¼ 22; random effects).

The HRs for OS were available in 19 studies with meta-static breast cancer. The estimated pooled HR showed CTCpresencewas associatedwith a significantly increased risk ofdeath (HR, 2.33; 95%CI, 2.09–2.60; P¼ 0.00; df¼ 22; fixedeffects; Fig. 2D). The heterogeneity among studies wasabsent (P ¼ 0.33, I2 ¼ 9.8%). There was no publicationbias (PBegg ¼ 0.49; PEgger ¼ 0.31; Supplementary Fig. S1D).

One-way sensitivity analysis confirmed the stability ofour results (Supplementary Fig. S2C and D).

Influence of sampling timepoint anddetectionmethodAs shown in Table 3, subgroup analyses stratified by

detection method and time point of blood sampling con-firmed that the CTC presence was a strong prognosticator inall subgroups except the subgroup designated "other ICC."

The prognostic value of CTC for PFS/DFS was significantin the "RT-PCR" subgroup (HR, 2.58; 95% CI, 1.99–3.35)and the "CellSearch" subgroup (HR, 1.85; 95% CI, 1.53–2.25), whereas it was not significant in the subgroup "otherICC" (HR, 1.11; 95% CI 0.95–1.29; Fig. 3A). In the sub-group "other ICC," Serrano’’s study was overweighted(87.01%), which may influence the pooled result. There-fore, pooled analysis was conducted again omitting thisstudy, and the results showed that the subgroup HR did notcross the 1.0 line (HR, 2.42; 95% CI, 1.62–3.62), therebyalso indicating the significant prognostic impact of CTC

Figure 2. Meta-analysis of theHR for PFS/DFS andOS for early-stage andmetastatic breast cancer patients depending on the absence or presence ofCTC. A,DFS for early-stage patients (M0), random-effects analysis. B, OS for early-stage patients (M0), fixed-effects analysis. C, PFS for metastatic patients (M1),random-effects analysis. D, OS formetastatic patients (M1), fixed-effects analysis. Values of HR greater than 1.0 indicate that patients with CTChave a poorersurvival compared with those without CTC.

Zhang et al.





detection in this subgroup. The analysis of OS as endpointreveals similar results (Fig. 3B).

DiscussionThe present analysis is based on a large pool of clinical

studies (6,825 patients) and substantially differs from theother 2 meta-analysis published in 2011 (56, 57), whichconsidered smaller series, and evaluated studies analyzing

blood samples only by RT-PCR. Here, we identified 49studies that assessed the prognostic value of CTC detectionby RT-PCR as well as the CellSearch system or other ICCmethods. It provides evidence that the presence of CTC inperipheral blood is significantly associated with poorerprognosis both in early-stage and metastatic breast cancer.The pooled results are fairly stable andnot influenced by theCTC detectionmethod and time point of blood withdrawal(Table 3). In addition, this meta-analysis of pooled data

Table 3. Results of subgroup analyses for effects of CTC presence on progression-free/disease-free andoverall survival

PFS/DFS OS

Study selection dfHR (95% CI;random effects)

Test forheterogeneity (I2) df

HR (95% CI;random effects)

Test forheterogeneity (I2)

Detection methodRT-PCR 27 2.58 (1.99–3.35) 67.7% 19 2.38 (1.98–2.85) 0.0%CellSearch 12 1.85 (1.53–2.25) 56.3% 18 2.45 (2.10–2.85) 28.8%Other ICC 7 1.11 (0.95–1.29) 71.8% 11 2.35 (1.49–3.86) 73.0%

Time point of blood withdrawalBaseline 33 2.39 (1.96–2.91) 64.7% 34 2.37 (2.10–2.67) 5.6%Mid-therapy 6 1.98 (1.23–3.20) 89.6% 8 2.51 (1.86–3.40) 48.0%Posttherapy 7 1.62 (1.03–2.54) 84.3% 6 2.03 (1.23–3.67) 69.0%

All 48 2.07 (1.80–2.39) 85.5% 50 2.45 (2.07–2.90) 70.8%

Abbreviations: CTC, circulating tumor cells; RT-PCR, reverse transcriptase polymerase chain reaction; ICC, immunocytochemistry;OS, overall survival; PFS, progression-free survival; DFS disease-free survival.

Figure 3. Influence of the CTC detection method on the prognostic impact of CTC. Subgroup analysis of the HR for survival in the total population of patients(M0 and M1 stages) depending on the presence or absence of CTC, stratified by the detection method. A, PFS/DFS. B, OS.






confirmed that the presence of CTC represents a significantrisk factor for both PFS/DFS and OS, even after adjustmentfor publication bias.

The 2main approaches for CTC detection are ICC and RT-PCR. One advantage of ICC methods over nucleic acidmethods is the preservation of the cell during the process,which enables further characterization of CTC by additionalICC and molecular methods. Compared with ICC, RT-PCRseems to be more sensitive (58). In our analysis, the numberof studies using RT-PCR to detect CTC is nearly equal to thenumber of studies using ICC. Particularly, among thosestudies using ICC methods, the CellSearch system was usedin 18 studies (18/29, 62.1%). At present, the CellSearchsystem is the only assay cleared by the U.S. FDA for clinicaluse. Insubgroupanalysis,pooledHRs forPFS/DFSandOSarestably statistically significant in RT-PCR and CellSearch sub-groups. In the "other ICC" subgroup, pooled HR was notstatistically significant (HR,1.11; 95%CI,0.95–1.29), andwededuced that the heterogeneity might be the main cause forthis finding. Omitting the overweighted study, adjustedpooled HR was 2.42 (95% CI, 1.62–3.62) and the heteroge-neity disappeared. These results indicate that, although withdifferent sensitivity andspecificity,differentdetectionsystemsare able todetect prognostically relevantCTC inbreast cancer.However, it is desirable that future studies on the clinicalutility of CTC use standardized detection methods.

Most studies assessed the clinical significance of CTCdetected at baseline and showed that the presence of CTCat this time point was an independent prognostic factor.How about CTC detected during or after systemic therapy?Our present results indicate that CTC detection at mid-therapy or posttherapy cannot be only used for monitoringtherapeutic effects but has also prognostic relevance. In ourmeta-analysis, CTC detected before, during and after sys-temic therapy showed similar prognostic value (Table 3),suggesting that the presence of CTC at different time pointsis a strong prognostic indicator, which is not significantlyinfluenced by systemic therapy. The fact, that CTCdetectionat baseline (before therapy) is as predictive for an unfavor-able outcome as the detection during or after therapy,suggests that most CTC might be rather resistant to thecurrent forms of therapy in breast cancer. However, wecould not evaluate the effects of therapy on the prognosticvalue of CTC, which is an important parameter in prog-nostic studies (59). The reason for this drawback of ourstudy is that the therapeutic regimens were various amongmost studies included in this meta-analysis.

The prognosis of patients with early-stage and metastaticbreast cancer is obviously quite different. Therefore, mostpublished studies researched these 2 patient cohorts sepa-rately. However, in this meta-analysis, there are 8 studiesthat analyzed patients without separation into M0 and M1

stages. Meta-regression analysis revealed that it was tumorstage that was the main source of heterogeneity for thismeta-analysis. When these 8 studies were omitted, hetero-geneitywas decreased or disappeared, which confirmed thatearly-stage and metastatic breast cancer patients should beanalyzed separately to obtain more accurate results.

Some limitations of this meta-analysis need to be dis-cussed. First, our meta-analysis is based on data from trialswhose results have been published, and we did not obtainupdated individual patient data. Use of individual patientdata may further enhance the accuracy and reduce theuncertainty of the estimates. Second, significant heteroge-neity was found in our study. Although at meta-regression,only tumor staging was significantly associated with HRestimates, variability in definitions of end point, measure-ments, and experimental design may also contribute to theheterogeneity. Therefore, validation of the prognostic pow-er of CTC should be conducted through large multicenterprospective studies based on homogeneous populations.Third, publication bias is a concern. We tried to identify allrelevant data, but it is unavoidable that somedata could stillbemissing.Missing informationmay reflect negative resultsthat could reduce the prognostic power of CTC. However,we conducted the trim-and-fill analysis and found that evenif these missing studies were published the association ofCTC and poorer survival was still significant.

In conclusion, the present results support the notion of astrong prognostic value of CTC in early-stage andmetastaticbreast cancer patients. In the future, the detection of CTC atdifferent time points before and during systemic therapymight serve as a tool to guide treatment in cancer patients.To achieve clinical utility of CTC in breast cancer, moreintervention trials in which therapy decision making isbased on CTC results need to be initiated, such as theongoing multicenter S0500 trial led by the SouthwestOncologyGroup and theGermanDETECT-III trial focusingon anti-HER2 therapywith lapatinib in patients withHER2-positive CTC.

Disclosure of Potential Conflicts of InterestK. Pantel receives an honorarium from Speakers Bureau. No potential

conflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: L. Zhang, G. Wu, P. KlausAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): K. YangAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): L. Zhang, S. Riethdorf, T. Wang, K. Yang,G. Peng, J. Liu, P. KlausWriting, review, and/or revision of the manuscript: L. Zhang, S. Rieth-dorf, G. Wu, T. Wang, K. Yang, G. Peng, J. Liu, P. KlausAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): L. Zhang, T. Wang, G. Peng, J. LiuStudy supervision: G. Wu, P. Klaus

AcknowledgmentsThe authors thank Dr. Daniel F. Hayes for his critical comments on the

manuscript, and thank statistician Dr. Sheng Wei at Department of Epide-miology and Statistics for his critical support with statistics.

Grant SupportThis work was supported by National Natural Science Foundation of

China 81101573 (L. Zhang).The costs of publication of this article were defrayed in part by the

payment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received May 23, 2012; revised July 31, 2012; accepted August 8, 2012;published OnlineFirst August 20, 2012.

Zhang et al.





References1. BraunS, Vogl FD,NaumeB, JanniW,OsborneMP,CoombesRC, et al.

A pooled analysis of bonemarrowmicrometastasis in breast cancer. NEngl J Med 2005;353:793–802.

2. Ashworth TA. A case of cancer in which cells similar to those in thetumors were seen in the blood after death. Aust Med J 1869;14:146–9.

3. Harris L, Fritsche H, Mennel R, Norton L, Ravdin P, Taube S, et al.American Society of Clinical Oncology 2007 update of recommenda-tions for the use of tumor markers in breast cancer. J Clin Oncol2007;25:5287–312.

4. Cristofanilli M, BuddGT, EllisMJ, StopeckA,Matera J,MillerMC, et al.Circulating tumor cells, disease progression, and survival inmetastaticbreast cancer. N Engl J Med 2004;351:781–91.

5. Cristofanilli M, Hayes DF, Budd GT, Ellis MJ, Stopeck A, Reuben JM,et al. Circulating tumor cells: a novel prognostic factor for newlydiagnosed metastatic breast cancer. J Clin Oncol 2005;23:1420–30.

6. Hayes DF, Cristofanilli M, Budd GT, Ellis MJ, Stopeck A, Miller MC,et al.Circulating tumor cells at each follow-up timepoint during therapyof metastatic breast cancer patients predict progression-free andoverall survival. Clin Cancer Res 2006;12:4218–24.

7. BuddGT,Cristofanilli M, EllisMJ, StopeckA, BordenE,MillerMC, et al.Circulating tumor cells versus imaging - Predicting overall survival inmetastatic breast cancer. Clin Cancer Res 2006;12:6403–9.

8. Bidard FC, Vincent-Salomon A, Sigal-Zafrani B, Dieras V, Mathiot C,Mignot L, et al. Prognosis of women with stage IV breast cancerdepends on detection of circulating tumor cells rather than dissem-inated tumor cells. Ann Oncol 2008;19:496–500.

9. YagataH,NakamuraS, ToiM,BandoH,OhnoS,KataokaA.Evaluationof circulating tumor cells in patients with breast cancer: multi-institu-tional clinical trial in Japan. Int J Clin Oncol 2008;13:252–6.

10. Botteri E, Sandri MT, Bagnardi V, Munzone E, Zorzino L, Rotmensz N,et al. Modeling the relationship between circulating tumour cellsnumber and prognosis of metastatic breast cancer. Breast CancerRes Treat 2010;122:211–7.

11. Wulfing P, Borchard J, Buerger H, Heidl S, Zanker KS, Kiesel L, et al.HER2-positive circulating tumor cells indicate poor clinical outcome instage I to III breast cancer patients. Clin Cancer Res 2006;12:1715–20.

12. Xenidis N, Perraki M, Kafousi M, Apostolaki S, Bolonaki I, Stathopou-lou A, et al. Predictive and prognostic value of peripheral bloodcytokeratin-19mRNA-positive cells detected by real-time polymerasechain reaction in node-negative breast cancer patients. J Clin Oncol2006;24:3756–62.

13. Xenidis N, Ignatiadis M, Apostolaki S, Perraki M, Kalbakis K, Agelaki S,et al. Cytokeratin-19 mRNA-positive circulating tumor cells after adju-vant chemotherapy in patients with early breast cancer. J Clin Oncol2009;27:2177–84.

14. PiergaJY,BonnetonC, Vincent-SalomonA, deCremouxP,NosC,BlinN, et al. Clinical significance of immunocytochemical detection oftumor cells using digital microscopy in peripheral blood and bonemarrow of breast cancer patients. Clin Cancer Res 2004;10:1392–400.

15. Benoy IH,ElstH, PhilipsM,WuytsH,VanDamP,ScharpeS, et al. Real-time RT-PCR detection of disseminated tumour cells in bone marrowhas superior prognostic significance in comparison with circulatingtumour cells in patients with breast cancer. Br J Cancer 2006;94:672–80.

16. Serrano MJ, Sanchez-Rovira P, Delgado-Rodriguez M, Gaforio JJ.Detection of circulating tumor cells in the context of treatment Prog-nostic value in breast cancer. Cancer Biol Ther 2009;8:671–5.

17. Marques AR, Teixeira E, Diamond J, Correia H, Santos S, Neto L, et al.Detection of human mammaglobin mRNA in serial peripheral bloodsamples from patients with non-metastatic breast cancer is not pre-dictive of disease recurrence. Breast Cancer Res Treat 2009;114:223–32.

18. Ntoulia M, Stathopoulou A, Ignatiadis M, Malamos N, Mavroudis D,Georgoulias V, et al. Detection of Mammaglobin A-mRNA-positivecirculating tumor cells in peripheral blood of patients with operablebreast cancer with nested RT-PCR. Clin Biochem 2006;39:879–87.

19. Chen Y, Zou TN, Wu ZP, Zhou YC, Gu YL, Liu X, et al. Detection ofcytokeratin 19, humanmammaglobin, and carcinoembryonic antigen-positive circulating tumor cells by three-marker reverse transcription-

PCRassay and its relation to clinical outcome in early breast cancer. IntJ Biol Markers 2010;25:59–68.

20. Gradilone A, Naso G, Raimondi C, Cortesi E, Gandini O, Vincenzi B,et al. Circulating tumor cells (CTCs) inmetastatic breast cancer (MBC):prognosis, drug resistance and phenotypic characterization. AnnOncol 2011;22:86–92.

21. Mego M, De Giorgi U, Dawood S, Wang X, Valero V, Andreopoulou E,et al. Characterization of metastatic breast cancer patients with non-detectable circulating tumor cells. Int J Cancer 2011;129:417–23.

22. Parmar MK, Torri V, Stewart L. Extracting summary statistics toperform meta-analyses of the published literature for survival end-points. Stat Med 1998;17:2815–34.

23. Apostolaki S, Perraki M, Pallis A, Bozionelou V, Agelaki S, Kanellou P,et al. Circulating HER2 mRNA-positive cells in the peripheral blood ofpatients with stage I and II breast cancer after the administration ofadjuvant chemotherapy: evaluation of their clinical relevance. AnnOncol 2007;18:851–8.

24. Apostolaki S, Perraki M, Kallergi G, Kafousi M, Papadopoulos S,Kotsakis A, et al. Detection of occult HER2mRNA-positive tumor cellsin the peripheral blood of patients with operable breast cancer: eval-uation of their prognostic relevance. Breast Cancer Res Treat2009;117:525–34.

25. Stathopoulou A, Vlachonikolis I, Mavroudis D, Perraki M, KouroussisC, Apostolaki S, et al. Molecular detection of cytokeratin-19-positivecells in the peripheral blood of patients with operable breast cancer:evaluation of their prognostic significance. J Clin Oncol 2002;20:3404–12.

26. Gaforio JJ, Serrano MJ, Sanchez-Rovira P, Sirvent A, Delgado-Rodri-guez M, Campos M, et al. Detection of breast cancer cells in theperipheral blood is positively correlated with estrogen-receptor statusand predicts for poor prognosis. Int J Cancer 2003;107:984–90.

27. Weigelt B, Bosma AJ, Hart AA, Rodenhuis S, van't Veer LJ. Markergenes for circulating tumour cells predict survival in metastasizedbreast cancer patients. Br J Cancer 2003;88:1091–4.

28. Xenidis N, Vlachonikolis I, Mavroudis D, Perraki M, Stathopoulou A,Malamos N, et al. Peripheral blood circulating cytokeratin-19 mRNA-positive cells after the completion of adjuvant chemotherapy inpatients with operable breast cancer. Ann Oncol 2003;14:849–55.

29. Giatromanolaki A, Koukourakis MI, Kakolyris S, Mavroudis D, Kour-oussis C, Mavroudi C, et al. Assessment of highly angiogenic anddisseminated in the peripheral blood disease in breast cancer patientspredicts for resistance to adjuvant chemotherapy and early relapse. IntJ Cancer 2004;108:620–7.

30. Masuda TA, Kataoka A,Ohno S,Murakami S,Mimori K, Utsunomiya T,et al. Detection of occult cancer cells in peripheral blood and bonemarrow by quantitative RT-PCR assay for cytokeratin-7 in breastcancer patients. Int J Oncol 2005;26:721–30.

31. WiedswangG, Borgen E, Schirmer C, Karesen R, KvalheimG,NeslandJM, et al. Comparisonof the clinical significanceof occult tumor cells inblood and bone marrow in breast cancer. Int J Cancer 2006;118:2013–9.

32. Wong NS, Kahn HJ, Zhang L, Oldfield S, Yang LY, Marks A, et al.Prognostic significance of circulating tumour cells enumerated afterfiltration enrichment in early and metastatic breast cancer patients.Breast Cancer Res Treat 2006;99:63–9.

33. Cristofanilli M, Broglio KR,Guarneri V, JacksonS, FritscheHA, IslamR,et al. Circulating tumor cells in metastatic breast cancer: biologicstaging beyond tumor burden. Clin Breast Cancer 2007;7:471–9.

34. Ignatiadis M, Perraki M, Apostolaki S, Politaki E, Xenidis N, Kafousi M,et al. Molecular detection and prognostic value of circulating cytoker-atin-19 messenger RNA-positive and HER2 messenger RNA-positivecells in the peripheral blood of women with early-stage breast cancer.Clin Breast Cancer 2007;7:883–9.

35. Xenidis N, Markos V, Apostolaki S, Perraki M, Pallis A, Sfakiotaki G,et al. Clinical relevance of circulating CK-19 mRNA-positive cellsdetected during the adjuvant tamoxifen treatment in patients withearly breast cancer. Ann Oncol 2007;18:1623–31.

36. Dawood S, Broglio K, Valero V, Reuben J, Handy B, Islam R, et al.Circulating tumor cells in metastatic breast cancer from prognostic






stratification to modification of the staging system? Cancer 2008;113:2422–30.

37. IgnatiadisM, Kallergi G, Ntoulia M, Perraki M, Apostolaki S, Kafousi M,et al. Prognostic value of the molecular detection of circulating tumorcells using amultimarker reverse transcription-PCR assay for cytoker-atin 19, mammaglobin A, andHER2 in early breast cancer. Clin CancerRes 2008;14:2593–600.

38. Tkaczuk KH, Goloubeva O, Tait NS, Feldman F, Tan M, Lum ZP, et al.The significance of circulating epithelial cells in Breast Cancer patientsby a novel negative selection method. Breast Cancer Res Treat2008;111:355–64.

39. Daskalaki A, Agelaki S, Perraki M, Apostolaki S, Xenidis N, Statho-poulos E, et al. Detection of cytokeratin-19 mRNA-positive cells in theperipheral blood and bone marrow of patients with operable breastcancer. Brit J Cancer 2009;101:589–97.

40. De Giorgi U, Valero V, Rohren E, Dawood S, Ueno NT, Miller MC, et al.Circulating tumor cells and [18F]fluorodeoxyglucose positron emis-sion tomography/computed tomography for outcome prediction inmetastatic breast cancer. J Clin Oncol 2009;27:3303–11.

41. Liu MC, Shields PG, Warren RD, Dawood S, Ueno NT, Miller MC, et al.Circulating tumor cells: a useful predictor of treatment efficacy inmetastatic breast cancer. J Clin Oncol 2009;27:5153–9.

42. TewesM,AktasB,WeltA,MuellerS,HauchS,KimmigR,etal.Molecularprofiling and predictive value of circulating tumor cells in patients withmetastatic breast cancer: an option for monitoring response to breastcancer related therapies. Breast Cancer Res Treat 2009;115:581–90.

43. Bidard FC, Mathiot C, Delaloge S, Brain E, Giachetti S, de Cremoux P,et al. Single circulating tumor cell detection and overall survival innonmetastatic breast cancer. Ann Oncol 2010;21:729–33.

44. Bidard FC, Mathiot C, Degeorges A, Etienne-Grimaldi MC, Delva R,Pivot X, et al. Clinical value of circulating endothelial cells and circu-lating tumor cells in metastatic breast cancer patients treated first linewith bevacizumab and chemotherapy. Ann Oncol 2010;21:1765–71.

45. De Giorgi U, Valero V, Rohren E, Mego M, Doyle GV, Miller MC, et al.Circulating tumor cells andbonemetastasesasdetectedbyFDG-PET/CT in patientswithmetastatic breast cancer. AnnOncol 2010;21:33–9.

46. Hu Y, Fan L, Zheng J, Cui R, Liu W, He Y, et al. Detection of circulatingtumor cells in breast cancer patients utilizing multiparameter flowcytometry and assessment of the prognosis of patients in differentCTCs levels. Cytometry A 2010;77:213–9.

47. Nakamura S, Yagata H, Ohno S, Yamaguchi H, Iwata H, Tsunoda N,et al. Multi-center study evaluating circulating tumor cells as a surro-gate for response to treatment and overall survival in metastatic breastcancer. Breast Cancer 2010;17:199–204.

48. Consoli F, Grisanti S, Amoroso V, Almici C, Verardi R, Marini M, et al.Circulating tumor cells as predictors of prognosis in metastatic breast

cancer: clinical application outside a clinical trial. Tumori 2011;97:737–42.

49. Giordano A, Giuliano M, De Laurentiis M, Arpino G, Jackson S, HandyBC, et al. Circulating tumor cells in immunohistochemical subtypes ofmetastatic breast cancer: lack of prediction in HER2-positive diseasetreated with targeted therapy. Ann Oncol 2012;23:1144–50.

50. Hayashi N, Nakamura S, Tokuda Y, Shimoda Y, Yagata H, Yoshida A,et al. Prognostic value of HER2-positive circulating tumor cells inpatients with metastatic breast cancer. Int J Clin Oncol 2012;17:96–104.

51. Molloy TJ, Devriese LA, Helgason HH, Bosma AJ, Hauptmann M,Voest EE, et al. A multimarker QPCR-based platform for the detectionof circulating tumour cells in patientswith early-stage breast cancer. BrJ Cancer 2011;104:1913–9.

52. Muller V, Riethdorf S, Rack B, Janni W, Fasching PA, Solomayer E,et al. Prospective evaluation of serum tissue inhibitor of metallopro-teinase 1 and carbonic anhydrase IX in correlation to circulating tumorcells in patients with metastatic breast cancer. Breast Cancer Res2011;13:R71.

53. Pierga JY, Hajage D, Bachelot T, Delaloge S, Brain E, Campone M,et al. High independent prognostic and predictive value of circulatingtumor cells comparedwith serum tumormarkers in a large prospectivetrial in first-line chemotherapy for metastatic breast cancer patients.Ann Oncol 2012;23:618–24.

54. Reinholz MM, Kitzmann KA, Tenner K, Hillman D, Dueck AC, HobdayTJ, et al. Cytokeratin-19 and mammaglobin gene expression in circu-lating tumor cells from metastatic breast cancer patients enrolled inNorth Central Cancer Treatment Group trials, N0234/336/436/437.Clin Cancer Res 2011;17:7183–93.

55. Serrano MJ, Lorente JA, Delgado Rodriguez M, Fernandez A, Fernan-dez M, de la Torre C, et al. Circulating tumour cells in peripheral blood:potential impact on breast cancer outcome. Clin Transl Oncol2011;13:204–8.

56. Zhao S, Liu Y, ZhangQ, Li H, ZhangM,MaW, et al. The prognostic roleof circulating tumor cells (CTCs) detected by RT-PCR in breast cancer:a meta-analysis of published literature. Breast Cancer Res Treat2011;130:809–16.

57. Zhang L, Wu G, Pantel K. Detection of circulating tumor cells by RT-PCR significantly associated with poor prognosis in breast cancer.Breast Cancer Res Treat 2011;130:359–64.

58. Ring AE, Zabaglo L, Ormerod MG, Smith IE, Dowsett M. Detection ofcirculating epithelial cells in the blood of patients with breast cancer:comparison of three techniques. Br J Cancer 2005;92:906–12.

59. McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M, Clark GM.Reporting recommendations for tumor markers prognostic studies.J Clin Oncol 2005;23:9067–72.

Zhang et al.





2012;18:5701-5710. Published OnlineFirst August 20, 2012.Clin Cancer Res Liling Zhang, Sabine Riethdorf, Gang Wu, et al. Breast CancerMeta-Analysis of the Prognostic Value of Circulating Tumor Cells in

Updated version

10.1158/1078-0432.CCR-12-1587doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2012/08/20/1078-0432.CCR-12-1587.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/18/20/5701.full#ref-list-1

This article cites 59 articles, 14 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/18/20/5701.full#related-urls

This article has been cited by 21 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/18/20/5701To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-12-1587

http://clincancerres.aacrjournals.org/content/suppl/2012/08/20/1078-0432.CCR-12-1587.DC1

http://clincancerres.aacrjournals.org/content/18/20/5701.full#ref-list-1

http://clincancerres.aacrjournals.org/content/18/20/5701.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/18/20/5701


Documents

Meta-Analysis of the Prognostic Value of Circulating Tumor Cells … · Imaging, Diagnosis, Prognosis Meta-Analysis of the Prognostic Value of Circulating Tumor Cells in Breast Cancer